Thesis Diagnosing Cerebral Visual Dysfunctions in Children-Looking Beyond Visual Acuity and Visual Field

Diagnosing Cerebral
Visual Dysfunctions in
Children:
visual ﬁeld
visual acuity and
looking beyond
van der Zee

IJmkjen Janneke
Diagnosing Cerebral Visual Dysfunctions in
Children: looking beyond visual acuity and
visual field
IJmkjen Janneke van der Zee

De uitgave van dit proefschrift is mogelijk gemaakt door de financiële onder-
steuning van: Koninklijke Visio
https://epubs.ogc.nl/?epub=y.vanderzee
Cover design: Optima Grafische Communicatie, Rotterdam

Printed by: Optima Grafische Communicatie, Rotterdam
ISBN: 978-94-6361-255-5
Copyright © 2019, IJmkjen Janneke van der Zee

All rights reserved. No part of this thesis may be reproduced or transmitted
in any form or by any means, including photocopying, recording, or other
electronic or mechanical methods, without the prior written permission of the
holder of the copyright.
Diagnosing Cerebral Visual Dysfunctions in Children:
looking beyond visual acuity and visual field
Het vaststellen van cerebrale visuele stoornissen bij kinderen:

verder kijken dan de gezichtsscherpte en het gezichtsveld
Proefschrift
ter verkrijging van de graad van doctor aan de

Erasmus Universiteit Rotterdam
op gezag van de
rector magnificus
Prof.dr. R.C.M.E. Engels
en volgens besluit van het College voor Promoties.

De openbare verdediging zal plaatsvinden op
woensdag 15 mei 2019 om 13:30 uur
door
IJmkjen Janneke van der Zee

geboren te Oosstellingwerf
Promotiecommissie
Promotor: Prof.dr. H.M. Evenhuis
Overige leden: Prof.dr. N.E. Schalij-Delfos

Dr. C.E. Catsman-Berrevoets
Prof.dr. H.A.M. Middelkoop
Copromotor: Dr.ir. J.J.M. Pel

Contents
Chapter 1 General introduction 7
Chapter 2 Should we add visual acuity ratios to referral criteria for 19

potential cerebral visual impairment?
Chapter 3 Visual motion perception: brain localisation, terms and 37

paradigms
Chapter 4 Applicability of computerised motion perception tasks 45

in clinical practice for children. A systematic review
Appendix 1 67
Appendix 2 75
Chapter 5 Chronological age versus developmental age in 99

evaluating patients’ performances on motion perception
tests
Chapter 6 Clinical assessment of visual motion perception 121

in children with brain damage and the risk of
overdiagnosis
Chapter 7 Object recognition and dorsal stream vulnerabilities in 135

children with early brain damage
Appendix 1 151
Chapter 8 Remote eye tracking assesses age dependence 157

processing of coherent motion in typically developing
children
Chapter 9 Gestalt perception in children with visual impairments: 173

item-specific performance and looking behaviour
Chapter 10 General discussion 195

Appendix 1 209
Summary 213
Samenvatting 223
Dankwoord 235
About the author 241
PhD Portfolio 245
Chapter
Chapter 1
1
General introduction
General introduction 9
Introduction
1
To understand and interact with the world around us, we humans are highly
dependent on the quality of our senses (vision, hearing, touch, smell, and
taste), cognition and motor abilities. An extensive network is involved in the
processing of sensory information, and includes the sensors, such as eyes,
ears and skin, and our brain. People are inclined to use their own sensorial
input and interpretation of the world around them as reference. This makes it
difficult or even impossible for people with effectively working senses to imag-
ine, understand and acknowledge what it is like to be sensorially impaired,
as in people with a visual impairment. To understand and recognise sensorial
impairments, we need knowledge of sensorial systems and functions, and the
effect that function loss or impairment has on interaction with our environ-
ment. Vision in particular, is regarded as the primary sense for developing
many skills. Therefore, in this thesis, I focus on visual impairments in children.
Children that need special support because of a visual

impairment
1
It is generally recognised that congenital low vision or blindness negatively
influences a child’s development of cognitive, motor, communication and social
[1-6]
skills. To stimulate development in children with low vision and blindness,
early intervention and habilitation is preferred. In the Netherlands, early
intervention and habilitation are offered by special services for people with
visual impairments. Because such services are funded by health insurances
- originally the Exceptional Medical Expenses Act (AWBZ) and since January
2015 the mandatory health insurances -, eligibility requirements have to be
met. These eligibility requirements, originally designed for adults, are based
on the definitions of low vision and blindness as described in the International
th
Statistical Classification of Diseases and Related Health Problems 10 Revi-
[7]
sion (ICD-10) and involve two visual functions: visual acuity and visual field
(Table 1).
Visual acuity is the ability to see or resolve details of a retinal image when
tested at maximum contrast (black/white), whereas visual field is defined as
the area that can be seen when looking straight forward.
1 Although this term suggests that there is a total lack of vision, there may be some
vision left.
10 Chapter 1
Table 1. Definition of low vision and blindness as described in ICD-10 H54, 1993.
Definition
In the better eye with best possible
correction Low vision Blindness
Visual acuity less than 6/18 and equal to or better less than 3/60
than 3/60
or or
Visual field less than 30 degrees and less than 10
equal to or better than 10 degrees degrees
Although visual acuity and visual field are important properties, there is
more to this than meets the eye when it comes to visual perception, the
ability to recognise and interpret visual aspects of the surrounding environ-
ment. Visual perception highly depends on the development of an extensive
brain network and several visual functions, which are assigned to different
brain areas. It starts with processing light that falls on the retina. The retinal
information is conveyed via the optic nerve and lateral geniculate nucleus
(LGN) to the primary visual cortical areas, such as the occipital lobe (V1), and
extrastriate areas (V2). Next, the visual brain can be described in a simpli-
fied hierarchical model with two anatomically and functionally distinguishable
[8, 9]
main streams: the ventral stream and the dorsal stream. The ventral
[10, 11]
stream ends in the inferior temporal cortex and is involved in object
[11]
and face recognition. The dorsal stream leads into the posterior parietal
cortex and extends towards the anterior parietal and posterior frontal cortex,
[12, 13] [10]
and is involved in motion perception, visuomotor integration, visual
[14, 15] [16, 17]
attention, and object recognition in suboptimal representations.
Habilitation specialists generally acknowledge that a lack or loss of cerebral
visual functions can impede a child’s development and participation as well.
Thus cerebral visual disfunctions, for example, may hinder the ability to visu-
[18-20]
ally recognise persons or objects or to perceive motion. Difficulties in the
perception of a person’s identity hinder discrimination of familiar and unfamiliar
persons, and proper social communication. Object recognition problems not
only complicate learning by means of objects, but make exercises in school-
books, which often contain illustrations, difficult or even impossible to do.
Visual spatial attention and motion perception are essential for understanding
and predicting the constantly changing world around us. For blind people or
people with impaired visual motion perception, crossing a street without aids
such as traffic lights with a signal system is very risky, especially now that
car engines become more silent. The combination of spatial awareness and
motion perception also helps us to direct our object handling, such as catching
a ball or pouring a cup of tea. From a social interaction perspective, it also

enables us to predict what other people intend to do.
1
Visual impairments, either mild, moderate or severe, can have a major im-
pact on daily life functioning. When children with mild visual impairments grow
older, they may experience more difficulties when it comes to educational
methods towards more self-regulation, and demands on communicative and
social skills. This may lead to an increase of referrals to specialised services
[21]
and special schools, similar as has been suggested for children with autism.
A thorough but tailored diagnostic process and the ability to diagnose mild
cerebral visual impairments become of a greater importance.
Diagnosing cerebral visual dysfunctions in children

In 2005, when we started this project, the ophthalmological guideline for refer-
ral to special services for people with visual impairments exclusively mentioned
low and subnormal vision (visual acuity ≤ 0.5), visual field defects, reading
problems as result of high hyperopia (farsightedness) and glare (interference
[22] [23]
of light) as referral criteria. In the updated guideline of 2011 specific at-
tention is asked for cerebral visual impairment, but no specific referral criteria
are given and only risk groups with ‘mental retardation’, low birth weight, and
cerebral paresis were mentioned. Once children are referred, they enter a
diagnostic process, which routinely includes history taking and an assessment
of oculomotor functions, visual acuity, visual fields and contrast sensitivity.
Only if cerebral visual dysfunctions (e.g. problems with visual attention, object
or face recognition, motion perception or visuomotor abilities) are suspected
to cause problems in daily life, further assessment is done.
For the decision to start further assessment, the cause and type of visual
impairment in combination with the experienced problems in daily practice,
as reported by parents or teachers, are considered relevant. Visual impair-
ments are grossly divided into cerebral (or central) visual impairments (CVI)
and ocular (or peripheral) visual impairments (OVI). CVI is mainly caused
by brain damage or brain dysfunction. Low vision and visual field loss may
also be caused by brain damage (CVI) as well as by ocular disorders (OVI).
It is important to realise that children with CVI can have normal visual acu-
[24]
ity and normal visual fields. Currently, the following groups of children
are considered at risk for cerebral visual dysfunctions: children with cerebral
[24]
palsy, children with intellectual disability and prematurely born children.
The prevalence of CVI in developed and developing countries is still unknown.
Brain damage or brain dysfunction is considered the most common cause of
[25, 26]
low vision and blindness in developed countries.
12 Chapter 1
A few reasons why epidemiological knowledge on cerebral visual dysfunc-

tions in children is limited are:
1. the effect of brain plasticity: the course of development in children with
early brain damage is unpredictable, a certain brain area can take over the
function of affected brain areas. As a result, the presentation of problems,
including visual problems, may be heterogeneous in a group with compa-
rable brain damage.
2. a-specific testing of visual functions: several brain functions interact and
thereby making it difficult to disentangle the specific function that caused
the abnormal or weak performance, e.g. a child might be unable to name a
picture due to a language problem, a visual perceptual problem or did not
pay attention to the picture;
3. the difficulty in recognition of problems: cerebral visual dysfunctions may
be difficult to recognise in children with other prominent problems, such as
motor or intellectual disabilities;
4. the availability and applicability of diagnostic tests: not all cerebral visual
functions can be assessed with diagnostic tests with quantitative outcomes;
moreover, many tests are not applicable in very young children;
5. the limited accessibility of certain groups of children to expert services:
many children with unrecognised cerebral visual dysfunctions, especially
among those visiting special schools, rehabilitation centres, and day care
centres for children with intellectual disabilities are not referred yet.
Aim of this scientific project and contents of this thesis

Without a valid basis of scientific knowledge on the epidemiology of cerebral
visual dysfunctions and evidence on their impact on daily life, adaptation of the
eligibility requirements is very unlikely. As a result, children with cerebral visual
dysfunctions, especially those without a diagnosis of low vision or blindness,
are at risk for denial of the support they are in need of. Prior to investigating
the prevalence of cerebral visual dysfunctions, it is required to get a clear
overview of the wide spectrum of functional impairments. During a meeting of
a Dutch expert working party in 2006, professionals involved in the signalling,
referral and early assessments, such as youth healthcare physicians, pae-
diatricians, intellectual disability physicians and paediatric ophthalmologists,
indicated that they needed clear guidelines how to recognise children at risk
for cerebral visual dysfunction, apart from the above-mentioned risk groups.
They requested a test or screening questionnaire, because most paediatri-
cians and youth healthcare physicians were unfamiliar with even the existence
of cerebral visual impairment. Many professionals had difficulties with the
[27]
recognition of typical behaviours and problems associated with it. Profes-
sionals involved in the expert diagnostic process indicated multiple problems,
1
such as the lack of neuropsychological tests for children under the age of
four years. As a result, they are forced to rely on structural observations with
no or limited quantitative outcome measures. In addition, they stressed the
lack of quantitative tests for specific cerebral visual functions applicable to
children older than four years, with a special emphasis on motion perception
tests. Still, most tests have been developed for normally sighted children,
excluding children with low vision. Therefore, the effect of low vision on test
outcomes is unknown. The presence of a cognitive impairment can also affect
the performance of perception tests. That raises a very practical question:
what entry should be used when using norm tables: the developmental or the
chronological age of a child?
Study approach
Neuropsychology aims to understand how behaviour and cognition are in-
fluenced by brain functioning. During several tests, the processing of visual
information plays an important role, because many tasks consist of visual
representations. Here, the ability to visualize a complete whole even when
presented incomplete information or a partial picture, also known as Gestalt,
is crucial for optimum performance. The general aim of my thesis was to
provide new information to advance evidence-based diagnostic procedures in
children at risk for cerebral visual dysfunction (children with brain damage),
which ultimately should be efficient (burden and costs, use of valid tests).
I decided to focus on school-aged children with a developmental age be-
tween 4 and 7 years. The selected tests are relatively easily to integrate in
the assessment and provide a quantitative outcome of cerebral visual func-
tions associated with dorsal stream vulnerability of the visual system. Here, I
consider object recognition under suboptimal conditions, motion perception,
visuomotor abilities and visual attention as dorsal stream functions.
Data collection
We assessed oculomotor functions, visual acuity, visual fields and contrast
sensitivity, performance age, object recognition under suboptimal conditions,
motion perception, visuomotor abilities and visual attention in children at
risk for cerebral visual dysfunction (children with brain damage) as well as
in children not at risk (normally developing children and children with ocular
disorders).
14 Chapter 1
In Chapter 2, we focus on early recognition by e.g. youth healthcare physi-

cians, paediatricians, intellectual disability physicians, ophthalmologists. We
studied whether the crowding ratio (ratio between single optotype and multiple
optotype acuity), or ratios between grating and optotype acuity, could help to
discriminate between children considered at risk for cerebral visual dysfunc-
tions (children with brain damage) and children not at risk for cerebral visual
dysfunction (normally developing children and children with ocular disorders),
given the different detection goals in the care supply chain.
In Chapter 3-6 we focus on the expert assessment of cerebral visual dys-
functions. Chapter 3 is a short introduction into motion perception, including
current insight into its representation in the brain, different aspects of motion
perception, and available diagnostic approaches. In chapter 4, a systematic
review is presented of the scientific literature on motion perception tasks, to
evaluate whether there is substantial evidence that motion perception tasks
are valid and reliable to be used in clinical practice as well as epidemiological
research, i.e. tasks are applicable, norm values are available and reliable, and
motion perception problems have been demonstrated in at risk groups.
Consequently, motion perception becomes a central part of my studies.
Based on neuropsychological assessment, I address the following topics: 1. the
relation between chronological age, performance age and motion perception
task outcomes and whether we should use chronological age or performance
age as an entry of norm tables in the diagnostic process in at risk groups
(Chapter 5), and 2. whether specific (isolated) motion perception problems
are present in children with brain damage (Chapter 6).
Chapter 7 addresses object recognition, and specifically the question
whether children with brain damage and problems with object recognition
under suboptimal conditions either have a general dorsal stream dysfunction
or have selectively impaired dorsal stream functions.
Chapter 8 and 9 result from a collaboration with the Vestibular and oculo-
motor research group at the department of Neuroscience, Erasmus MC. The
rationale for this collaboration was the availability of a method to test visual
processing functions in children aged younger than 4 years and in children
as well as adults with severe intellectual disabilities. The method depends on
automated tracking of reflexive and goal-directed eye movements towards
specific visual stimuli using a remote eye tracker system. In Chapter 8,
the viewing behaviour to three different coherent motion stimuli is assessed
[28]
in terms of reaction time and fixation quality. The data are obtained in
typically developing children aged 0-12 years and may serve as normative
data. In Chapter 9, I investigated the ability of children with OVI and CVI to
visually integrate parts of the representation into a whole. In clinical practise,

the subtest Gestalt Closure of the Kaufman Assessment Battery for Children
1
second edition (KABC-II) is frequently used for this purpose. Here, this test
was combined with eye tracking to be able to combine the performance with
the obtained viewing behaviour.
Finally, in the general discussion, Chapter 10, the outcomes and limitations
of this study are discussed, as well as implications for clinical practice and
research.
Remarks on terminology
Terms like cortical visual impairment or cerebral visual impairment (CVI) can
cause confusion about the visual functions of interest. The term visual impair-
ment is often used to indicate low vision and blindness. In the above text and
in the following chapters the term visual impairment is used for both low vision
and blindness and cerebral visual dysfunctions.
16 Chapter 1
References
1. Brambring, M., Divergent development of verbal skills in children who are blind or
sighted. Journal of Visual Impairment & Blindness, 2007. 101(12): p. pp.
2. Houwen, S., et al., Gross motor skills and sports participation of children with
visual impairments. Res Q Exerc Sport, 2007. 78(2): p. 16-23.
3. Houwen, S., et al., Motor skill performance of school-age children with visual
impairments. Dev Med Child Neurol, 2008. 50(2): p. 139-45.
4. Dyck, M.J., et al., Emotion recognition/understanding ability in hearing or vision-
impaired children: do sounds, sights, or words make the difference? Journal of
Child Psychology and Psychiatry, 2004. 45(4): p. 789-800.
5. Tadić, V., L. Pring, and N. Dale, Are language and social communication intact in
children with congenital visual impairment at school age? Journal of Child Psy-
chology and Psychiatry, 2010. 51(6): p. 696-705.
6. O’Donnell, L.M. and R.L. Livingston, Active exploration of the environment by
young children with low vision: A review of the literature. Journal of Visual Im-
pairment & Blindness, 1991. 85(7): p. 287-291.
7. World Health Organization, International Statistical Classification of diseases,
injuries and causes of death, tenth revision., in H54 Blindness and low vision.
1993, WHO: Geneva.
8. Merigan, W.H. and J.H. Maunsell, How parallel are the primate visual pathways?
Annu Rev Neurosci, 1993. 16: p. 369-402.
9. Goodale, M.A. and A.D. Milner, Separate visual pathways for perception and ac-
tion. Trends in Neurosciences, 1992. 15(1): p. 20-5.
10. James, T.W., et al., Ventral occipital lesions impair object recognition but not
object-directed grasping: an fMRI study. Brain, 2003. 126(Pt 11): p. 2463-75.
11. Atkinson, J., et al., Neurobiological models of visuospatial cognition in children
with Williams syndrome: measures of dorsal-stream and frontal function. Devel-
opmental Neuropsychology, 2003. 23(1-2): p. 139-72.
12. Braddick, O.J., et al., Brain areas sensitive to coherent visual motion. Perception,
2001. 30(1): p. 61-72.
13. Stiers, P., et al., Mapping multiple visual areas in the human brain with a short
fMRI sequence. Neuroimage, 2006. 29(1): p. 74-89.
14. Pollmann, S., et al., Separating distractor rejection and target detection in poste-
rior parietal cortex--an event-related fMRI study of visual marking. Neuroimage,
2003. 18(2): p. 310-23.
15. Marini, F. and C.A. Marzi, Gestalt Perceptual Organization of Visual Stimuli Cap-
tures Attention Automatically: Electrophysiological Evidence. Front Hum Neurosci,
2016. 10: p. 446.
16. Vuilleumier, P., et al., Multiple levels of visual object constancy revealed by event-
related fMRI of repetition priming. Nat Neurosci, 2002. 5(5): p. 491-9.
17. Eger, E., et al., Mechanisms of top-down facilitation in perception of visual objects
studied by FMRI. Cereb Cortex, 2007. 17(9): p. 2123-33.
18. Ahmed, M. and G.N. Dutton, Cognitive visual dysfunction in a child with cerebral
damage. Dev Med Child Neurol, 1996. 38(8): p. 736-9.
19. Dutton, G.N., R.E. Day, and D.L. McCulloch, ‘Who is a visually impaired child? A
model is needed to address this question for children with cerebral visual impair- 1
ment’. Dev Med Child Neurol, 1999. 41(3): p. 212-3.
20. Jacobson, L., et al., Visual and perceptual characteristics, ocular motility and stra-
bismus in children with periventricular leukomalacia. Strabismus, 2002. 10(2): p.
179-83.
21. Besseling, J., et al., Toename gebruik ondersteuning voor jongeren met een
gezondheidsbeperking. 2007, TNO KvL: Hoofddorp.
22. Oogheelkundige richtlijn. Verwijzing van slechtzienden en blinden. 2004, Neder-
lands Oogheelkundig Gezelschap.
23. van Rens, G., H. Vreeken, and R. van Nispen, Richtlijn Visusstoornissen, revalidatie
en verwijzing (Ophthalmological guideline Visual impairments, Rehabilitation and
Referral). 2011, Nederlands Oogheelkundig Gezelschap (Dutch Ophthalmological
Society).
24. Dutton, G.N., E.C. McKillop, and S. Saidkasimova, Visual problems as a result of
brain damage in children. Br J Ophthalmol, 2006. 90(8): p. 932-3.
25. Gronqvist, S., et al., Association between visual impairment and functional and
morphological cerebral abnormalities in full-term children. Acta Ophthalmol
Scand, 2001. 79(2): p. 140-6.
26. Ravenscroft, J., et al., A novel method of notification to profile childhood visual
impairment in Scotland to meet the needs of children with visual impairment.
British Journal of Visual Impairment, 2008. 26(2): p. 170-189.
27. Evenhuis, H., et al., Detection of Childhood Visual Impairment in At-Risk Groups.
Journal of Policy and Practice in Intellectual Disabilities, 2007. 4(3): p. 165-169.
28. Pel, J., et al., Effects of visual processing and congenital nystagmus on visually
guided ocular motor behaviour. Dev Med Child Neurol, 2011. 53(4): p. 344-9.
Chapter
Chapter 2
2
Should we add visual acuity ratios

to referral criteria for potential
cerebral visual impairment?
van der Zee, Y.J., P. Stiers, and H.M. Evenhuis, Should we add
visual acuity ratios to referral criteria for potential cerebral visual
impairment? J Optom, 2017. 10(2): p. 95-103.
20 Chapter 2
Abstract
Purpose
To determine whether the assessment of visual acuity ratios might improve
the referral of children with (sub)normal visual acuity but at risk of cerebral
visual impairment.
Methods
In an exploratory study, we assessed visual acuity, crowding ratio and the
ratios between grating acuity (Teller Acuity Cards-II) and optotype acuity
(Cambridge Crowding Cards) in 60 typically developing school children (mean
age 5y8m ± 1y1m), 21 children with ocular abnormalities only (5y7m ± 1y9m)
and 26 children with (suspected) brain damage (5y7m ± 1y11m). Sensitivities
and specificities were calculated for targets and controls from the perspective
of different groups of diagnosticians: youth health care professionals (target:
children with any visual abnormalities), ophthalmologists and low vision ex-
perts (target: children at risk of cerebral visual impairment).
Results
For youth health care professionals subnormal visual acuity had the best
sensitivity (76%) and specificity (70%). For ophthalmologists and low vision
experts the crowding ratio had the best sensitivity (67%) and specificity (79
and 86%).
Conclusion
Youth health care professionals best continue applying subnormal visual acuity
for screening, whereas ophthalmologists and low vision experts best add the
crowding ratio to their routine diagnostics, to distinguish children at risk of
visual impairment in the context of brain damage from children with ocular
pathology only.
Key words: visual acuity; visual acuity ratios; brain damage; referral criteria
Visual acuity ratios as referral criteria? 21
Introduction
Childhood cerebral visual impairment often goes unrecognised, especially

in case of mild brain damage which has remained obscure, delaying reha- 2
bilitation to support normal development. Dutch youth health care physicians
(physicians providing preventive services), paediatricians and paediatric oph-
thalmologists indicated difficulties in recognising higher visual problems and
expressed the need for clear diagnostic guidelines, whereas experts of low
vision services indicated that it is not efficient to perform time-consuming as-
sessments of cerebral visual functions, e.g. object and face recognition, in all
referred children. They expressed the need for a quantitative, evidence-based
[1]
measure that could help with the selection.
[2]
Hård et al suggested that ‘crowding’ might be an indicator for visual
perception problems. Crowding is the phenomenon of impaired discrimination
[3]
of visual stimuli in the middle of other stimuli. Abnormal crowding may
show in daily life as difficulties handling complex visual situations, for example
finding a toy in a toy box or coping with busy environments like supermarkets;
reading may be hampered, as well as visual acuity testing with routinely used
[4]
cards. It may either concern a selective attention problem, negatively in-
[5]
fluencing specific higher perceptual functions, or it might be a failure in the
[6]
neural development of small integration fields. In both instances, crowding
might predict (unidentified) brain damage and, indirectly, visual perception
[7-9]
problems.
Children without any neurological impairment who have developmental
[10, 11]
visual problems, like strabismic amblyopia, are also at risk of crowding
as well as of cerebral visual impairments, for example motion-defined form
[12] [13]
perception. All these visual problems in both children with amblyopia
and children with cerebral visual impairment are associated with anatomical
[13, 14]
and functional abnormalities of the brain’s visual system.
If, indeed, abnormal crowding is more common in children with brain dam-
age than in other children, a small addition to routine visual acuity assessment
might improve the detection of children at risk of cerebral visual impairment.
Instead of a single acuity test, two tests at the same test distance would have
to be performed, namely a test with a single optotype and a test with multiple
optotypes, after which the difference can be expressed as a ratio. Studies
comparing grating acuity (Teller Acuity Cards) with optotype acuity (single
and/or multiple optotypes) also indicate that larger differences between acu-
[11]
ities are found in children with strabismic amblyopia and children with
22 Chapter 2
[15]
brain damage than in normally developing children and children with ocular
[11, 15]
disorders.
Therefore, the first aim of this exploratory study was to assess whether
the crowding ratio, estimated by the Cambridge Crowding Cards (optotypes),
and the ratio between grating acuity and optotype acuity, differentiate better
than visual acuity alone between children without any neurological and ocular
abnormalities, children with ocular disorders only, and children with (indica-
tions of) brain damage.
In clinical practice, different levels in the care supply chain, i.e. youth health
care physicians or nurses, ophthalmologists, and experts in low vision services,
see different populations and have different detection goals and indications for
referral.
Therefore, our second aim was to study which parameter (visual acuity
alone, different ratios) for each level has the best sensitivity and specificity to
meet their detection goal.
Methods
For this cross-sectional study, a heterogeneic group of 107 children in the

age range 3-12 years was recruited in the Rotterdam area through primary
schools, two rehabilitation centres, and the Rotterdam Eye Hospital. For all
participants, informed consent was obtained from their parents or guardians.
Medical files and the children’s histories were used to categorise participants
as ‘typically developing without indications of neurological or ocular abnor-
malities’ (group 1), ‘ocular abnormalities without indications of neurological
abnormalities’ (group 2) and ‘(indications of) brain damage with or without
ocular abnormalities’ (group 3). Cerebral palsy (CP) as well as visual problems
not explained by ocular abnormalities were considered indications of brain
damage. Children without neurological symptoms with high myopia or hypero-
pia (spherical error ≥ 6D) were included in group 2, and those with smaller
refractive errors in group 1.
To meet study aim 2, we had to compose different control and target groups
per discipline. In order to refer properly, youth health care physicians want to
identify all children with visual problems irrespective of their cause (target:
group 2 & 3, controls: group 1), whereas ophthalmologists and specialists at
low vision services want to differentiate between children at risk of cerebral vi-
sual impairment (target: group 3) and children with ocular abnormalities only
(controls ophthalmologist: group 1 & 2, controls low vision centre: group 2).
To this end, the following child characteristics were collected from the clinical
files: visual acuity, refractive errors, use of glasses or lenses, strabismus, con-
trast sensitivity, other ocular abnormalities, and if present, aetiology of brain
damage. Visual fields were checked using the confrontation test. Although 2
refractive errors and strabismus are associated with crowding problems, these
children were not excluded from the analyses, because this reflects clinical
practice.
The study adhered to the Declaration of Helsinki and the protocol has been
approved by the Medical Ethical Testing Committee of the Erasmus University
Medical Center Rotterdam (MEC nr: 2006-056).
In the period June 2006-September 2010, visual acuity was assessed in all
[16]
participants with the Cambridge Crowding Cards to assess uncrowded and
[17]
crowded optotype acuity, and the Teller Acuity Cards-II to assess grating
acuity. The uncrowded Cambridge Crowding Cards show a single letter and the
crowded cards show a target letter surrounded by four letters, with a distance
between letters of half a letter width. To minimise the effect of differences in
test distance, the Teller Acuity Cards as well as the Cambridge Crowding Cards
were presented at a distance of 3 meters. Outcomes of the Teller Acuity Cards
were corrected for the greater than recommended presentation distance, using
the following formula: cycl/deg = (test distance/57)*cycl/cm. Children were
assessed binocularly, if applicable wearing their own glasses. Grating acuity
was assessed first, followed by single and crowded optotype acuities. During
Teller Acuity Card assessment, children were asked to actively point out the
location of the stripes (left or right), and during optotype assessment, they
had to name or match the target letter. All tests were started with coarse grat-
ings/large letter sizes. Finer gratings/smaller letter sizes were presented until
the child was unable to identify the correct position/letter. Maximally three
cards were presented per visual acuity level and the visual acuity threshold
was defined as the highest visual acuity level with two out of three correctly
identified grating positions/letters. All assessments were performed by trained
orthoptists.
To enable comparisons between grating and optotype acuities, we expressed
all acuities in cycl/deg. We determined two cut-off values for crowded visual
acuity, because in youth health care, children with ‘subnormal vision’ should be
referred for ophthalmological examination, whereas ‘low vision’ is a traditional
eligibility requirement for low vision services. Subnormal visual acuity was
defined as a crowded acuity lower than 20 cycl/deg (Snellen equivalent 6/9)
for 3½-4½ year olds and lower than 30 cycl/deg (Snellen equivalent 6/6) for
older children (norms from The Cambridge Crowding Cards leaflet), whereas
24 Chapter 2
low vision was defined as a crowded acuity equal to or lower than 10 cycl/deg
[18]
(Snellen equivalent 6/18).
Crowding was quantified as a crowding ratio (CR), which is defined as the
acuity presenting a single optotype, divided by the acuity presenting several
optotypes (linear optotype acuity). Because linear acuity tests mostly result in
lower visual acuities than single optotype testing, the crowding ratio is usually
[16]
higher than 1. In typically developing children, there is a developmental
effect, with a mean crowding ratio (CR 1.8) in 3-4 year olds, decreasing with
[19]
age. Therefore, a crowding ratio ≥ 2 (i.e. halving of the visual acuity out-
come) was considered as the standard for abnormal crowding, irrespective of
age or developmental level.
The following ratios were calculated for each participant: crowding ratio
(uncrowded/crowded optotype acuity), grating acuity/uncrowded acuity, grat-
ing acuity/crowded acuity. Mean (SD) and median ratios were calculated for
the three participant groups. Between-test differences were analysed with
Friedman and Wilcoxon Signed Rank tests and between-group differences with
Kruskall-Wallis tests.
To answer the second study question, per specialist discipline, the propor-
tion of correctly identified children-at-risk within the target group (sensitivity)
and the proportion of correctly identified non-targets within the control group
(specificity) were calculated. We plotted ROC-curves for crowded visual acu-
ity and each ratio, to check which outcome and cut-off value could be used
best by each discipline. If visual acuity, measured during the study, appeared
different from the received information, the child was not excluded from the
analysis or transferred to a different group, because we wanted to act on clini-
cally defined risks. To be certain that the presence of abnormal crowding could
not be explained by the presence of strabismus – risk factor for amblyopia -,
we calculated the correlation (Spearman’s ρ) between a crowding ratio ≥ 2
and manifest or latent strabismus in the total study population.
Results
Participants
Sixty of the 107 children (25 boys, 35 girls) were included in group 1, 21 (11
boys, 10 girls) in group 2, and 26 (13 boys, 13 girls) in group 3. Ages of groups
2 and 3 did not differ significantly, whereas group 1 included significantly more
young children (H(2) = 9.17, p = .01). Ocular abnormalities in groups 2 and
3 are presented in Table 1.
Five children in group 1 had refractive errors and 12 latent strabismus. In

group 2, eight children had multiple ocular abnormalities other than refractive
errors or strabismus. Fifteen wore glasses and/or contact lenses, including
seven because of pseudophakia or aphakia. Nine had strabismus and eight 2
had other ocular abnormalities. In 3 children, contrast sensitivity was reduced
(> 5% contrast needed), whereas in one child the right visual field seemed to
be somewhat impaired (confrontation test). There was no significant correla-
tion between a crowding ratio ≥ 2 and latent or manifest strabismus.
The aetiology of brain damage in group 3 was: hypoxic ischemic encepha-
lopathy (11: 4 intraventricular haemorrhage, 6 periventricular leucomalacia,
1 both), asphyxia (4), brain malformation (2), leukoencephalopathy with optic
neuropathy (1), Beckwith Wiedemann syndrome (1), meningitis (2) and un-
known (3 cerebral palsy, 2 visual problems not explained by ocular abnormali-
ties). In 7/26 children, no ocular abnormalities were found. Six had refractive
errors and fourteen strabismus. In addition to a child with opticopathy, four
children had structural ocular abnormalities other than abnormalities due to
retrograde optic atrophy or refractive errors. Contrast sensitivity was reduced
in three and visual field defects seemed present in ten children.
Table 1. Ocular abnormalities present in subgroups of children with ocular abnormalities only and
with brain damage
Group 2 Group 3
Ocular abnormalities n = 21 n = 17
Microphthalmus unilateral - 1
Lens 6 1
Congenital cataract 6 1
Stickler Syndrome 2
Uvea and Fundus 8 2
Idiopathic chronic uveitis 1
Choreoretinitis 1
Aniridia 1
Iris and peripheral choridea coloboma 1
x-linked retinoschizis 1
Retinoblastoma 1
Cone dysfunction 1
Cone-rod dysfunction 1
Cone-rod dystrophy 1
Red-green deficiency 1
Macula 2 1
abnormality 1 1
Bull’s eye 1
Oculocutaneous albinism 1
26 Chapter 2
Table 1. Ocular abnormalities present in subgroups of children with ocular abnormalities only and
with brain damage (continued)
Group 2 Group 3
Ocular abnormalities n = 21 n = 17
Neuro-ophthalmological 9 8
Nystagmus 8 5
Congenital nystagmus 4
Manifest 3 4
Latent 1 1
Torticollis 5 4
Optic nerve hypoplasia 1
Septo-optic dysplasia 1*
Opticopathy 1
Retrograde optic nerve atrophy 2
Strabismus 11 14
Esotropia (only trace) 6 (0) 5
Exotropia (only trace) 3 (1) 5
Intermittent mixed (only trace) 1
Hypertropia (only trace) 1 (1) 1
Esophoria (only trace) 1 (1)
Exophoria (only trace)
Upshoot in adduction 2
Convergence limited 1
Refractive errors 8 6
Myopia
high 1
moderate 2
Hyperopia
high 4 1
moderate 1 2
mild 2
Anisohyperopia 1
* this child was missing the right eye and had a prechiasmatic abnormality of the optic nerve
Visual acuities
All visual acuity tasks were completed by 59/60 children in group 1, 15/21 in
group 2 and 19/26 in group 3. Grating acuity was missing in 9 participants
because of logistical reasons; optotype acuities were missing in five children
with brain damage, because of low vision, short attention or problems with
the test distance.
Results of the acuity tests are reported in Table 2. As was to be expected,
within all groups, visual acuity varied with the test used: Teller Acuity Cards
resulted in a significantly higher median visual acuity than the uncrowded
Cambridge cards (p < .01) and median uncrowded acuity was significantly
higher than median crowded acuity (p < .01). Remarkably, low vision (< 10
cycl/deg) was found in two children of group 1. One of them was known in the
Eye Hospital and wore glasses to correct mild hyperopia; we referred the other
for ophthalmological assessment. 2
Table 2. Visual acuity results of the Teller Acuity Cards (grating acuity) and the Cambridge Crowd-
ing Cards (uncrowded and crowded acuity) in typically developing children (group 1), children with
ocular disorders (group 2), children with brain damage (group 3).
Group
1 2 3
Test (n = 60) (n = 21) (n = 26)
Teller Acuity Cards
Grating acuity n 59 15 24
Median Snellen equivalent * 20/11.6 20/17.5 20/11.6
< 20/20 † (n) 2 7 5
≤ 20/30 ‡ (n) 0 0 1
Cambridge Crowding Cards
Uncrowded acuity n 60 21 24
Median Snellen equivalent 6/4 6/12 6/6
<6/6 † (n) 2 8 2
≤ 6/18 ‡ (n) 0 9 4
Crowded acuity n 60 21 21
Median Snellen equivalent 6/6 6/60 6/9
< 6/9 or 6/6 § (n) 16 4 7
≤ 6/18 ‡ (n) 2 15 6
* cycl/deg = (20/Snellen denominator) x 30. † Subnormal vision, low vision excluded. Snellen
equivalent for 30 cycl/deg is 6/6 or 20/20. ‡ Low vision ≤ 10 cycl/deg. Snellen equivalent for 10
cycl/deg is 6/18 or 20/30. § Age dependent cut-off value for subnormal vision: 20 cycl/deg (Snellen
equivalent 6/9) for age < 4y6m; < 30 cycl/deg (Snellen equivalent < 6/6) for older children. Low
vison excluded.
Groups performed significantly different on the visual acuity tests (Kruskal-

Wallis tests p < .01). To control for unequal variance we performed post-hoc
[20]
tests on ranks using the Welch t-test. Median grating acuity only differed
significantly between group 1 and 2 (tw(16)= 2.43, p = .03), whereas the
differences in median uncrowded and crowded acuities between groups were
all significant (p ≤ .02). To study a possible confounding effect by age, we cal-
culated the Spearman rank correlations between age and test outcomes. Only
in group 1, a significant relation was found between test outcomes and age
(range ρ = 0.44 - 0.52, p < .01), which may indicate a developmental trend.
Because group 1, which included younger children than the other groups,
performed equal to or better than group 2 and 3, confounding by age is not
likely.
28 Chapter 2
Acuity ratios
In Table 3, we present the observed crowding and grating/optotype ratios
within each group. Significant group differences were found for all ratios
(Kruskal-Wallis tests: p < .01).
Table 3. Observed ratios in typically developing children (group 1), children with ocular disorders
(group 2) and children with brain damage (group 3)
Group
1 2 3
Ratio (n = 60) (n = 21) (n = 26)
Uncrowded acuity/Crowded acuity
n 59 21 21
Median 1.50 1.50 2.00
Mean 1.60 1.42 1.94
SD 0.42 0.49 0.66
Ratio ≥ 2 (n) 14 3 14
Grating acuity/Uncrowded acuity
n 59 15 22
Median 1.14 2.51 1.60
Mean 1.30 2.55 2.00
SD 0.33 1.98 1.91
Ratio ≥ 2 (n) 3 10 6
Grating acuity/Crowded acuity
n 59 15 19
Median 1.71 3.35 2.57
Mean 2.04 3.75 3.96
SD 0.60 2.02 4.22
Ratio ≥ 2 (n) 23 12 13
Uncrowded acuity/crowded acuity
Post-hoc analysis on the ranks indicated that the ranks of the ratios in group
1 were significantly higher than in group 2, (tw(31.7) = 2.25, p = .03), al-
though median score for the uncrowded/crowded acuity ratio of group 1 and
2 was equal. This difference cannot be explained by the different step sizes in
the Cambrigde Crowding Cards and its ratios, but is the result of the higher
proportion of participants with a crowding ratio ≥ 2 in group 1.
The difference between group 1 and 3 was nearly significant (tw(28.9) = -
2.09, p = .05) and the ratio in group 3 was significantly higher than that in
group 2 (tw(40) = 3.20, p < .01). A crowding ratio of 1.5 (crowded acuity is
a half octave or one step lower than uncrowded acuity) or lower was found in
76% of group 1, 86% of group 2, and 33% of group 3.
Only in group 2, a significant relationship between age and the uncrowded/

crowded ratio was found (ρ = -0.53, p = .01). Ratios in older children were
lower than those in younger children, which might be indicative of non-parallel
trends in acuity development: a relatively stable uncrowded acuity (ρ = 0.06) 2
over the age range and a small but positive trend for crowded acuity (ρ =
0.27). Mean uncrowded/crowded ratios per age group can be found in Table
4. Correlations between age and other ratios were not significant (range ρ =
0.05 - 0.27).
Table 4. Mean ratio between uncrowded acuity and crowded acuity per age group in typically de-
veloping children (group 1), children with ocular disorders (group 2) and children with brain damage
(group 3)
Group
1 2 3
(n = 59) (n = 21) (n = 21)
3-4 years old
n 16 4 2
Mean 1.67 1.63 2.13
SD 0.51 0.25 0.18
5-7 years old
n 43 12 13
Mean 1.57 1.48 1.92
SD 0.39 0.56 0.44
Older than 7 years
n 5 6
Mean 1.10 1.93
SD 0.22 0.66
Grating acuity/uncrowded acuity

Group 2 had significantly higher ratio of grating acuity to uncrowded acuity,
compared to group 1 (tw(72) = 4.90, p < .01) and group 3 (tw(35) = 2.24, p
= .03), whereas ratios for group 1 and 3 did not differ significantly.
Grating acuity/crowding acuity

The ratios between grating acuity and crowded acuity of group 2 and 3 did
not differ significantly, but were significantly higher than in group 1 (tw(72) =
3.90, p < .01 and tw(24.8) = 2.23, p = .04).
30 Chapter 2
Referral criteria for different specialists

In order to decide which parameter could be used best by youth health care
physicians versus specialist groups, we used results presented in Table 5 and
ROC-curves, shown in Figure 1.
Youth health care physicians best continue to use subnormal crowded acuity,
i.e. routinely used acuity tests, to discriminate between children with and
without any visual abnormalities (area under the curve (AUC) = 0.80, 95%-
CI: 0.71-0.89, p < .01). Results of the crowding ratio (uncrowded acuity/
crowded acuity) are equal to a random guess (AUC = 0.50).
For ophthalmologists, the primary choice for the detection of children at
risk of cerebral visual impairments would be the crowding ratio (AUC = 0.68,
95%-CI = 0.54-0.83, p = .01). Subnormal vision and the ratio between grat-
ing acuity and crowded acuity resulted in comparable sensitivities, but lower
specificities.
Table 5. Sensitivity and specificity for cut-off values evaluated for different specialist groups
Crowded acuity Ratios
Subnormal Low vision unC/C TAC/unC TAC/C
Specialist < 6/6 or 6/9 * 6/18 † ≥2 ≥2 ≥2
Youth health care physician
Sensitivity .76 .50 .40 .43 .74
(95%-CI) (.61-.87) (.36-.64) (.27-.56) (.29-.59) (.57-.85)
Specificity .70 .97 .77 .95 .61
(95%-CI) (.58-.80) (.89-.99) (.65-.86) (.86-.98) (.36-.64)
Ophthalmologist
Sensitivity .62 .29 .67 .27 .68
(95%-CI) (.41-.79) (.14-.50) (.45-.83) (.13-.48) (.46-.85)
Specificity .54 .79 .79 .82 .53
(95%-CI) (.43-.64) (.69-.87) (.69-.86) (.72-.89) (.42-.64)
Orthoptist special services
Sensitivity .62 .29 .67 .27 .68
(95%-CI) (.41-.79) (.14-.50) (.45-.83) (.13-.48) (.46-.85)
Specificity .10 .29 .86 .33 .20
(95%-CI) (.03-.29) (.14-.50) (.65-.95) (.15-.58) (.07-.45)
C = crowded acuity, unC = uncrowded acuity, TAC = grating acuity; * subnormal vision is age
dependent, 6/6 = 30 cycl/deg for children age 4y6m and older and 6/9 = 20 cycl/deg for younger
children; † 6/18 = 10 cycl/deg.
Figure 1. ROC-curves for crowded visual acuity and acuity ratios from the perspective of different
groups of diagnosticians. A. Youth health care physicians (target group: children at risk of visual
problems due to ocular abnormalities or brain damage; control group: children without visual prob-
lems). B. Ophthalmologists (target group: children with cerebral visual problems; control group:
children without visual problems and children with ocular visual problems). C. Low vision centres
(target: group: children with cerebral visual problems; control group: children with ocular visual
problems)
C = crowded acuity; unC = uncrowded acuity; TAC = grating acuity
Each marker is a different cut-off point. Filled makers indicate cut-off points used in the current
study: low vision (grey diamond), subnormal acuity (black diamond) and ratios equal to 2.0 (other
black markers). Crowded acuity markers to the right of these filled markers are higher cut-off values
and markers to the left are lower cut-off values. Ratio markers to the right of the filled markers are
lower cut-off values, markers to the left are higher cut-off values.
32 Chapter 2
Within low vision services, too, the crowding ratio with a cut-off value of 2 is
currently the best option for screening for a risk of cerebral visual impairment
(AUC = 0.76, 95%-CI = 0.60-0.91, p < .01). Table 5 shows that subnormal
vision and a ratio of 2 or higher between grating acuity and crowded acuity
were relatively frequent in both patient groups 2 and 3 (high sensitivity, low
specificity). Low vision and a ratio of 2 or higher between grating acuity and
uncrowded acuity were relatively frequent in group 2, but relatively uncom-
mon in group 3 (low sensitivity and low specificity).
We conclude that the results so far indicate that in groups referred for
ophthalmological assessment or specialist evaluation in a low vision service,
the crowding ratio is sensitive to detect children at risk of cerebral visual
impairment, with and without low vision.
Discussion
This study explored an opportunity to improve clinical identification of children

at risk of cerebral visual impairment, based on visual acuity ratios, in order
to advance effective referrals for specialist neuropsychological assessment.
The outcomes indicate, that the crowding ratio (ratio between uncrowded
and crowded acuity) has a favourable sensitivity (67%) and specificity (79-
86%), independent of visual acuity, to distinguish children at risk of cerebral
visual impairment from children with ocular pathology only. Current referral
for specialist low vision assessment, based on visual acuity, is less sensitive
and specific for this aim: the ‘low vision’ criterion only detects 29% of the
at-risk children. A crowding ratio ≥ 2, easily checked by orthoptists, might be
a valuable referral criterion. However, for screening of unselected groups of
children, such as in youth health care, it does not provide useful information.
Youth health care professionals best continue applying subnormal crowded
acuity as an indication for referral.
The present study indicates that the Cambridge Crowding Cards are highly
applicable (80 -100%) and better accepted than the addition of a grating
acuity test to routine assessment. Further studies have to prove whether or
not ratios differ using different crowded acuity tests, because acuity tests vary
in letter distances and step sizes (Snellen vs. log-based acuity). A published
study in 3-year olds comparing results of different crowded acuity tests, indi-
[21]
cates that differences can be expected, although these differences might
have been age-dependent.
Even though crowding has been extensively studied during more than 80
[22]
years and is considered clinically relevant, little is known about develop-
mental changes and normal limits in children. We were surprised to find that
14 out of 59 typically developing children had crowding ratios ≥ 2. This might 2
be explained by the relatively large number of younger children in this group:
a significant relationship between crowding ratio and age in the group without
brain damage suggests a developmental phenomenon.
Because in an earlier study applying Cambridge Crowding Cards in children
[16]
without ocular and neurological abnormalities standard deviations have
not been reported, we cannot statistically compare findings. Mean crowding
ratios found in 3-4 year olds seem comparable (1.7 vs. 1.8), whereas the
mean crowding ratio in 5-7-year olds in the present study is slightly higher
than in the earlier report (1.6 vs. 1.2).
Although the use of crowding ratio ≥ 2 may help to detect children with
unconfirmed brain damage, especially those with subnormal vision that are
otherwise not referred for or entitled to specialised diagnostics and low vi-
sion care, we did not address the question whether or not the crowding ratio
directly predicts perception problems and can be used as selection criterion
in low vision services. Studying the developmental changes of the crowding
ratio in relation to results on tasks with a high attentional load or with stimuli
that can be considered ‘crowded’ (i.e. visual search tasks and object percep-
tion tasks under suboptimal conditions), might help to answer the question
whether or not the crowding ratio can be used as selection criterion and pre-
dict higher visual impairments. If target identification in ‘crowded’ stimuli is
[5]
modulated by attention and determined by the attentional resolution, task
performance might be dependent on age and task duration. Young children
with brain damage might have a higher crowding ratio as well as weak per-
ception performances, whereas older children might perform normally on the
short acuity task but worse on perception tests, with their longer duration i.e.
with a higher attentional load.
Acknowledgements
We thank the participating children and their parents, the primary schools
(Eduard van Beinum, Openbare basisschool Charlois, De Bergse Zonnebloem,
Visio-school), Rijndam Rehabilitation Centre, and Royal Dutch Visio, all located
in Rotterdam, the Netherlands, for their participation and help.
34 Chapter 2
References
2. Hard, A.L., E. Aring, and A. Hellstrom, Subnormal visual perception in school-
aged ex-preterm patients in a paediatric eye clinic. Eye, 2004. 18(6): p. 628-34.
3. Flom, M.C., F.W. Weymouth, and D. Kahneman, Visual Resolution and Contour
Interaction. J Opt Soc Am, 1963. 53: p. 1026-32.
4. Dutton, G.N. and L.K. Jacobson, Cerebral visual impairment in children. Semin
Neonatol, 2001. 6(6): p. 477-85.
5. He, S., P. Cavanagh, and J. Intriligator, Attentional resolution and the locus of
visual awareness. Nature, 1996. 383(6598): p. 334-7.
6. Pelli, D.G., M. Palomares, and N.J. Majaj, Crowding is unlike ordinary masking:
distinguishing feature integration from detection. J Vis, 2004. 4(12): p. 1136-69.
7. Pike, M.G., et al., Patterns of visual impairment associated with lesions of the
preterm infant brain. Dev Med Child Neurol, 1994. 36(10): p. 849-62.
8. Jacobson, L., et al., Visual impairment in preterm children with periventricular
leukomalacia--visual, cognitive and neuropaediatric characteristics related to
cerebral imaging. Dev Med Child Neurol, 1996. 38(8): p. 724-35.
9. Mercuri E, A.S., Guzzetta A, Barnett A, Haataja L, Rutherford M, Cowan F, Dubowitz
L, Braddick O, Atkinson J., Visual function at school age in children with neonatal
encephalopathy and low Apgar scores. Arch Dis Child Fetal Neonatal Ed, 2004.
89: p. 258-262.
10. Bonneh, Y.S., D. Sagi, and U. Polat, Local and non-local deficits in amblyopia:
acuity and spatial interactions. Vision Res, 2004. 44(27): p. 3099-110.
11. Rydberg, A., et al., Assessment of visual acuity in children aged 1 1/2-6 years,
with normal and subnormal vision. Strabismus, 1999. 7(1): p. 1-24.
12. Giaschi, D.E., et al., Defective processing of motion-defined form in the fellow eye
of patients with unilateral amblyopia. Invest Ophthalmol Vis Sci, 1992. 33(8): p.
2483-9.
13. Mendola, J.D., et al., Voxel-based analysis of MRI detects abnormal visual cortex
in children and adults with amblyopia. Hum Brain Mapp, 2005. 25(2): p. 222-36.
14. Flodmark, O., J.E. Janc, and P.K.H. Wongc, Computed Tomography Of The Brains
Of Children With Cortical Visual Impairment. Developmental Medicine & Child
Neurology, 1990. 32(7): p. 611-620.
15. Stiers, P., R. Vanderkelen, and E. Vandenbussche, Optotype and grating visual
acuity in patients with ocular and cerebral visual impairment. Investigative Oph-
thalmology & Visual Science, 2004. 45(12): p. 4333-9.
16. Atkinson, J., et al., Visual acuity testing of young children with the Cambridge
Crowding Cards at 3 and 6 m. Acta Ophthalmol (Copenh), 1988. 66(5): p. 505-8.
17. Teller, D.Y., The forced-choice preferential looking procedure: A psychophysical
technique for use with human infants. Infant Behavior and Development, 1979.
2(0): p. 135-153.
18. International Classification of Diseases 10th edition online version. 12 december];
Available from: http://apps.who.int/classifications/icd10/browse/2016/en#/H54.
19. Larsson, E.K., A.C. Rydberg, and G.E. Holmstrom, A population-based study on
the visual outcome in 10-year-old preterm and full-term children. Arch Ophthal-
mol, 2005. 123(6): p. 825-32.
20. Zimmerman, D.W. and B.D. Zumbo, Rank Transformations and the Power of the
Student t Test and Welch t’ Test for Non-Normal Populations With Unequal Vari- 2
ances. Canadian Journal of Experimental Psychology, 1993. 47(3): p. 523-539.
21. Ciner, E., et al., Threshold visual acuity testing of preschool children using the
crowded HOTV and Lea Symbols acuity tests. Journal of American Association for
Pediatric Ophthalmology and Strabismus, 2003. 7(6): p. 396-399.
22. Levi, D.M., Crowding--an essential bottleneck for object recognition: a mini-
review. Vision Res, 2008. 48(5): p. 635-54.
Chapter
Chapter 3
3
Visual motion perception: brain

localisation, terms and paradigms
Introduction motion perception 39
Motion perception is the specific mental function of recognizing and interpret-

ing dynamic visual information. Its representation in the brain is complicated.
Motion perception can be selectively impaired as a result of damage to the
[1]
lateral occipital-temporal cortex, although it seems never completely absent.
Lesions acquired in adulthood illustrate the importance of motion perception
[2]
in daily activities. Scientific studies indicate, that motion perception deficits
also emerge in various developmental disorders, such as Williams syndrome 3
[3]
(WS), autism and developmental dyslexia.
Different aspects of motion perception can be distinguished, e.g. the percep-
tion of coherent or global motion, motion-defined form, and biological motion.
In this chapter, we describe the current state of knowledge on brain areas,
involved in motion perception, followed by an introduction and explanation of
terms and motion perception paradigms, which are frequently used in funda-
mental research on visual motion perception.
Hierarchical brain model of motion perception

Through cognitive neuroscience research, several visual functions have been
differentiated and assigned to different brain areas or neural systems.
Visual functions of the brain can be described in a hierarchical model with
two anatomically and functionally distinguishable main streams: the ventral
stream and the dorsal stream. From the retina, most signals pass through the
lateral geniculate nucleus (LGN) into the primary visual cortex (occipital or
striate cortex (V1)) and the extra-striate cortex (V2). The ventral stream ends
in the inferior temporal cortex, and is involved in object and face recognition.
The dorsal stream leads into the posterior parietal cortex, including extrastri-
ate visuocortical areas V3a, V5+ (which encompasses V5, medial superior
temporal area (MST) or associated motion processing regions in humans
[4]
(hMT+)), and extends towards the anterior parietal and posterior frontal
[5]
cortex, where the integration of visual and motor functions is accomplished.
The dorsal stream is activated during motion perception, object location
and visual spatial attention. hV5+ is located in the ascending limb of the
inferior temporal sulcus (ITS) at the temporal-parietal occipital junction, and
[6-9]
is highly sensitive to and activated by motion stimuli, for example those
presented by random dot kinematograms (RDK) (see below). Simple global
[10, 11]
motion perception (linear displacement) is related to the activity in V5.
Stimuli with relatively slow speed, e.g. < 6 deg/s, activate the V1-V5 route.
When stimuli with relatively high speeds, e.g. > 15 deg/s, are presented,
V5 is activated before V1. This suggests, that V5 also gets input through a
[12, 13]
colliculo-cortical pathway. Cells within hV5+ are motion direction sensi-
40 Chapter 3
[14, 15]
tive and stimulus characteristics, such as luminance contrast, influence
[6]
hV5+ activity.
Perception of complex motions, e.g. rotation, expansion contraction and
[16-20]
combination of these motions, is related to activity in MST. Biological
[21-23]
motion additionally activates the superior temporal sulcus (STS) region,
a ventral stream area.
Similarly, the perception of motion-defined forms does not only activate
dorsal stream motion areas, but also gives activity in ventral stream form
[24-26]
processing areas.
It is therefore clear, that visual motion perception is subserved by many
different cortical areas, and that various motion paradigms will differ in the
number and nature of the visual motion areas that they rely on for adequate
performance. Generally, it is presumed that motion perception tasks are valid
to study higher-level dorsal stream functioning and to identify motion percep-
tion problems in patients.
Different motion perception paradigms

A common method to measure motion perception makes use of the direc-
[3]
tion of motion, a basic motion perception function, in so-called random dot
kinematograms (RDK). These consist mostly of a black field with moving white
dots, presented on a screen. One portion of dots moves in a uniform direction,
[27]
while the remaining dots move in random order. Coherent motion can be
perceived as a result of perceptual grouping of multiple local motion signals,
which is an unconscious process. An observer will be able to see coherent
motion, also called global motion, as long as the proportion of dots moving in
the uniform direction exceeds the observer’s individual perception threshold.
If global motion is perceived, the observer is able to see and indicate in which
direction the global motion moves. The individual threshold can be assessed
quantitatively, for example by measuring the minimum proportion of coherent
moving dots, needed to identify the motion direction correctly (Figure 1A).
The perception of motion direction can also be affected by the displace-
ment of dots between consecutive frames within the RDK (distance in degrees
between old and new position of a dot). If the displacement exceeds the
individual threshold, no (global) motion will be perceived. This threshold is
known as maximum displacement or Dmax and is the displacement level at
[28]
which global motion and motion direction still can be perceived.
Figure 1. Three types of motion perception. A. Global motion. An RDK with upwards global motion
with a coherence level of 20%. Dots with an arrow move upwards, dots without an arrow move in
random direction. B. Biological motion. A point-light human walker, walking to the right. C. Motion-
defined form. Dots within the square move upwards. Dots outside the square move downwards.
Boundary of the square is only illustrative and not visible in the test condition.
Another, more complex, perceptual function is the ability to recognize forms

or objects only through motion. An object can be recognised by making only
the contour of the object visible by the use of global motion in an RDK or
through the presentation of object-specific motion, for example a human can
be recognised by a presentation of its walking motion only (Figure 1B). Terms
like form-from-motion or motion-defined form are used to describe such mo-
tion perception stimuli. The motion of motion-defined forms can be subdivided
in two distinct classes: biological motion and non-biological motion. Animate
objects, like humans and animals, produce biological motion, whereas non-bi-
ological motion is the motion of non-animate objects or fictive forms. Not only
can biological motion be distinguished from non-biological motion, animate
objects and their activities can even be recognised and distinguished through
their specific motion patterns only.
Motion-defined form, its diversity and its use are illustrated by the following
[29]
studies. Johnson and Mason used non-animate forms with a random dot
texture placed on a random dot texture background to study motion-defined
form perception in infants. Only when the forms moved, they became visible.
After habituation to a motion-defined form, luminance-defined forms were
presented, of which one form was new and the other was presented earlier
as the motion-defined form. Longer looking times for a newly presented form
than for an earlier presented form, also called novelty preference, indicated
that form perception had taken place.
Global motion in an RDK, just as in the perception of a motion direction
task, can also be used to study motion-defined form (Figure 1C). In the study
[30]
of Regan et al dots in the contour of a letter were moving opposite to the
42 Chapter 3
background dots. The contour of the letter became clearer when the dots
moved faster. The proportion correctly named letters was used to determine
the motion sensitivity threshold in terms of motion speed and in terms of
visual acuity based on motion-defined letters.
Tasks for biological motion often consist of point-light figures (see Figure 2
right) made by placing white dots or small lights at the joints of the otherwise
[31]
invisible human or animal. In the study by Grossman et al, for instance,
observers had to discriminate a human walker from a scrambled walker and
indicate the position of the human walker. In order to make the task more
difficult, noise (extra dots) was added. A proportion correctly indicated human
walkers was used to determine the amount of additional dots the observer
could deal with.
A range of computerised motion perception tasks have been developed around
the world, but they are primarily applied in research settings. The advantage
of such tasks is, that they have a quantitative outcome measure: a perception
threshold. The use of motion perception tasks and thresholds is a way to
[31, 32]
objectify the presence and severity of a motion perception impairment,
which would be a relevant addition to clinical low vision diagnostics. Therefore,
we have dedicated a major part of this thesis to motion perception, performing
a systematic review of the literature to identify available tasks and judge their
qualities for use in clinical practice, followed by experimental research.
References
1. Baker, C.L., Jr., R.F. Hess, and J. Zihl, Residual motion perception in a “motion-
blind” patient, assessed with limited-lifetime random dot stimuli. J Neurosci,
1991. 11(2): p. 454-61.
2. Zihl, J., D. von Cramon, and N. Mai, Selective disturbance of movement vision
after bilateral brain damage. Brain, 1983. 106 (Pt 2): p. 313-40.
3. Braddick, O., J. Atkinson, and J. Wattam-Bell, Normal and anomalous develop-
3
ment of visual motion processing: Motion coherence and ‘dorsal-stream vulner-
ability’. Neuropsychologia, 2003. 41(13): p. 1769-1784.
4. Merigan, W.H. and J.H. Maunsell, How parallel are the primate visual pathways?
Annu Rev Neurosci, 1993. 16: p. 369-402.
5. Goodale, M.A. and A.D. Milner, Separate visual pathways for perception and ac-
tion. Trends Neurosci, 1992. 15(1): p. 20-5.
6. Tootell, R.B., et al., Functional analysis of human MT and related visual cortical
areas using magnetic resonance imaging. J Neurosci, 1995. 15(4): p. 3215-30.
7. Rees, G., K. Friston, and C. Koch, A direct quantitative relationship between the
functional properties of human and macaque V5. Nat Neurosci, 2000. 3(7): p.
716-23.
8. Watson, J.D., et al., Area V5 of the human brain: evidence from a combined study
using positron emission tomography and magnetic resonance imaging. Cereb
Cortex, 1993. 3(2): p. 79-94.
10. Alexander, I. and A. Cowey, The cortical basis of global motion detection in blind-
sight. Exp Brain Res, 2009. 192(3): p. 407-11.
2001. 30(1): p. 61-72.
12. Rodman, H.R., C.G. Gross, and T.D. Albright, Afferent basis of visual response
properties in area MT of the macaque. II. Effects of superior colliculus removal. J
Neurosci, 1990. 10(4): p. 1154-64.
13. Ffytche, D.H., C.N. Guy, and S. Zeki, The parallel visual motion inputs into areas
V1 and V5 of human cerebral cortex. Brain, 1995. 118 ( Pt 6): p. 1375-94.
14. Huk, A.C., D. Ress, and D.J. Heeger, Neuronal basis of the motion aftereffect
reconsidered. Neuron, 2001. 32(1): p. 161-72.
15. Tootell, R.B., et al., Visual motion aftereffect in human cortical area MT revealed
by functional magnetic resonance imaging. Nature, 1995. 375(6527): p. 139-41.
16. Duffy, C.J. and R.H. Wurtz, Sensitivity of MST neurons to optic flow stimuli. II.
Mechanisms of response selectivity revealed by small-field stimuli. J Neurophysiol,
1991. 65(6): p. 1346-59.
17. Duffy, C.J. and R.H. Wurtz, Sensitivity of MST neurons to optic flow stimuli. I.
A continuum of response selectivity to large-field stimuli. J Neurophysiol, 1991.
65(6): p. 1329-45.
18. Smith, A.T., et al., Sensitivity to optic flow in human cortical areas MT and MST.
Eur J Neurosci, 2006. 23(2): p. 561-9.
44 Chapter 3
19. Wall, M.B., et al., Selective visual responses to expansion and rotation in the hu-
man MT complex revealed by functional magnetic resonance imaging adaptation.
Eur J Neurosci, 2008. 27(10): p. 2747-57.
20. Dukelow, S.P., et al., Distinguishing subregions of the human MT+ complex using
visual fields and pursuit eye movements. J Neurophysiol, 2001. 86(4): p. 1991-
2000.
21. Bonda, E., et al., Specific involvement of human parietal systems and the amyg-
dala in the perception of biological motion. J Neurosci, 1996. 16(11): p. 3737-44.
22. Beauchamp, M.S., et al., Parallel visual motion processing streams for manipu-
lable objects and human movements. Neuron, 2002. 34(1): p. 149-59.
23. Grossman, E., et al., Brain areas involved in perception of biological motion. J
Cogn Neurosci, 2000. 12(5): p. 711-20.
24. Marcar, V.L., et al., An fMRI study of the cerebral macro network involved in
‘cue invariant’ form perception and how it is influenced by stimulus complexity.
Neuroimage, 2004. 23(3): p. 947-55.
25. Wang, J., et al., Relationship between ventral stream for object vision and dorsal
stream for spatial vision: an fMRI + ERP study. Hum Brain Mapp, 1999. 8(4): p.
170-81.
26. Klaver, P., et al., Dorsal stream development in motion and structure-from-motion
perception. Neuroimage, 2008. 39(4): p. 1815-23.
27. Williams, D.W. and R. Sekuler, Coherent global motion percepts from stochastic
local motions. Vision Res, 1984. 24(1): p. 55-62.
28. Baker, C.L., Jr. and O.J. Braddick, Eccentricity-dependent scaling of the limits for
short-range apparent motion perception. Vision Res, 1985. 25(6): p. 803-12.
29. Johnson, S.P. and U. Mason, Perception of kinetic illusory contours by two-month-
old infants. Child Dev, 2002. 73(1): p. 22-34.
30. Regan, D., A.C. Kothe, and J.A. Sharpe, Recognition of motion-defined shapes
in patients with multiple sclerosis and optic neuritis. Brain, 1991. 114 ( Pt 3): p.
1129-55.
31. Grossman, E.D., L. Battelli, and A. Pascual-Leone, Repetitive TMS over posterior
STS disrupts perception of biological motion. Vision Res, 2005. 45(22): p. 2847-
53.
32. Grossman, E.D. and R. Blake, Brain Areas Active during Visual Perception of
Biological Motion. Neuron, 2002. 35(6): p. 1167-75.
Chapter
Chapter 4
4
Applicability of computerised
motion perception tasks in clinical
practice for children. A systematic
review
46 Chapter 4
Abstract
Problem
A variety of computerised tasks is being used in research to assess motion
perception in children. Which can be applied in clinical diagnostics?
Methods
Published studies in children aged 2-12 years were systematically reviewed to
judge 1. types of tasks, technical specifications and quantitative outcomes; 2.
reliability of normal limits and developmental trends; 3. differences between
typically developing children and risk groups (mean performance and percent-
age abnormal performers).
Key findings
Global motion has been studied most extensively, whereas a smaller number
of studies focussed on motion-defined form and biological motion. There was a
wide variety in task characteristics, and confidence intervals for normal limits
were wide as a result of small age groups. Developmental trends were observed
for global motion and motion-defined form. Several risk groups performed
significantly worse than typically developing children, but taking confidence
intervals into account, there is only some evidence for an increased risk of
abnormal performance in children with autism, amblyopia, and prematurely
born children.
Interpretation
Computerised motion perception tasks can be used in clinical practice as ob-
servational instruments with quantifiable outcomes, but because of unreliable
normal limits, abnormal performers cannot yet be identified reliably. Interna-
tional collaboration might be needed to study larger patient groups in order to
estimate the prevalence of motion perception problems.
Systematic review motion perception 47
Introduction
Motion perception is the specific mental function of recognizing and interpret-

ing dynamic visual information. Motion perception can be selectively impaired
following damage to the lateral occipital-temporal cortex, although it seems
[1]
never completely absent. Acquired lesions in adults illustrate the importance
[2]
of motion perception in daily activities.
Various aspects of motion perception can be distinguished, e.g. the percep-
tion of coherent or global motion, biological motion, and motion-defined form.
4
Motion perception deficits are associated with various developmental dis-
orders, e.g. Williams syndrome (WS), autism, hemiplegia and developmental
[3]
dyslexia. However, the computerised motion perception tasks used in these
studies are still uncommon in clinical practice. In Dutch low vision centres,
motion perception is assessed occasionally by means of observational meth-
ods, e.g. observing the reaction of a child to an approaching rolling ball. The
advantage of motion perception tasks is that they have a quantitative outcome
measure, i.e. a perception threshold, that allows to diagnose the presence and
[4, 5]
severity of a motion perception impairment. Early detection of motion
perception impairment is preferred to allow early intervention. Therefore,
making such tasks available to the clinic is of great advantage.
A wide range of computerized motion perception tasks used in scientific
studies address different aspects of motion perception in different ways. As a
result, it remains unclear which of the currently available tasks can be intro-
duced in low vision centres for clinical diagnostic purposes.
Therefore, we performed a systematic review of the literature in order to:
(1) provide an overview of the types and characteristics of motion perception
tasks used in children, with their main technical specifications and quantitative
outcome measures (thresholds); (2) evaluate current knowledge on normal
limits, their precision, and developmental trends; (3) give an overview of vali-
dation studies, which show differences between patient groups and typically
developing children, i.e., the evidence for motion perception impairment of
different kinds in paediatric populations. Based on the results, we will discuss
which of the available tasks have the best characteristics and could be intro-
duced safely as a diagnostic test in low vision centres. In this review we focus
on both preschool and school-aged children (age between 2 and 13 years),
because most motion perception tasks require understanding of task instruc-
tions and at least some cooperation of the child.
48 Chapter 4
Methods
Literature search
We searched for English-language articles and reviews on motion perception,
published until March, 2009, using several public databases. In ISI Web of
Knowledge we searched in all citation databases, setting the time span at
all years. In PubMed we only used a search string without additional limits.
For the search in Embase, the databases EMbase and Unique MEDLINE were
selected. A fourth and last search was done in PsycInfo. Search strings and
numbers of hits are shown in Table 1.
Table 1. Search strings used in each database and number of hits.

Database Search string Hits
ISI Web of TS = (motion perception) AND TS = (“motion coherence” OR “global motion” OR 59
Knowledge “motion direction” OR “biological motion” OR “form from motion” OR “motion
defined form”) AND TS = (children) NOT TS=(infants) AND Language=(English)
AND Document Type=(Article OR Review)
Databases=SCI-EXPANDED, SSCI, A&HCI Timespan=All Years
PubMed ((“Child, Preschool”[Mesh] OR “Child”[Mesh])) AND (“motion perception”[MESH] 42
AND (“biological motion”[Title/Abstract] OR “motion coherence”[Title/Abstract]
OR “global motion”[Title/Abstract] OR “motion direction”[Title/Abstract] OR “form
from motion”[Title/Abstract] OR “motion defined form” [Title/Abstract])) AND
((“Child, Preschool”[Mesh] OR “Child”[Mesh])) AND (“motion perception”[MESH]
AND (“biological motion”[Title/Abstract] OR “motion coherence”[Title/Abstract] OR
“coherent motion”[Title/Abstract] OR “global motion”[Title/Abstract] OR “motion
direction”[Title/Abstract] OR “form from motion”[Title/Abstract] OR “motion defined
form” [Title/Abstract]))
Embase (‘movement perception’/exp/mj OR ‘movement perception’) AND ([article]/lim OR 16
[review]/lim) AND ([preschool]/lim OR [school]/lim) AND (‘biological motion’ OR
‘motion coherence’ OR ‘global motion’ OR ‘motion direction’ OR ‘form from motion’
OR ‘motion defined form’)
PsycInfo (Motion perception and (“biological motion” or “motion coherence” or “global 30
motion” or “motion direction” or “form from motion” or “motion defined form”)).mp.
limit to (all journals and (160 preschool age <age 2 to 5 yrs> or 180 school age
<age 6 to 12 yrs>))
Eligibility of studies
Based on the abstracts, or if needed the complete article, two authors inde-
pendently judged the articles using the following inclusion criteria: original
article published in English; subgroups of participants containing at least 5
participants; children aged 2 to 12 years included in the study; typically de-
veloping children and/or patients studied; if patients were studied, at least a
typically developing control group had to be included too; motion perception,
specified as either motion coherence/global motion, motion defined form, or
biological motion was tested; biological motion tasks used stimuli express-
ing global motion of the whole human body; quantitative outcomes, other
than quality judgements of the imitation of presented motions, EEG and fMRI
results, were reported for the motion perception tasks; if typically developing
children older than 12 years of age were also studied, quantitative data for
a subgroup of children aged up to 12 years had to be extractable from the
published results. If a group of patients including children with a chronological
age above 12 years was matched by developmental age to a group of typically
developing children, results of patients with a developmental age up to 12
years had to be extractable from the article. If one of the above criteria was
not met, the article was excluded. We discussed all discrepancies in judge-
ment, until consensus was reached on the eligibility of the study.
4
Data extraction and quality assessment
The following data were extracted: date of publication, authors, recruitment
location for participants, and participant selection methods; type and number
of participants; type of task and task characteristics; type of quantitative re-
sults and formulae used; methods of statistical analysis; group results, group
differences and significance of differences.
Quality assessment, based on the Quality Assessment Tool for Quantitative
[6]
Studies, was done by two authors independently. Difficulties in data extrac-
tion and quality assessment were discussed, until consensus was reached.
For the evaluation of possible confounders, comparability of groups was as-
sessed. In comparative studies (patients vs. controls), we focussed on age,
visual acuity and social economical background (SES) or intelligence (IQ). In
developmental studies (only typically developing children of different ages)
we focussed on visual acuity and social economical background (SES) or intel-
ligence (IQ). We considered SES and IQ as one variable, because SES and IQ
are significantly related and SES seems to influence cognitive development
[7-9]
in children. In comparative studies, the confounding risk was rated ‘low’
if groups were comparable on all three variables or analysis had shown that
confounding was not likely, ‘moderate’ if confounding was not likely for two,
and ‘high’ if confounding was not likely for one variable. In developmental
studies, the confounding risk was rated ‘low’ if studies were longitudinal and
both variables were reported and controlled for; ‘moderate’ if studies were
cross-sectional and both variables were reported and controlled for; and ‘high’
if one variable was reported and controlled for.
Bias was assessed by judging the recruitment setting, selection procedure
and the level of participation (% agreement of invited persons). Because it
appeared, that such information was absent or limited in all articles, whereas
participant groups of all studies were too small to obtain reliable normal limits,
we ignored bias risk in the analyses and do not present the rating here.
50 Chapter 4
For the description of task characteristic and thresholds, all studies were
included. We examined the following task characteristics: response by partici-
pant, stimulus type, motion direction per trial, dot speed and dot size, as well
as task distance, monocular or binocular testing. For the evaluation of normal
limits and differences between risk groups and typically developing children,
we included studies with a low or moderate confounding risk.
Analysis
To study normal limits for the general population, we focussed on typically
developing children. Normal limits or cut-off values are crucial for diagnostic
use: a task performance outside the normal limits is labelled as abnormal
or deviant. Sample limits were used to estimate normal limits. To estimate
the precision of reported sample limits, we calculated the 95% confidence
intervals (95%-CI) for the intended percentage excluded participants with
[10, 11]
the exact binomial method of Clopper-Pearson. This confidence interval
indicates how many of the general population may be considered abnormal if
the reported normal limits are used.
To study and illustrate developmental trends for mean performance and
sample limits, we combined the data of typically developing children of differ-
ent studies in a graph, provided that outcome measures were identical, and
stimuli, responses and procedures were comparable, and drew a regression
th
line. To compare sample limits of different studies, we estimated p90 (90
percentile) for each sample, defined as 1.28 SD above the mean. We as-
sumed that data were distributed normally and that the reported mean and
SD were the true population values. The relation between age and outcome
was studied with the Spearman rho correlation and we drew illustrative trend
lines. Differences in performance on motion perception tasks between various
patient groups and controls were studied in two ways.
First, we evaluated group differences in mean performance. If summary
data, but no statistics were reported, group means were compared by means
of a t-test, provided that groups were comparable.
Second, we evaluated differences in the percentages of abnormal performers.
Only if the patient groups show a significantly higher frequency of abnormal
performers, it is likely that patients have a higher risk of motion perception
problems. To evaluate whether the percentage abnormal performers, observed
in patients, was significantly higher than that observed in controls, a binomial
proportion test for independent samples or Fisher’s exact test was used. These
tests were also used to evaluate whether more patients than controls would
be labelled as abnormal, under the assumption that the upper limit of the
95%-CI of the intended percentage of excluded controls is the true population

frequency of abnormally performing controls if the estimated normal limits
are used. If this difference is significant, the patient group can be considered
at risk if the normal limits are adjusted to reduce the number of abnormally
performing controls in the general population.
To indicate the possible range of abnormally performing patients in the pa-
tient population given the sample limit estimated in the control group, we also
calculated the 95%-CI of the percentage abnormally performing patients, as
had been done for the controls.
4
Differences were considered significant at p ≤ .05.
Results
Study selection
Totally, 82 unique articles were found, of which 56 were excluded, some for
multiple reasons (Table 2). The remaining 26 articles were included in the
review, after discussion of six cases, primarily on the type of motion percep-
tion task and the extractability of data. Two tasks using global motion-like
stimuli were not included, because they were based on motion discrimination,
[12]
motion speed discrimination and motion direction discrimination, and not
[13]
on motion detection. In one article, no data were extractable for one of
two experiments on biological motion, therefore only one experiment was
[12-16]
included. For five studies also including children older than 12 years,
data of patients with a chronological age over 12 years but a younger mental
age were extracted, whereas for their control groups, only data of children
with a chronological age under 12 years were used.
Table 2. Main reason for exclusion of the articles not included in this review
Reason for exclusion # articles
Only abstract published 1
Qualitative review 7
Number of participants < 5 2
Age of participants outside range 2-12 years 18
Other task characteristics than specified for review 7
No quantitative outcome measures (fMRI, EEG, qualitative assessment motion imitation) 6
No quantitative data extractable for children in age range 2-13 years 14
Duplicate data 1
Total 56
52 Chapter 4
Study characteristics and quality assessment

Table 3 presents the 26 included studies, the study design and the evaluation
of confounding risk. All articles were published after 1994 and 22/26 focused
on a single motion perception aspect. All studies, including six developmental
studies, were cross-sectional. In only 6 articles, information on age as well as
visual acuity and SES or IQ was given and sufficiently controlled for, resulting
in a low confounding risk. Confounding risk was judged as moderate in 10 and
high in another 10 studies.
Table 3. Rating of confounding risk

Task type Year Study Design Confounding risk rating
[17]
GM 1998 Raymond & Sorensen Comp moderate
[18]
2000 Spencer et al Comp high
[19]
2002 Gunn et al Comp low
[20]
O’Brien et al Comp low
[21]
Ellemberg et al Comp high
[22]
2003 Atkinson et al Comp high
[23]
2004 Ellemberg et al Devc high
[24]
2005 MacKay et al Comp low
[25]
Chow & Ho Comp moderate
[12]
Mendes et al Comp moderate
[26]
2006 Milne et al Comp moderate
[15]
2006 Del Viva et al Comp moderate
[27]
2008 Pellicano & Gibson Comp low
[28]
GM & MDF 2005 Ho et al Comp moderate
[29]
Parrish et al Devc high
[30]
2007 Wang et al Comp moderate
[31]
MDF 1992 Giaschi et al Comp moderate
[32]
1999 Schrauf et al Devc high
[33]
2006 Jakobson et al Comp low
[13]
BM 1995 Moore et al Comp moderate
[34]
2001 Pavlova et al Devc high
[16]
2002 Jordan et al Comp moderate
[35]
2003 Blake et al Comp high
[36]
2006 Freire et al Devc high
[37]
2008 Lichtensteiger Devc high
[14]
GM, MDF & BM 2005 Reiss et al Comp low
GM = global motion or motion coherence, MDF = motion-defined form; BM = biological motion (in-
cluding global human motion. Comp = comparative study, comparing patients with controls, Devc
= cross-sectional developmental study, comparing different age groups of typically developing chil-
dren.
Motion perception tasks and thresholds

In the 26 studies, a wide variety of tasks were applied (tables with detailed
specification of threshold and response types, stimulus characteristics and test
distance of all studies are presented in the Appendix 1). Global motion was
studied most extensively, with motion coherence as the primary threshold.
The most common threshold for motion-defined form was minimum dot speed
and for biological motion an accuracy indicator. Most stimuli consisted of white
dots on a black background, slowly moving dots and a dot size greater than 6
arcmin, which should be visible for people with a visual acuity of 0.17 or higher
4
at test distance. Test distance varied between 40 centimetres and 6 metres.
[17, 23, 26, 32]
In 4 studies it was reported that testing was done binocularly and
[21, 24, 28-31]
in 6 studies monocularly, whereas this remained undefined in the
other 16 studies. Apart from these aspects, tasks also differed in number of
dots per random-dot kinematogram (RDK) or target, dot density range and
size of targets. Additionally, studies differed by whether or not they controlled
for undesirable effects, such as the use of local motion cues or tracking of a
single dot. To prevent this effect, the lifetime of signal dots was limited in ten
[14, 15, 17, 19, 21-24, 26, 27]
studies. The range of signal dot lifetimes was 13.3
[19, 21, 22, 24, 26, 27] [21]
ms-120 ms. One of 17 global motion studies controlled
for possible bias as a result of horizontal nystagmus, which is associated with
[38]
a reduced sensitivity for horizontal motion, by presenting vertical motion.
Taking all task aspects into account, only two global motion studies with
[28, 30] [18, 20]
motion direction and two studies with target location as response
used exactly the same task and task procedure. Additionally, three studies
on motion-defined form used identical stimuli and thresholds and comparable
[28-30]
task procedures. Biological motion studies were not comparable.
Estimated normal limits

In 8 of 16 studies with a low and moderate confounding risk, some kind of
[17, 24, 26-28, 30, 31, 33]
sample limit or cut-off value was used (Table 4a-b).
Sample limits were reported for global motion and motion-defined form tasks,
but not for biological motion tasks.
In the second column of Table 4a-b, we summarize study characteristics
(number of controls, mean age or age range of participants), formulas used
to estimate sample limits and the cut-off values, the percentage of the normal
population that authors were willing to label as abnormal and its 95%-CI as
calculated by us. In 7 of 8 studies, a z-score or SD was used to estimate the
sample limit, whereas the percentage intended to exclude in the general popu-
lation ranged from 0.1 to 5% (M + 3 SD to M + 1.64 SD). This was often done
54
Table 4a. Estimated normal limits in global motion studies, the intended percentages of persons in the general population that are excluded by use of the
estimated limit and labelled as abnormal performers, precision of sample limits (95%-CI), and percentages of abnormal performing patients according to the
sample limits.
Chapter 4
Controls Patients Statistical test

% intended % abnormal, p1-p2 pUL-p2
Formula for SL to exclude score outside
a
Study n Age SL-values (number) 95%- CI group n SL (number) 95%-CI p-value
Coherence level
Raymond & Sorensen 10 9.9 M + 2.82 SD = 27.8% 0.2% (n = 0) 0-31 dyslexia 10 60% (n = 6) 26-88 < .01 ns
[17]
(1998)
MacKay et al 19 9.8 deficit ratio > 2 - - VLBW 19 47% (n = 9) 24-71 - -
[24]
(2005)
Pellicano & Gibson 61 6.7 z > 1.65 5% (n = 3) 1-14 autism 20 40% (n = 8) 19-63 < .01 .01
[27]
(2008)
dyslexia 41 36% (n = 7) 15-59 .05 ns
Ho et al 25 4.6 M + 1.64 SD = 71% 5% (n = 1 per 25) 0-20 amblyopia FE 21 14% (n = 3) 3-36 ns ns
[28]
(2005)
25 7.5 M + 1.64 SD = 48%
25 10.6 M + 1.64 SD = 35%
Wang et al 32 7.3 M + 1.64 SD = 44% 5% (n = 2) 1-21 amblyopia FE 6 17% (n = 1) 0-64 ns ns
[30]
(2007)
Milne et al 23 10.3 z >1.65 5% (n = 1) 0-22 autism spectrum 23 22% (n = 5) 7-44 ns ns
[26]
(2006) disorder
Dmax
b
Ho et al 25 4.6 M ± 1.97 SD 2.5% (n = 1 per 25) 0-20 amblyopia FE 21 Dmax_e : 19% 5-42 .04 ns
[28]
(2005) = 0.35 & 1.45 deg (n = 4)
25 7.5 M ± 1.97 SD Dmax_d 10% 1-30 ns ns
= 0.50 & 1.56 deg (n = 2)
25 10.6 M ± 1.97 SD
= 0.57 & 1.43 deg
SL = sample limit, FE = fellow eye/non-amblyopic eye, Dmax_e = Dmax is elevated, Dmax_d = Dmax is depressed, p1-p2 is the difference between observed
proportions between patients and controls. The intended percentage excluded was used as observed proportion for controls, pUL-p2 is the difference between
a b
the observed proportion excluded patients and the upper limit of 95%-CI of the normal population. mean age or age range in years; 2.5% is excluded by
using lower sample limit or upper sample limit. A total of 5% will be excluded if both SLs are used.
Table 4b. Estimated normal limits in motion-defined form studies, the intended percentages of persons in the general population that are excluded by use
of the estimated limit and labelled as abnormal performers, precision of sample limits (95%-CI), and percentages of abnormal performing patients according
to the sample limits.
Controls Patients Statistical test
% intended % abnormal, p1-p2 pUL-p2
Formula for SL to exclude score outside
a
Study n Age SL-values (number) 95%- CI group n SL (number) 95%-CI p-value
Minimum dot speed
Ho et al 25 4.6 M + 1.64 SD 5% (n = 1 per 25) 0-20 amblyopia FE 21 24% (n = 5) 8-47 .01 ns
[28]
(2005) = 0.50 deg/s
25 7.5 M+ 1.64 SD
= 0.26 deg/s
25 10.6 M + 1.64 SD
= 0.23 deg/s
Wang et al 32 7.3 M + 1.64 SD 5% (n = 2) 1-21 amblyopia FE 6 67% (n = 4) 22-96 < .01 .05
[30]
(2007) = 0.26 deg/s
Giaschi et al 10 4-6 M + 2.5 SD 0.6% (n = 0) 0-31 amblyopia FE 20 90% (n = 18) 68-99 < .01 < .01
[31]
(1992) = 0.24 deg/s
10 7-9 M + 2.5 SD amblyopia AE 95% (n = 19) 75-100 < .01 < .01
= 0.10 deg/s
10 10-12 M + 2.5 SD
= 0.14 deg/s
Percentage correct
Jakobson et al 19 6.1 M - 2 SD 2.3% (n = 0) 0-18 preterm 43 47% (n = 20) 31-62 < .01 .02
[33]
(2006)
19 6.1 M - 3 SD 0.1% (n = 0) 0-18 33% (n = 14) 19-49 < .01 ns
SL = sample limit, FE = fellow eye/non-amblyopic eye, AE = amblyopic eye, p1-p2 is the difference between observed proportions between patients and con-
trols. The intended percentage excluded was used as observed proportion for controls, pUL-p2 is the difference between the observed proportion excluded in
a
patients and the upper limit of 95%-CI of the normal population. mean age or age range
Systematic review motion perception
55
4
56 Chapter 4
without providing evidence for symmetrical or normal distributions of scores.

As a result of the relatively small sample sizes per age group (mean n = 23),
wide precision intervals were observed; the difference between the intended
percentage to exclude and the upper limit of its 95%-CI varied from 9 to 31%.
The only normal limit, not based on group measurements but on individual per-
[24]
formances, was the deficit ratio. The patient’s performance was considered
abnormal if the threshold was at least twice that of their matched controls.
Developmental trends
To study possible developmental trends, we compared studies using similar
tasks, based on published results. On average, control groups consisted of 27
participants per author-defined age group (range 8-93, median 21). No stud-
ies were found for children aged 2-3 years, whereas data were very limited for
the ages 3-7 years and above 10 years.
In Figure 1A, mean results of control groups in global motion direction studies
[17, 25, 27, 28, 30]
and their estimated p90 are represented. Mean scores and the
estimated p90 seem to decrease significantly with increasing age, suggesting
that the minimum proportion of coherent moving dots, needed to identify the
motion direction correctly, becomes smaller with age. Mean coherence level
and the estimated normal limit are negatively related to age (mean ρ = -0.61;
p < .03; p90 ρ = -0.64; p < .02). This is illustrated by the linear trend lines
in the figure, because the quadratic effect was not significant. Additionally,
residual analysis of the trend line for p90 indicated the risk of overestimation of
performances in multiple age groups, with a maximum of 14% coherence. The
overestimation of performances could lead to an undesirably high number of
performers labelled as abnormal. Because of the unreliability of the estimators
and the risk of overestimation of performance, we consider the trend line not
the best way to evaluate the performance of clinically evaluated children. One
should use all information in the graph, position each clinically tested child
against the data from the presented studies.
In Figure 1B, means and estimated p90s are illustrated for two global motion
target location studies with motion coherence thresholds. No sample limits
[19, 20]
were reported in these articles. Mean coherence level and the estimated
p90 are negatively related to age (ρ = -0.71; p = .06). Again, the mean scores
and the estimated p90 seem to decrease significantly with increasing age and
visual inspection seems to indicate that there is a quadratic trend, although
the regression estimators for age are not significant, probably due to the low
number of data points. Therefore, the trend lines only illustrate the overall
mean coherence level and the overall mean of the p90’s.
Figure 1. Mean results and mean + 1.28 SD for studies with comparable motion perception stimuli
with a black background and white dots
1A. Global motion studies with a single target RDK and motion direction as response are represent-
[28] [30] [27]
ed. Symbols represent different studies: Ho et al, Wang et al, Pellicano & Gibson, Raymond
[17] [25]
& Sorensen and Chow & Ho. Dot speed, test distance varied per study. Linear regression trend
was significant.
1B. Global motion studies with two RDKs and with target location as response are represented:
[20] [19]
O’Brien et al and Gunn et al. Regression estimators, except constant, were not significant.
[30]
1C. Motion-defined form studies with naming as response are represented: Wang et al and Ho et
[28]
al. Curvilinear trend was significant, but is considered not generalisable.
In Figure 1C, we present a comparison of two studies on motion-defined

[28, 30]
form. There seems to be a clear curvilinear negative trend: a perfect
negative Spearman correlation was found (ρ = -1.00; p < .01) between these
four data points and age. Younger children need a higher dot speed in order to
name a motion-defined form correctly.
58 Chapter 4
Patients’ performance compared to typically developing

children
Below, we summarize the results for each patient group, focussing on mean
differences and differences in abnormal performers. On average, control
groups consisted of 27 participants per author-defined age group (range 8-93,
median 21) and the sample sizes of patient groups varied between 6 and
43. The percentage of abnormal performing patients and applied statistics for
comparisons can be found in the last column of Table 4a-b. Tables with mean
differences between patients and controls have been included in the Appendix
2 (Included studies – comparative studies).
• Children with amblyopia, tested in the non-amblyopic eye, performed
significantly worse than age-matched controls on motion-defined form
[28] [30] [28, 30]
tasks (p < .01, p < .04 ), but not on global motion tasks.
The percentage abnormal performers was significantly higher in one global
[28] [28, 30, 31]
motion task and all three motion-defined form studies.
• Children with autism perform worse, compared to age-matched controls,
[27]
on a motion direction task using vertical motion (p < .01) or rotational
[15] [15]
motion (p < .01), but not on a horizontal motion task. Compared to
verbal mental age-matched controls, children with autism perform equally
well or even outperform the controls on motion direction tasks (horizontal
[15]
motion: p < .05). Children with autism spectrum disorder did not per-
form significantly different from age-matched controls on global motion
[26]
location tasks. In one of two studies reporting the percentage abnormal
performers, the percentage abnormally performing patients was signifi-
[27] [27]
cantly higher than in controls. The percentage abnormal performers
was even significantly higher than the population limit of 14% (the upper
limit of the controls’ 95%-CI), indicating that children with autism have an
increased risk to perform abnormally on global motion tasks.
• Children with dyslexia, compared to age-matched controls, perform worse
[17, 25, 27]
on global motion direction tasks (all p < .01). Their performance
[17]
seems to deteriorate with increasing number of frames and to improve
[25]
if blue dots are used instead of red or white dots. The percentage ab-
normally performing patients is significantly higher than that observed in
[17, 27]
the control group.
• Currently, there is no evidence that children with pure dyspraxia perform
worse than age-matched controls on a global motion target location
[20]
task.
• To assess whether or not children with hemiplegia of heterogeneous ae-
tiology performed worse than their controls on a global motion location
task, we performed a post-hoc t-test on the summary data. Children with

[19]
hemiplegia performed significantly worse (p < .01).
• There is no evidence, that children with intellectual disability perform worse
[13]
than mental age-matched controls on a biological motion task.
• Prematurely born children with complications, like retinopathy of prematu-
rity and/or periventricular brain injury, seem to perform worse on motion-
defined form tasks (p < .05), whereas those without complications do not
[33]
perform significantly different from controls. Although the observed
percentage abnormal performers was significantly higher than observed
4
in controls for both sample limits, the percentage abnormally performing
patients was only significantly higher than expected in the general popula-
tion for the sample limit ‘mean minus 2 SD’.
• Children with very low birth weight (<1500 g), compared to age-matched
[24]
controls, perform worse on a global motion direction task (p < .01).
• Children with Williams syndrome, as compared to mental age-matched
controls, perform significantly worse on a motion-defined form task, in
[14]
which a target had to be located (p < .01) , and on a 3D global motion
[12] [12, 14]
task (p < .01), but not on 2D or simple global motion tasks. They
perform equally well or even outperform mental age-matched controls on
[14, 16]
biological motion tasks (p < .05).
We conclude that, considering group mean performances, multiple patient

groups perform significantly worse than controls on one or more motion per-
ception tasks.
In a majority of studies reporting a sample limit, the percentage of abnormal
performers, too, was significantly higher in patients than in controls. However,
if the upper limit of the 95%-confidence interval of mean performance, as
calculated by us (Table 4a and 4b), would be considered as the true popula-
tion limit, there is only limited evidence for an increased risk of abnormal
performance on global motion direction tasks in children with autism, and on
motion-defined form tasks in children with monocular amblyopia and prema-
turely born children.
Discussion
This systematic review shows that investigators around the world have in-
vested considerable energy into the development and evaluation of tests,
in order to enable quantified judgement of aspects of motion perception in
60 Chapter 4
children. It appears that global motion has been studied most extensively,
whereas a smaller number of studies focussed on motion-defined form and
biological motion. Because this field is still in a pioneering phase, there is a
wide variety of task characteristics, procedures and thresholds. Sample sizes
of age-groups are still small, so currently available estimates of normal limits
are insufficiently precise. First results suggest that developmental trends
are present for global motion and motion-defined form, whereas insufficient
data are available for biological motion. Based on comparison studies, first
potential risk groups have been identified. However, we must conclude that at
this stage, no motion perception tasks have been evaluated satisfactorily for
application as diagnostic instruments in clinical practice.
In our opinion, building further upon these pioneering activities, further
research should focus on international consensus development, test standar-
disation and collaboration of research groups, in order to include satisfactory
study samples.
Consensus is needed on: 1. tasks characteristics and task procedures;
2. (estimated) normal limits and age groups; 3. confounding risks for task
performances and group differences. We have formulated the following first
suggestions:
Ad 1. Because stimuli with a black background and white dots are most com-
mon, we suggest that at least a black and white stimulus should be included
in all studies.
Dot life-time should be limited, to prevent single dot tracking and dot speed
and motion direction should also be standardised.
The choice of dot speed might depend on the brain systems one wants
to study, and on the presence of additional disorders, like nystagmus. The
indication that the activity of V1 decreases with increasing motion speed and
that a direct route from the retina to the superior colliculus and pulvinar to the
[39, 40]
prestriate cortex might also be involved in motion perception, could be
a good reason to study multiple speeds. In this way, the integrity or develop-
mental status of multiple routes could be studied. We suggest, that at least a
slow speed (< 6 deg/s), at which V1 is activated before V5, and a high speed
task (> 15 deg/s), at which the colliculo-prestriate cortical route is assumed
to be the primary route, are studied.
The co-presence of nystagmus warrants special attention. In patients with
congenital nystagmus, adaptive perceptual processes that are considered to
prevent oscilliscopia (illusory motion of the visual world) are suggested to
[38, 41, 42]
result in a loss of sensitivity to motion of a stimulus. Elevated thresh-
olds as a result of adaptive perceptual processes seem to be more likely at low
[38, 42]
speeds, horizontal motion direction, which is often the direction of the
[38, 41] [38]
major component in congenital nystagmus, and larger stimulus size.
If the retinal image motion in patients with congenital nystagmus is simulated,
by adding continuous sinusoidal or ramp motion to stimuli for controls, motion
[41]
sensitivity of patients with nystagmus seems similar to that of controls. If
congenital nystagmus is present in the patient group, as additional disorder,
one might consider studying vertical motion with a high dot speed, or an
analysis including and excluding patients with nystagmus, for an indication of
the effect of nystagmus on the results.
4
Considering the stimuli described in this review and the above mentioned
criteria, we conclude for pragmatic reasons that the stimuli described by the
following authors may be eligible for standardisation of slow speed stimuli:
[27]
Pellicano and Gibson for global motion direction if dot speed is slightly
[26] [20]
reduced, Milne and co-authors or O’brien and co-authors, for global
motion target location if limited dot life-time is guaranteed, and Reiss and
[14]
co-authors for biological motion.
Target location stimuli for global motion and biological motion are preferable
in young children, because target detection tasks appear to be easier to ac-
complish than motion direction tasks at that age.
Although no limited dot life-time seems to be used in motion-defined form
[28]
stimuli, the stimuli described by Ho et al are studied most extensively and
therefore may be eligible for standardisation.
At this moment, we have no suggestions for high-speed stimuli with a black
background and white dots, because this has infrequently been studied and
only was used in motion direction tasks.
Ad 2. Current sample sizes are too small to set reliable sample limits and to es-
timate the prevalence of motion perception problems. Reliable percentile norms
[43]
would require groups of at least 120 participants per age category. Which
age groups should be studied, could be determined by studying the develop-
mental trend more extensively, ideally by doing a longitudinal study. However,
this would take many years and participants’ performances might improve over
time, due to growing task experience. The alternative is to cross-sectionally
study large study populations and identify relevant age groups in the analysis.
We realise that large samples, specifically of patients, are difficult to recruit and
therefore might require international collaboration of research groups.
Other normal limits than sample-based limits, such as a deficit ratio, seem
to be unsuitable for diagnostic purposes, because if a patient is matched to a
single control, normal variation in controls is disregarded and there is a risk of
erroneous judgements.
62 Chapter 4
Ad 3. Another matter that needs consideration concerns factors that may

confound task performances. We presume that age, visual acuity and IQ
or socio-economic status (SES) are possible confounders or indicators for
confounders. Age needs no discussion, because of the developmental trends
observed in many studies. Visual acuity, SES and IQ might be less obvious.
Most studies used a dot size larger than 3 arcmin, indicating that a single dot
should be visible for children with low vision (visual acuity < 0.3). However,
the presence of multiple dots might influence visibility, perception and task
performance negatively. The assumption of normal visual acuity and therefore
visibility of the stimulus is realistic in controls, but not in patients. Subnormal
visual acuity or even low vision can be missed easily, especially in children with
motor or intellectual disabilities. Either it should be checked that a stimulus
and its details at near distance are equally visible for patients and controls,
or the effect of visual acuity should be specifically addressed in the analysis.
Information on SES and/or IQ should be included, too. Although we could
not find articles on SES or IQ in relation to motion perception, there is some
evidence that SES and environmental factors related to SES may influence
[44, 45]
IQ as well as neural system development and cognitive functioning.
[8, 46]
Small group differences were found between low SES en moderate SES
on tasks for visual cognition (ventral stream) and spatial cognition (dorsal
[8, 46]
stream). Information on IQ or SES in the control group could also be
used to rule out selection bias for the control group.
At this stage, what can we do in clinical practice? We cannot yet reliably
label individual children as normal or abnormal performers. By calculating
95%-confidence intervals for sample limits of typically developing children,
we were able to indicate maximum percentages of the typical population that
are likely to score above the sample p90-score. Accepting the fact that the
maximum percentage could be true, one is able to say that a score above this
limit indicates that a participant belongs to the weakest performers, as an
indication of risk. So, we may already introduce motion perception tasks as
observational instruments with quantifiable outcomes, and deliver very cau-
tious statements about children’s performances.
Acknowledgements
We thank Paul Looijestein for his expert contribution to the discussion of the
results.
References
blind” patient, assessed with limited-lifetime random dot stimuli. J Neurosci,
1991. 11(2): p. 454-61.
2. Zihl, J., D. von Cramon, and N. Mai, Selective disturbance of movement vision
after bilateral brain damage. Brain, 1983. 106 (Pt 2): p. 313-40.
3. Braddick, O., J. Atkinson, and J. Wattam-Bell, Normal and anomalous develop-
ment of visual motion processing: Motion coherence and ‘dorsal-stream vulner-
ability’. Neuropsychologia, 2003. 41(13): p. 1769-1784.
4. Grossman, E.D., L. Battelli, and A. Pascual-Leone, Repetitive TMS over posterior
STS disrupts perception of biological motion. Vision Res, 2005. 45(22): p. 2847-
4
53.
5. Grossman, E.D. and R. Blake, Brain Areas Active during Visual Perception of
Biological Motion. Neuron, 2002. 35(6): p. 1167-75.
6. Thomas, B.H., et al., A process for systematically reviewing the literature: provid-
ing the research evidence for public health nursing interventions. Worldviews Evid
Based Nurs, 2004. 1(3): p. 176-84.
7. Duyme, M., A.C. Dumaret, and S. Tomkiewicz, How can we boost IQs of “dull
children”?: A late adoption study. Proc Natl Acad Sci U S A, 1999. 96(15): p.
8790-4.
8. Noble, K.G., M.F. Norman, and M.J. Farah, Neurocognitive correlates of socioeco-
nomic status in kindergarten children. Dev Sci, 2005. 8(1): p. 74-87.
9. Tong, S., et al., Socioeconomic position, maternal IQ, home environment, and
cognitive development. J Pediatr, 2007. 151(3): p. 284-8, 288 e1.
10. Newcombe, R.G., Two-sided confidence intervals for the single proportion: com-
parison of seven methods. Stat Med, 1998. 17(8): p. 857-72.
11. Herrera, L., The precision of percentiles in establishing normal limits in medicine.
J Lab Clin Med, 1958. 52(1): p. 34-42.
12. Mendes, M., et al., Visual magnocellular and structure from motion perceptual
deficits in a neurodevelopmental model of dorsal stream function. Brain Res Cogn
Brain Res, 2005. 25(3): p. 788-98.
13. Moore, D.G., R.P. Hobson, and M. Anderson, Person Perception - Does It Involve
Iq-Independent Perceptual Processing. Intelligence, 1995. 20(1): p. 65-86.
14. Reiss, J.E., J.E. Hoffman, and B. Landau, Motion processing specialization in Wil-
liams syndrome. Vision Research, 2005. 45(27): p. 3379-3390.
15. Del Viva, M.M., et al., Spatial and motion integration in children with autism.
Vision Research, 2006. 46(8-9): p. 1242-1252.
16. Jordan, H., et al., Intact perception of biological motion in the face of profound
spatial deficits: Williams syndrome. Psychological Science, 2002. 13(2): p. 162-
167.
17. Raymond, J.E. and R.E. Sorensen, Visual motion perception in children with dys-
lexia: Normal detection but abnormal integration. Visual Cognition, 1998. 5(3):
p. 389-404.
18. Spencer, J., et al., Motion processing in autism: evidence for a dorsal stream
deficiency. Neuroreport, 2000. 11(12): p. 2765-2767.
64 Chapter 4
19. Gunn, A., et al., Dorsal and ventral stream sensitivity in normal development and
hemiplegia. Neuroreport, 2002. 13(6): p. 843-847.
20. O’Brien, J., et al., Form and motion coherence processing in dyspraxia: evidence
of a global spatial processing deficit. Neuroreport, 2002. 13(11): p. 1399-1402.
21. Ellemberg, D., et al., Better perception of global motion after monocular than
after binocular deprivation. Vision Res, 2002. 42(2): p. 169-79.
with Williams syndrome: measures of dorsal-stream and frontal function. Dev
Neuropsychol, 2003. 23(1-2): p. 139-72.
23. Ellemberg, D., et al., Putting order into the development of sensitivity to global
motion. Vision Research, 2004. 44(20): p. 2403-2411.
24. MacKay, T.L., et al., Deficits in the processing of local and global motion in very
low birthweight children. Neuropsychologia, 2005. 43(12): p. 1738-48.
25. Chow, E.M.-C. and C.S.-H. Ho, Visual motion perception in Chinese dyslexic chil-
dren. Journal of Psychology in Chinese Societies, 2005. 6(2): p. 161-178.
26. Milne, E., et al., Motion and form coherence detection in autistic spectrum dis-
order: Relationship to motor control and 2 : 4 digit ratio. Journal of Autism and
Developmental Disorders, 2006. 36(2): p. 225-237.
27. Pellicano, E. and L.Y. Gibson, Investigating the functional integrity of the dorsal
visual pathway in autism and dyslexia. Neuropsychologia, 2008. 46(10): p. 2593-
2596.
28. Ho, C.S., et al., Deficient motion perception in the fellow eye of amblyopic chil-
dren. Vision Research, 2005. 45(12): p. 1615-1627.
29. Parrish, E., et al., The maturation of form and motion perception in school age
children. Vision Research, 2005. 45(7): p. 827-837.
30. Wang, J., C.S. Ho, and D.E. Giaschi, Deficient motion-defined and texture-defined
figure-ground segregation in amblyopic children. Journal of Pediatric Ophthalmol-
ogy & Strabismus, 2007. 44(6): p. 363-371.
31. Giaschi, D.E., et al., Defective Processing of Motion-Defined Form in the Fellow
Eye of Patients with Unilateral Amblyopia. Investigative Ophthalmology & Visual
Science, 1992. 33(8): p. 2483-2489.
32. Schrauf, M., E.R. Wist, and W.H. Ehrenstein, Development of dynamic vision
based on motion contrast. Exp Brain Res, 1999. 124(4): p. 469-73.
33. Jakobson, L., V. Frisk, and A. Downie, Motion-defined form processing in ex-
tremely premature children. Neuropsychologia, 2006. 44(10): p. 1777-1786.
34. Pavlova, M., et al., Recognition of point-light biological motion displays by young
children. Perception, 2001. 30(8): p. 925-933.
35. Blake, R., et al., Visual recognition of biological motion is impaired in children with
autism. Psychological Science, 2003. 14(2): p. 151-157.
36. Freire, A., et al., The development of sensitivity to biological motion in noise.
Perception, 2006. 35(5): p. 647-57.
37. Lichtensteiger, J., et al., Role of dorsal and ventral stream development in biologi-
cal motion perception. Neuroreport, 2008. 19(18): p. 1763-1767.
38. Shallo-Hoffmann, J., et al., Reduced duration of a visual motion aftereffect in
congenital nystagmus. Doc Ophthalmol, 1998. 95(3-4): p. 301-14.
40. Chawla, D., et al., Speed-dependent motion-sensitive responses in V5: an fMRI
study. Neuroimage, 1998. 7(2): p. 86-96.
41. Bedell, H.E., Sensitivity to oscillatory target motion in congenital nystagmus.
Invest Ophthalmol Vis Sci, 1992. 33(5): p. 1811-21.
42. Lappin, J.S., et al., Spatial and temporal limits of motion perception across varia-
tions in speed, eccentricity, and low vision. J Vis, 2009. 9(1): p. 30 1-14.
43. Reed, A.H., R.J. Henry, and W.B. Mason, Influence of statistical method used on
the resulting estimate of normal range. Clin Chem, 1971. 17(4): p. 275-84.
44. Kaplan, G.A., et al., Childhood socioeconomic position and cognitive function in 4
adulthood. Int J Epidemiol, 2001. 30(2): p. 256-63.
45. Turkheimer, E., et al., Socioeconomic status modifies heritability of IQ in young
children. Psychol Sci, 2003. 14(6): p. 623-8.
46. Farah, M.J., et al., Childhood poverty: specific associations with neurocognitive
development. Brain Res, 2006. 1110(1): p. 166-74.
Chapter
Chapter 4:
4: Appendix
Appendix 1
1
Table 1. Characteristics of the global motion studies
Table 2. Characteristics of motion-defined form studies
Table 3. Characteristics of the biological motion studies

Table 1. Characteristics of the global motion studies: thresholds, responses, stimulus characteristics (stimulus type, motion direction, speed, dot size) and
test distance.
Test
Motion direction Dot size distance
Threshold Response Stimulus type per trial Speed (deg/s) (arcmin) (cm)
[4, 5] [9, 10]
1. Motion coherence 1. Motion direction 1. RDK, black background white dots 1. up or down Slow (< 6 deg/s) - 0.84 - 40
[1-17] [1, 3-6, 11-14, 16] [1-10, 13, 15, 17] [3-6, 11, 12, 14] [6] [1]
level - 1.2 - 2.5 - 50
[4, 6] [4, 5] [13] [1, 3, 7, 8, 14]
2. Dmax 2. Target location - Single RDK 2. left or right - 1.26 - 3
[2, 7-10, 15] [1, 3-6, 13] [1, 2, 13, 16] [11] [15] [17]
o Target RDK - 1.5 - 6 - 56
[17] [17] [3]
3. Target or non-target o Target RDK or noise-RDK 3. oscillating horizontal - 2.0 - 6.6 - 57
[17] [7-10, 15] [2] [11, 15, 16]
o 3D target RDK or (2D) noise RDK motion - 2.5 - 6.7x10.1
[17] [17] [17] [2]
4. 0-360 deg (pseudo- - 3.0 - 61
[1] [17] [10] [2] [13]
- RDK adjacent to noise-RDK random) - 5.3 - 22.8 - 65
[7, 8] [16] [4-6]
- RDKs left and right of the centre 5. rotation on spherical - 5.8 - 24 - 140
[2, 15] [17] [12, 14]
o Target RDK vs noise-RDK surface Intermediate - 30
o RDK with target strip vs 6. rotation clockwise or (6-15 deg/s)
[7-10] [16] [9, 11]
undivided-RDK anti-clockwise - 6
[3]
2. Single RDK, black background colour 7. contraction or - 6.3
[13] [16] [15]
dots expansion - 7.0
[11]
3. Single RDK, white background black - 9.0
[12, 14] [16]
dots - 10
[1]
4. Single RDK, grey background black and - 11
[16]
white dots High (> 15 deg/s)
[12, 14]
5. Single RGK (circular random gabor - 18
kinematogram) with static noise
[11]
background
- Gabors luminance modulated
- Gabors contrast modulated
Appendix 1 Characteristics of motion perception studies
69
4A1
70
Table 2. Characteristics of motion-defined form studies: thresholds, responses, stimulus characteristics (stimulus type, motion direction, speed, dot size)
and test distance.
Test
Motion direction per Dot size distance
Threshold Response Stimulus type trial Speed (deg/s) (arcmin) (cm)
[4-6] [2]
1. Minimum dots speed 1. Identification 1. RDK, black background, white dots 1. motion in form opposite Slow (< 6 deg/s) - 3.8 - 61
[4-6, 18] [4-6] [20] [18] [19]
- Form - With 1 of 5 forms (circle, duck, to background motion - 0.021 - 0.68, - 15.6 - 100
[18] [2, 18, 20] [4-6] [2] [20]
2. Accuracy - Letter fish, heart, gingerbread man) - left or right - 0.039 - 1.26 - 22.8 - 200
Chapter 4: Appendix 1
[4-6] [4-6] [18] [4, 5]

- Weighted percentage 2. Gap position - up or down - ≤0.45 - 560
[19] [19, 20] [18] [19] [6]
correct response - With 1 of 10 letters 2. left or right in form and - 1.3 - 580
[2] [19] [2] [18]
- percentage correct 3. Target location - With Landolt C in 1 of 4 static background - 2.51 - 600
[20] [19]
response orientations,
[2] [20]
3. Motion coherence level - With E in 1 of 4 orientations
- RDK left and right of centre,
target RDK contains vertical
rectangular figure, non-target
RDK contains horizontal
[2]
rectangular figure
Table 3. Characteristics of the biological motion studies: thresholds, responses, stimulus characteristics (stimulus type, motion direction, speed, dot size)
and test distance.
Motion Dot Test
direction per size distance
Threshold Response Stimulus type trial Speed (deg/s) (arcmin) (cm)
[21-25] [2, 22-24, 26] [25] [21]
1. Accuracy 1. Naming 1. Point-light figure, black background, white dots object motion Slow (< 6 deg/s) - 10 - 41
[2] [21] [25]
- number - Naming only - Single figure, walking human or object (bicycle, ball, direction - 2.51 - 12 - 60
[26] [26] [21, 25] [24] [2]
of correct scissors, chair) - across - 4 - 13.8 - 61
- [22] [22]
responses - Naming and - Single figure, 1 of 4 figures (walking man, running dog, screen gait cycle 15 - 92
[23-25] [24] [24] [2]
report facing walking dog, bird) o left or 1.76 sec - 22.8 - 200
[2] [24]
- percentage direction - Target, 1 of 3 human activities (walking, jumping, right 2.0 sec
[26] [24]
correct of figure waving), or non-target, phase scrambled point-light figure fly cycle
[22] [23]
response and motion - like in 1.98 sec
[24]
- d’ , unbiased direction o without noise treadmill
measure of 2. Target or non- o with pseudo-random noise o left or
[21] [21, 23, 25] [22]
sensitivity target - target, walking man, in noise right
[2, 21-25]
2. Minimum 3. Walking direction o static noise
[22]
exposure o random moving noise o frontal
[26]
duration 4. Target location o yoked moving noise (bird)
[2] [24]
3. Reaction time - Target, walking man, vs non-target, phase scrambled
[23]
point-light figure, left and right of the centre
[2]
4. Noise at certain o With yoked moving noise
[21, 25]
accuracy level 2. Point-light figure, grey background, black dots
- Number of - Target, 1 of 5 human activities (running, kicking,
[25]
noise dots climbing, throwing, jumping) or, phase scrambled point-
- Motion light figure
coherence o Without noise
[2]
level o With noise
1 arcmin = 1/60 degree. For a visual acuity of 1.0 one has to be able to discriminate an element 1 arcmin in size, a Snellen-E of 5 arcmin.
d’ , unbiased measure of sensitivity, is based on hits (target indicated as target) and false alarms (non-target indicated as target). A higher positive value
indicates a higher sensitivity to the target.
Appendix 1 Characteristics of motion perception studies
71
4A1
72 Chapter 4: Appendix 1
References
p. 389-404.
2596.
ogy & Strabismus, 2007. 44(6): p. 363-371.
Neuropsychol, 2003. 23(1-2): p. 139-72.
Brain Res, 2005. 25(3): p. 788-98.
Science, 1992. 33(8): p. 2483-2489.
Appendix 1 Characteristics of motion perception studies 73
167.
cal motion perception. Neuroreport, 2008. 19(18): p. 1763-1767. 4A1
Perception, 2006. 35(5): p. 647-57.
Chapter
Chapter 4:
4: Appendix
Appendix 2
2
Included studies - Comparative data

1. Amblyopia
2. Autism
3. Dyslexia
4. Dyspraxia
5. Hemiplegia
6. Intellectual disability
7. Premature born children
8. Children with very low birth weight
9. Williams syndrome
Included studies - Developmental data
Excluded studies - Comparative data
Excluded studies - Developmental data

Appendix 2 Outcomes of motion perception studies 77
Included studies – comparative studies
1. Amblyopia
Global motion & Motion-defined form

Global motion studies with motion direction as response and motion coherence level
(%;(Signal dots/(Signal dots + Noise dots)*100) and/or Dmax (deg) as outcome.
Motion-defined form studies with identification as response and minimum dots speed
(deg/s) as outcome.
[1]
Ho et al (2005)
Amblyopia,
non-amblyopic eye Controls Statistics
n 21 21
4A2
Gender (% male) - -
Age range in years 4.4-11.0 4.3-11.2
(mean; SD) (6.9; 1.7) (7.0; 1.9)
IQ information - -
Motion perception MANOVA
Group: F(3, 38) = 4.71, p < .01
Global motion
a b
Mean motion coherence % 26 25 F(1, 40) = 0.01,
(SD; range) (22; 3-81) (18) p = .91, f = 0.00
a b
Dmax deg 1.24 1.04 F(1, 40) = 0.23,
(SD; range) (0.79; 0.49-3.70) (0.26) p = .63; f = 0.08
Motion-defined form
a b
Minimum dots speed deg/s 0.20 0.10 F(1, 40) = 13.55,
(SD; range) (0.13; 0.07-0.61) (0,04) p < .01; f = 0.58
a b
Data was extracted from table with individual data and figure with group results; Results if abnor-
mal performers are excluded: Motion coherence %: 19 (14; 3-47); Dmax : 0.99 (0.24; 0.55-1.37);
deg/s: 0.16 (0.08; 0.07-0.34).
[2]
Wang et al (2007)
Amblyopia,
a
non-amblyopic eye Controls Statistics
n 6 32
Gender (% male) -
Age range in years 5.81-8.66 5-8
(mean; SD) (6.8; 1.1) (7.3; 1.1)
IQ information - -
Motion perception MANOVA
Group: F(2, 9) = 4.50, p < .04
Global motion
Mean motion coherence % 20 20 F(1, 10) = 4.133,
(SD; range) (16; 9-47) (15) p < .07, ns
Motion defined form
Mean minimum dot speed deg/sec 0.39 0.14 ns
(SD; range) (0.32; 0.07-0.91) (0.08)
a
an age matched control group (n = 6) was used for analysis (mean age 6.8 years, SD = 1.1),
post-hoc univariate F-test was controlled for multiple comparisons with Bonferroni adjustment set at
0.025 for each task. Also a texture defined task was used.
2. Autism
Global motion
Global motion study with target location as response and motion coherence level
(%;(Signal dots/(Signal dots + Noise dots)*100) as outcome.
[3]
Milne et al (2006)
Autism spectrum
disorder Controls Statistics
n 23 23
Gender (% male) 96 43
(mean; SD) (10.08; 1.67) (10.25; 1.08)
a
Mean non-verbal IQ 95 102
(SD; range) (14.2; 70-122) (14.1; 84-130)
Global motion
Mean motion coherence level % 17.09 10.26 Mann-Witney: ns
(SD; range) (15.22; 6.24-55.14) (4.13; 3.55-19.65)
a
Outcomes of Raven’s (standard) progressive matrices were used to calculate non-verbal IQ.
Global motion
Global motion studies with motion direction as response and motion sensitivity ((Signal dots +Noise
dots)/Signal dots) or motion coherence level (%;(Signal dots/Noise dots)*100) as outcome.
[4]
Del Viva et al (2006)
Verbal mental
Age matched age matched
Autism controls controls Statistics
n 10 14 12
Gender (% male) - - -
Age range in years 6.0-14.1 8.0-11.9 6.1-7.2
(mean; SD) (8.8; 3.0) (9.7;0.8) (6.6; 0.3)
Mean verbal 6.7 - matched
a
mental age years (2.5;4.4-12.3)
Global motion ANOVA
Mean motion Group: F = 5.8, p < .01
sensitivity (%) per Control groups: p < .05
test condition Patients vs. age-matched
controls: ns
Patients vs. verbal mental age
matched controls: ns
Rotation 14.4 23.1 18.2 Student t-test
(SD; range) (3.7; 7.9-21.2) (5.2;12.7-30.4) (6.0; 9.0-27.2) Patients vs. Age-matched
Horizontal 20.3 18.1 14.4 controls:
(SD; range) (7.8;12.0-36.8) (2.7;13.1-21.9) (2.8; 11.7-19.2) Rotation: p < .01
Expansion & horizontal: ns
Contraction/ 15.1 16.1 13.3
Patients vs. Verbal mental age
Expansion (3.0; 11.1-22.3) (3.8;11.7-26) (2.1; 9.1-16.5)
matched controls:
(SD; range)
Horizontal: p = .02
Expansion
(p = .1) & rotation (p = .05):
ns
a
Data of the WISC-R was used to estimate the verbal mental age ((chronological age*VIQ)/100).
The reported mean verbal age is that of the total patient group (n = 13, mean chronological age in
years = 10.9, SD = 4.2).
[5]
Pellicano & Gibson (2008)
ab
Autism Controls Statistics
n 20 61
(mean; SD) (9.59; 1.38) (9.88; 1.01)
c
Mean non-verbal IQ 107.10 106.82
(SD) (8.95) (12.90)
c
Verbal IQ 97.00 104.28
(SD) (15.25) (11.60)
Global motion 4A2
Mean motion coherence 22.40 14.02 ANOVA
level % (13.78; 5.31-54.44) (6.70; 4.41-40.53) Group: F(2, 119) =
(SD; range) 10.82, p < .001
t(119) = 2.41, p < .01,
d = 1.01
a
Effect of outliers was reduced by scores more than 3 SD above/below the group mean were re-
b
placed by a score at 2.5 SD before calculation of standard scores and analysis took place; Also a
c
group with dyslexia was included. Autism vs. dyslexia: t(119) = 1.08, ns; Outcomes of Raven’s
(standard) progressive matrices or Wechsler Abbreviated Scale of Intelligence (WASI) were used
to assess non-verbal IQ, outcomes of the Peabody Picture Vocabulary Test or Wechsler Abbreviated
Scale of Intelligence were used to assess verbal IQ.
3. Dyslexia
Global motion
Global motion studies with motion direction as response and motion coherence level (%;(Signal
dots/(Signal dots + Noise dots)*100) as outcome.
[6]
Raymond & Sorensen (1998)
a
Experiment 1a Dyslexia Controls Statistics
n 10 10
(Mean; SD) (9.9; 1.2) (9.9; 1.2)
Mean IQ 98 At least average
(SD) (9.8)
Reading level At least 1.5 years Assumed at age
below age level level
Global motion
Mean motion coherence level (%) per
stimulus condition
Noise and no-noise, 60 ms 39.8 19.9 p < .01
b
(SD; range ) (17.2; 18-70) (2.9; 17-26)
Experiment 2 Dyslexia Controls Controls Statistics
n 12 12 8
Gender (% male) 50 41 38
Age range in years - - -
(mean; SD) (11.6) (9.3) (8.0)
IQ Mean (SD) - -
Reading level At least 1.5 years Assumed at age
below age level level
Global motion
Mean motion coherence level (%) per
stimulus condition
2 short frames, 64 ms 26.3 25.4 - ns
b
(SD; range) (15; 14-56) (14; 9-48)
7 short-frames, 224 ms 19.9 10.5 - p < .01
b
(SD; range) (8; 8-34) (4; 4-16)
2 long-frames, 224 ms - - 25
b
(SD; range) (10; 11-41)
a b
Experiment 1b is excluded, because the sample size of the control group was too small; Data was
partially extracted from a figure with individual results.
[7]
Chow & Ho (2005)
Dyslexia Controls Statistics
n 24 24
Gender (% male) - -
Age range in years - -
(mean; SD) (9.16; 1.38) (9.12; 1.29)
a
(SD) (8.28) (7.10)
Global motion
Mean motion coherence level
(%) per test condition
White dots 36.59 24.73 Two-way mixed design
4A2
(SD) (8.63) (2.26) ANOVA:
Group x Colour:
Blue dots 31.97 23.30
F(2,41) = 5.83,
(SD) (7.91) (0.87)
p < .01
Red dots 35.83 24.81 Group: F(1,47) =
(SD) (10.62) (1.93) 37.58, p < .01
Total 34.80 24.28
(SD) (9.23) (1.90)
a
Outcomes of Raven’s (standard) progressive matrices were used to calculate non-verbal IQ.
[5]
Pellicano & Gibson (2008)
ab
Dyslexia Controls Statistics
n 41 61
(mean; SD) (9.97; 1.10) (9.88; 1.01)
c
(11.38) (12.90)
c
Verbal IQ 110.15 104.28
(SD) (10.05) (11.60)
Global motion
Mean motion coherence 26.38 14.02 ANOVA
level % (19.60; 5.19-75.50) (6.70; 4.41-40.53) Group: F(2, 119) =
(SD; range) 10.82, p < .001
t(119) = 4.54,
p < .001, d = 1.32
a
Effect of outliers was reduced by scores more than 3 SD above/below the group mean were re-
b
placed by a score at 2.5 SD before calculation of standard scores and analysis took place; Also a
c
group with dyslexia was included. Autism vs. dyslexia: t(119) = 1.08, ns; Outcomes of Raven’s
(standard) progressive matrices or Wechsler Abbreviated Scale of Intelligence (WASI) were used
to assess non-verbal IQ, outcomes of the Peabody Picture Vocabulary Test or Wechsler Abbreviated
Scale of Intelligence were used to assess verbal IQ.
4. Dyspraxia
Global motion
Global motion study with target location as response and motion coherence level
(%;(Signal dots/Noise dots)*100) as outcome.
[8]
O’Brien et al (2002)
Chronological age
and verbal mental
Pure dyspraxia age matched controls Statistics
n 8 50 F(1,56) = 1.2; ns
Gender (% male) 75 -
Age range in years 7-11 matched
(mean; SD) (8.2; 1.5) (8.4)
a
Verbal mental age matched matched
Global motion
Mean motion coherence level 14.9 17.5
% (SD) (7.9) (6.1)
a
Outcomes of the British Picture Vocabulary Scale were used to estimate mental age.
5. Hemiplegia
Global motion
Global motion study with target location as response and motion coherence level as
outcome.
[9]
Gunn et al (2002) Threshold: Motion coherence level (%); Response: target Location
a
Hemiplegia Controls Statistics
n 22 295 repeated measure
ANOVA with age as
Gender (% male)
covariate: interaction
b
Age range in years 3.2-12.4 4.10 –11.99 task*group :
(mean; SD) (-) (-) F(1,381) = 5.6;
p < .02
IQ information - -
t-test done by
Global motion reviewers for threshold
Mean coherence level % 36.8 28.8 difference
(SD; range) (10.4; 26.4-61.4) (8.4) p < .01
a
Two patients were excluded because of the age criterion, data was therefore estimated from figures
b
with individual data; also a global form task was done.
6. Intellectual disability
Biological motion
Biological motion study with identification as response and minimum exposure duration
as outcome.
[10]
Moore et al (1995) Threshold:Minimum exposure duration (ms), response naming
Intellectually Verbal mental age
disabled matched controls Statistics
n 15 15
Gender (% male) - -
Age range in years 9.75 -16.42 11.92-10.58
(mean; SD) (14.17; 1.67) (8.42;1.17) 4A2
a
Verbal mental age 6.00-10.33 6.25-10.33
(mean, SD) (8.08; 1.33) (8.33; 1.25)
a
Verbal IQ 61.5 100.4
(mean, SD) (7.3; 49-71) (4.3; 91-106)
Biological motion
Mean exposure duration in ms
b
per test condition
c
First presentation: 5PLW 203 309 Sign test: ns
(SD; range) - -
c
Second presentation: 10PLW 123 149 Sign test: ns
(SD; range) - -
Total 163 229 Sign test: p = .09, ns
(SD; range) - -
a
Outcomes of the British Picture Vocabulary Scale were used to match subjects pair wise and assess
b c
verbal IQ; Data was partially extracted from a figure with individual results; 5PLW = point-light
walker made of 5 dots, 10PLW is point-light walker made of 10 dots.
7. Premature born children
Motion-defined form
Motion-defined form study with gap position location as response and percentage correct responses
as outcome.
[11]
Jakobson et al (2006) Threshold: Percentage correct response &, response: gap position
Preterm Age
and SES
matched
No ROP PVBI ROP and full-term
and PVBI ROP only only PVBI Total controls Statistics
n 11 12 10 10 43 19
Gender (% male) - - - - 34 47
Age range in - - - - 5.25-6.83 5.25-6.83
years (6.08; (6.08;
(mean; SD) 0.52) 0.52)
IQ information - - - -
Performance IQ - - - - 90.5 109.3
(SD) (10.6) (13.0)
Verbal IQ - - - - 91.0 114.3
(SD) (10.8) (16.0)
Motion-defined
letter
a
correct % 61 47 46 37 48 71 ANOVA
(SD) (14) (11) (13) (13) (16) (19) Group:
F(4, 57) =
10.16,
p < .01,
2
η = .42
ROP vs.
control
PVBI vs.
control
ROP and
PVBI vs.
control: p
< .05
a
PVBI = periventricular brain injury; ROP = retinopathy of prematurity; Data partially extracted
from a figure with group results.
8. Children with very low birth weight (VLBW)
Global motion
Global motion study with motion direction as response and motion coherence level
(%;(Signal dots/(Signal dots + Noise dots)*100) as outcome.
[12]
MacKay et al (2005)
VLBW, without
major neurological
a
abnormalities A term controls Statistics
n 19 19
Age range in years 5.17-8.42 4.92-8.92 4A2
(mean; SD) (6.83; 0.98) (6.83; 1.28)
Mean Verbal IQ 103.8 110.9
(SD) (9.8) (6.9)
Global motion
Mean coherence level % 29.81 9.80 MANCOVA, controlling
(SD) (29.20) (11.72) for age
F(5,29) = 4.12,
p < .01
Group: p < .01
t(22) =−2.70, p < .05
a b
Number of participants for analysis is 18, 1 control and 1 patient were not tested on global motion;
Peabody Picture Vocabulary Test- Third Edition was used to assess Verbal IQ.
9. Williams syndrome
Global motion, motion-defined form and biological motion

Global motion, motion-defined form and biological motion study with target location as
response and motion coherence level (%;(Signal dots/(Signal dots + Noise dots)*100) as
outcome.
[13]
Reiss et al (2005)
Mental age matched
c
Williams Syndrome controls Statistics
n 10 10
Gender (% male) - -
(mean; SD) (14.25; -) (6.08; -)
a
IQ test information
Composite IQ-score 63.7 119.50
(SD) (15.97) (9.04)
Mean verbal score 38.30 34.40
(SD) (7.72) (5.22)
Mean matrices score 20.80 20.10
(SD) (4.21) (3.35)
Mean coherence level (%) Mixed-model repeated
b
per test condition measure ANOVA
Group: F(3, 36) =
3.10, p < .04
Global motion 8 8 t(18) < 1.54, ns
(SD; range) (5) (-)
Motion defined form 56 32 t(18) = 4.01,
(SD; range) (11) (7) p < .001
Biological motion 23 36 t(18) = 2.38,
(SD; range) (16) (13) p < .04
a
Outcomes of the Kaufman Brief Intelligence Test were used to match patients to controls on mental
b c
age; Data was extracted from a figure with group results; Study also included adult groups
Biological motion
Biological motion study with walking direction as response and percentage correct
responses as outcome.
[14]
Jordan et al (2002) Percentage correct & walking direction
Mental age matched
b
Williams Syndrome controls Statistics
n 10 10
Gender (% male) - -
(mean; SD) (11.58;-) (6.0;-)
a
IQ test information
4A2
Composite IQ-score 62.60 116.40
(SD) (15.72) (9.83)
Mean verbal score 33.15 31.60
(SD) (6.10) (4.59)
Mean matrices score 18.50 19.16
(SD) (2.56) (3.35)
Biological motion
Mean percentage correct per Mixed-model repeated
test condition measures ANOVA
Group x S/N ratio x
Noise Type
c
S/N ratio 1:1 97.54 89.20 Groups: F(2, 25) =
(SD; range) - - 9.11, p < .01
Post hoc comparison:
p < .01
c
S/N ratio 1:3 84.26 77.78 Groups: F(2, 25) =
(SD; range) - - 9.45, p < .01
p = .18, ns
total 90.90 83.49 Groups: F(2, 25) =
(SD; range) - - 11.33, p < .01
p < .05
a
Outcomes of the Kaufman Brief Intelligence Test was used to match patients to controls on mental
b
age; Study also included a group with adult controls, only group effect and post hoc comparison
c
for the patients vs. mental-age matched controls are reported; S/N = number of signal dots (S)/
number of noise dots (N)
Global motion
Global motion study with target or non-target as response and motion coherence level as outcome.
[15]
Mendes et al (2005)
c
Williams Syndrome Mental age matched Statistics
n 6 11
Age range in years 11-20 5-14
(mean; SD) (16.00;3.52) (9.09;3.08)
Mental age 6.5-10
(8.67;1.44)
a
IQ information
Performance IQ 44.3 -
(SD) (2.58)
Verbal IQ 52 -
(SD) (6.98)
Global motion
b
Mean coherence level (%) per test condition Mann-Whitney
Global motion or noise (SD) 36 28 ns
(12) (7)
3D global motion or 2D noise 35 12 p < .01
no time pressure (SD) (12) (2)
3D global motion or 2D noise 31 26 p = .02
200 ms per trial (SD) (16) (9)
a b c
WISC-3 or WAIS was used to assess IQ; Data was extracted from a figure with group results;
Control group consisted of 8, 9, 8 age matched controls for consecutive test conditions. The patient
group consisted of 5 patients in the 3D time constraint condition.
Included studies - Developmental data

Ordered by publication year
Global motion
Global motion study with letter identification as response and Minimum dots speed
(deg/s) as outcome.
[16]
Giaschi et al (1992)
Controls
4-6 year olds 7-9 year olds 10-12 year olds Total
n 10 10 10 30
Gender (% male) - - - -
4A2
Age range in years - - - 3.67 -13.5
(mean; SD) (8.42-2.75)
IQ information - - - -
Motion-defined letter
Mean minimum dot 0.14 0.05 0.06 -
speed deg/s (0.04) (0.02) (0.03)
(SD; range)
Global motion
Global motion study with target location as response and motion coherence level as outcome.
[9]
Gunn et al (2002)
a
Controls
4 year 5 year 6-7 year 8-9 year 10-11 year
olds olds olds olds olds Statistics
n 37 93 60 50 55
Gender (% male) - - - - -
Age range in years 4.10-5.29 5.30-5.99 6.00-7.99 8.00-9.99 10.00-11.99
(mean; SD) (-) (-) (-) (-) (-)
IQ information - - - - -
Global motion
Mean motion coherence 34.2 29.5 28.0 28.8 24.9 F(5,354) =
level % (11.3) (8.2) (7.0) (7.7) (6.2) 15.6;
(SD; range) p < .001
a
data of adults is excluded
Global motion & Motion-defined form

Global motion study with motion direction as response and motion coherence level (%;(Signal dots/
(Signal dots + Noise dots)*100) and/or Dmax (deg) as outcome. Motion-defined form study with
identification as response and minimum dots speed (deg/s) as outcome.
[1]
Ho et al (2005)
Controls
3-5 year olds 6-8 year olds 9-11 year olds
n 25 25 25
Gender (% male) - - -
(mean; SD) (4.63; 0.65) (7.46;0.88) (10.65; 0.93)
IQ information - - -
Motion tests
Global motion
Mean motion coherence level % 30 22 19
(SD; range) (25) (16) (10)
Mean Dmax deg 0.90 1.03 1.00
(SD; range) (0.28) (0.27) (0.22)
Motion-defined form
Mean Minimum dots speed 0.24 0.13 0.12
deg/s (0.16) (0.08) (0.07)
(SD; range)
Excluded studies - Comparative data

Global motion
Global motion study with motion direction as response and coherence level as outcome.
[17]_1
Ellemberg et al (2002)
Congenital bilateral Congenital unilateral
a a
(CB) cataract (CU) cataract
Non-
deprived Deprived
Best eye Worst eye eye eye Controls Statistics
4A2
n 6 6 9 9 12
Gender (% male)
Age range in 5.0-11.7 5.0-11.7 4.1-12.1 4.1-12.1 -
years (7.15; 2.42) (7.15; 2.42) (6.68; 2.24) (6.68; 2.24) (6.0; 0.17)
(mean; SD)
Global motion ANOVA
Group: F(4,
54) = 4.40,
p < .01
b
Mean motion 42.7 45.5 16.8 15.6 9.0 CB and CU
coherence level % (16.5; 19-63) (19.6; 26-75) (4.2; 7-22) (3.9; 7-19) both eyes
worse than
controls: p
c
<.01.
[17]_2
Developmental unilateral
a
(DU) cataract
Non-deprived
eye Deprived eye Controls Statistics
n 7 7 12
Gender (% male)
Age range in years 6.0-12.5 6.0-12.5 -
(mean; SD) (9.63; 2.09) (9.63; 2.09) (6.0; 0.17)
IQ information - - -
Global motion
Mean motion 10.1 11.1 9.0 ns
coherence level % (2.0; 8-14) (1.6; 9-14) (-) j
a
2 patients with bilateral congenital cataract, 5 patients with unilateral congenital cataract and 2
patients with developmental cataract were excluded because they were older than 12.99 years of
age. Bilateral developmental cataract was excluded because of the low number of remaining patients
b c
within the age criterion; Data extracted from a figure with group results; Patients with congenital
unilateral cataract performed better with the deprived eye than patients with congenital bilateral
cataract with either eye (p < .01). The performance with the deprived and the non-deprived eye of
the patients with congenital unilateral cataract did not differ significantly.
Global motion
Global motion studies with target location as response and motion coherence level
(%;(Signal dots/Noise dots)*100) as outcome.
[18]
Spencer et al (2000)
n 23 50
Gender (% male) - -
(mean; SD)
a
Verbal mental age in years 7-11 7-11
Global motion
Mean motion coherence 25.5 17.5 F(1,72) = 10.98,
level % p < .01
(SD)
a
Test used to assess verbal mental age is not reported
Global motion
Global motion study with target location as response and motion coherence level as
outcome.
[19]
Atkinson et al (2003)
Williams
a d
Syndrome Statistics
n - - - - 41
Gender (% male) -
Age range in years 4-5.5 5.5-6.9 7-8 10-11 4.75-12.99
(mean; SD) (-)
b
Mental age years ≥4
Global motion
Mean motion 32 -
c
coherence level % (12)
(SD)
Median motion 31 30 24 20 29
c
coherence level %
(SD)
P90 motion coherence 54 37 30 26 -
c
level%
a b
4 patients were excluded because of the age the criterion; Outcomes of the British Picture Vo-
c
cabulary Scale were used to assess mental age in patients; Data was extracted from a figure with
d
group results for controls and individual results for patients; Patients results were only compared
to p90 of controls (percentile 90) of controls: 45% of patients performed worse than p90 percentile, if
mental age was used for performance evaluations 18% of patients performed worse than p90
Biological motion
Biological motion study with target or non-target as unbiased measure of sensitivity (d’)
as outcome.
[20]
Blake et al (2003)
n 12 9
Gender (% male)
Age range in years 8-10 5-10
(SD) (-) (8.42; 1.89)
IQ information - -
Biological motion 4A2
unbiased measure of sensitivity d’ 2.52 1.17 t(19) = 2.68,
a
(SD) (0.5) (0.7) p = .02
a
Data was extracted from a figure with group results.
Excluded studies - Developmental data

Motion-defined form
Motion-defined form study with gap position as response and (weighted) percentage
Correct response as outcome.
[21]
Schrauf et al (1999)
Controls
4 5 6 7 8 9 10 11 12
year year year year year year year year year
olds olds olds olds olds olds olds olds olds
n - - - - - - - - -
Age range in years - - - - - - - - -
IQ information - - - - - - - - -
Motion-defined form
Males
Mean percentage correct 67 67 67 69 72 75 78 80 81
ab
% (7) (7) (8) (9) (3) (2) (3) (3) (4)
(SEM)
Females
Mean percentage correct 74 (7) 74 (6) 74 (7) 74 (7) 74 (3) 74 (4) 75 (5) 76 (5) 77 (5)
a
%
a b
Data was extracted from a figure with group results; In the age group 4-6 year olds also mean
percentage correct % per coherence level was illustrated: 100% coherence 94% (SEM = 1); 50%
coherence 90% (SEM = 3); 30% coherence 79% (SEM = 4); 20% coherence 63% (SEM = 5).
Biological motion
Biological motion study with naming and facing direction as response and percentage
correct as outcome.
[22]
Pavlova et al (2001)
Controls
3 year olds 4 year olds 5 year olds
n 16 16 16
Gender (% male) 63 56 50
(SD) (3.58; 0.17) (4.42; 0.25) (5.42; 0.33)
IQ information - -
Biological motion
Naming PLW
Mean percentage correct % 23 42 100
PLW = point light walker
Global motion
Global motion study with motion direction as response and coherence level (%;(Signal
gabors/(Signal gabors + Noise gabors)*100)as outcome.
[23]
a
Controls Statistics
n 24
Gender (% male)
Age range in years 4.75-5.25 (5.0; -)
(mean; SD)
IQ information -
Global motion
Mean motion coherence level % per test
condition
a
Luminance defined Age : F(1, 46) = 24.03, p < .01
speed 1.5 deg/s 24 Speed: F(2, 46) = 26.80,
(SD) (20) p < .01
speed 6 deg/s 7
speed 9 deg/s 6
a
Contrast defined Age : F(1, 46) = 27.51, p < .01
speed 1.5 deg/s 36 Speed: F(2, 46) = 40.53,
(SD) (29) p < .01
speed 6 deg/s 8
speed 9 deg/s 7
a
An adult group was included in the original study.
Motion-defined form
Global motion study with motion direction as response and coherence level as outcome.
Motion-defined form study with identification as response and minimum dots speed
(deg/s) as outcome.
[24]
Parrish et al (2005)
Controls
3-4 5-6 7-8 9-10 11-12
a
year olds year olds year olds year olds year olds Statistics
n 11-33 11-33 11-33 11-33 11-33
Gender (% male) - - - - -
Age range in
years
- - - - -
4A2
Motion perception
Global motion MANOVA
Age group:
n 23 23 23 23 23
F(10, 130) =
Mean coherence 33 30 23 24 30 1.37, ns
level %
n 11 11 11 11 11 MANOVA
Age group:
Mean Dmax deg 0.86 0.97 1.05 1.05 1.06
F(10, 118) =
2.39,
c
p < .05
ANCOVA with
covariate visual
acuity (LogMAR)
Age group: F(5,
61) = 3.85, p
< .05
Motion-defined MANOVA
form Age group: F(8,
188) = 2.95, p
n 13 13 13 13 13 b
< .01.
Mean minimum 0.28 0.18 0.14 0.09 0.14 ANCOVA with
dot speed deg/s (0.06) (0.03) (0.02) (0.02) (0.04) covariate visual
(SD) acuity (LogMAR)
Age group: F(4,
59) = 1.89, ns
a
Global motion studies included an adult group. Each age group for the mean coherence level analy-
b
sis consisted of 12 subjects; Performance of 3-4 year olds was significantly different from other
c
groups (p-value not available); Performance of 3-4 and 5-6 year olds was significantly different
from adults (p-value not available).
Biological motion
Biological motion study with target or non-target as response and percentage correct and
number of tolerated noise dots as outcome.
[25]
Freire et al (2006)
Controls
a
6 year olds 9 year olds Statistics
n 24 24
(SD) (6.0; 0.25) (9.0; 0.25)
IQ information -
Biological motion
Correct identification
Mean percentage correct % 95 94
(SD; range) (-; 90-100) (-; 86-100)
b
Mean number of noise dots 69.26 77.76 Age group: F(2, 69) =
c c
tolerated (12) (12) 6.00, p < .01
(SD) 6 year olds vs. adults:
p <.01
9 year olds vs. adults:
ns
a b c
Also an adult group was included in the study; One outliers was replaced by the group mean;
Data was extracted from a figure with group results.
Biological motion
Biological motion study with target or non-target as response and reaction time as
outcome.
[26]
Lichtensteiger et al (2008)
a
Controls Statistics
n 13
Gender (% male) 62
Age range in years -
(SD) (6.59; 1.2)
IQ information -
Biological motion Age group: F(1, 29) = 57.36,
p < .01. Children are slower
Target
than adults
Mean reaction time sec 1.36 (0.22)
(SD; range)
Non-target
Mean reaction time sec 1.50 (0.25)
(SD; range)
a
Also an adult group was included in the study.
References
ogy & Strabismus, 2007. 44(6): p. 363-371.
4A2
2596.
p. 389-404.
167.
Brain Res, 2005. 25(3): p. 788-98.
Science, 1992. 33(8): p. 2483-2489.

Neuropsychol, 2003. 23(1-2): p. 139-72.
Perception, 2006. 35(5): p. 647-57.
cal motion perception. Neuroreport, 2008. 19(18): p. 1763-1767.
Chapter
Chapter 5
5
Chronological age versus

developmental age in evaluating
patients’ performances on motion
perception tests
van der Zee, Y., Stiers, P., Lagae, L., Pel, J., & Evenhuis, H. (2018).
Chronological age versus developmental age in evaluating patients’
performances on motion perception tests. Accepted for publication
in Neuropsychological Trends.
100 Chapter 5
Abstract
In neuropsychological assessments, a patient’s raw score is frequently com-

pared to a large general population normative sample. It is common to use
the chronical age as entry of norm tables to assess a patient’s current cogni-
tive function. In individual patients with a developmental delay or cognitive
impairment this may result in misinterpretation of performance. The aim of
this study was to test the impact of chronological and developmental age
parameters on motion perception outcomes and to construct and evaluate
normal motion perception limits for clinical practice. In the present study,
the developmental age and four aspects of motion perception (biological mo-
tion, global motion, motion speed, motion-defined form) were assessed in 49
children with indications of brain damage and in 60 controls. Based on current
results we present the preliminary normal limits and we suggest the use the
developmental age as entry of norm tables.
Keywords: Motion perception assessment; Normal limits; Chronological age;

Developmental age; Brain damage
Assessing visual motion perception problems 101
Introduction
Congenital brain damage often leads to multiple disabilities, including impair-

ments in cognitive functioning. The presence of a cognitive impairment might
affect the performance on motion perception tasks. The visual perception of
motion is essential for navigating and interpreting a dynamically changing
visual environment. Valuable knowledge was obtained from a Patient (LM)
with selective impairments in visual motion perception. The patient reported
the inability to see other people and vehicles moving around, to see facial
movements and detect changes in the liquid level while pouring a cup of
[1, 2]
tea. Later studies on the neural mechanisms of visual motion perception
suggested that different aspects of visual motion are processed to a certain 5
extent in parallel, such as global motion, motion speed, motion-defined form,
[2-5]
body motion, and that such aspects can be selectively impaired. As a
result of these different aspects of motion perception, humans are able to pro-
cess multiple local motion signals, e.g. to perceive and distinguish directional
movement of dots in a display with random moving dots (global motion), to
discriminate speed differences (motion speed), to recognize objects of fictive
forms only through motion (motion-defined form) and to distinguish between
humans, animals and their activities through their specific motion patterns
only (biological motion).
Some studies on motion perception do control for an intellectual disability
and/or developmental delay by matching on verbal and/or non-verbal IQ-
[6, 7]
test performances, whereas others do not make a choice and report the
performances of patients in relation to chronological age and (verbal) mental
[8, 9]
age.
When controlling for mental age, normal performance levels can be found
[7]
for biological motion in intellectually disabled individuals and children with
[6] [8]
Williams Syndrome, and for global motion in autism and Williams Syn-
[6]
drome, while the performances were still weak for motion-defined form in
[6]
Williams Syndrome and for global motion in extremely prematurely born
[10]
children, mainly without clear signs of brain damage.
Since most controls were carried out on a group level, it remains unknown
what the effect of the control for mental age is on the evaluation of individual
[9]
results. A study by Atkinson et al suggests that a significant reduction of
50-67% in weak performances can be expected for global motion (from 19/45
to 8/45) and motion-defined form (from 18/45 to 6/45). Several studies in
children with early brain damage suggest that non-verbal/performance IQ is
[11]
more affected than verbal IQ. Therefore, their global or verbal cognitive
102 Chapter 5
level of functioning might not be the best predictor or control factor for their
performance level on a given task.
It might be a crucial factor to select the appropriate cognitive level, verbal
or non-verbal, when evaluating the performances of specific neuropsycho-
[12]
logical tests in multiply disabled patients. A study in Fragile X Syndrome
suggests that specific verbal and non-verbal working memory problems can
be detected if the verbal and non-verbal age equivalents, based on Verbal
IQ and Performance IQ results, are used as input for norm tables of verbal
tests and non-verbal tests, respectively. Specific impairments in dorsal stream
[13, 14]
functions, such as object recognition in suboptimal conditions can be
found in children with early brain damage by using the non-verbal mental age
as entry of the present norm tables. The use of chronological age masks these
specific neuropsychological deficits by a profile of weaknesses in several func-
[12, 14]
tions and thereby dramatically increases the number of impairments
[14]
in patients. This suggests that outcomes of dorsal stream function tests,
like motion perception test results obtained in children with congenital brain
damage could be either judged against chronological or non-verbal develop-
mental age. The aim of this study was to contribute to the discussion what age
parameter should be considered as entry of the norm tables when assessing
individual patients in clinical practice.
We examined four aspects of visual motion perception in a group of typically
developing children and children with brain damage: global motion, motion
speed, biological motion, motion-defined form. In this article, we focused on
the relation between age parameters (chronological age and developmental
age, i.e. the median age equivalent based on raw scores of nonverbal intel-
ligence subtests) and motion perception outcomes. We hypothesized that
motion perception scores are closely related to the non-verbal cognitive
level. In typically developing children, both measures are reflected in their
chronological age, suggesting that a clinician would be free to choose between
chronological age and performance age in the construction of norm values.
In children with brain damage, on the other hand, non-verbal cognitive level
and chronological age are dissociated. For the children with brain damage, we
hypothesized that motion perception performance is related to developmental
age rather than to chronological age. Here, we set and evaluated preliminary
norm values and considered the possible impact of the choice between the use
of chronological age or performance age as the entry of the norm table in the
evaluation of the performance of individual patients.
Methods
Participants
The patient group consisted of 49 children (24 boys, 25 girls). The patients
were recruited through rehabilitation centers in the Rotterdam area (Rijn-
dam Rehabilitation Centre and Royal Dutch Visio, n = 19) and the Leuven
University Hospital, Belgium (n = 30). Forty patients had abnormal imaging
results and nine had normal or no imaging results but had indications of brain
damage/dysfunction. The aetiology of brain damage was brain malformations
in three children, hypoxic-ischemic encephalopathy in twenty-three cases (19
periventricular leukomalacia (PVL), 3 intraventricular haemorrhage (IVH),
1 PVL + IVH), perinatal asphyxia in five, intracranial haemorrhage in two, 5
hydrocephalus in one and acquired brain injury in six (4 trauma, 1 meningitis,
1 tumour). Of the nine patients in whom no or normal imaging results were
present, three had a genetic disorder (Velo-Cardio-Facial syndrome; Beckwith
Wiedemann syndrome; 46XY + m), four had neurological signs such as cere-
bral palsy, one had visual problems not explained by ocular abnormalities and
one was dysmature probably due to prenatal drug exposure. Five patients had
ocular abnormalities other than refractive errors or oculomotor dysfunctions.
No patient had ophthalmological abnormalities to such a degree that it would
[15]
interfere with perceiving details of the motion stimuli. At the moment of
motion perception testing, chronological age ranged from 4.11 to 14.58 (M =
7.35, SD = 2.26 years).
The control group consisted of 119 typically developing children (54 boys,
65 girls) with no indication of neurological or visual impairments and normal
or corrected to normal visual acuity. Controls were recruited through primary
schools in the Netherlands (n = 79) and Belgium (n = 40). At the moment of
motion perception testing, their chronological age ranged from 3.50 to 7.86
years (M = 5.47, SD = 1.05 years), which was significantly lower than that of
the control group (U(166) = 1141.5, Z = -6.19, p < .01).
Studies were approved by the Ethics Committees of the Erasmus Medical
Center and the Leuven University Hospital. For all participants informed con-
sent was obtained from their parents or guardians.
Procedures
All Dutch and Belgian controls (n = 119), and the Dutch patients (n = 19)
were tested at the children’s primary schools. The Belgian patient group (n
= 30) was studied at the Leuven University Hospital. In the Dutch groups,
motion perception tasks were presented in a fixed order: biological motion,
104 Chapter 5
motion-defined form, global motion and motion speed. In a subgroup of Dutch

controls motion speed was not administered due to time constraints (n = 19).
In the Belgian control and patient groups tasks were administered in random
order.
Task administration was done by trained senior psychology students or neu-
ropsychologists. Tasks were presented on a 15-inch CRT monitor attached a
laptop. Participants were placed in front of the screen at approximately 40 cm.
Motion Perception tasks

Four different motion perception tasks were administered, covering the do-
mains of global motion, motion speed, biological motion and motion-defined
form (Figure 1A-D). All stimuli consisted of white dots on a black background,
with a resolution of 640 x 480 and refresh rate 25 frames/s. In the global
motion task, biological motion task and the motion speed task psychophysical
thresholds were estimated by calculating the mean of the values of the last
4 of 8 reversals, using a 2up-1down staircase procedure. In these tasks, a
correct answer was followed by a beep.
Figure 1. Schematic examples of mo-

tion perception tasks. The arrows indi-
cate the motion direction of the dots,
length of the area motion speed (longer
is higher speed). In the real task, the
background is black and dots are white
and the borders are not defined by lines.
A. M
otion coherence task with target
area on the right;
B. M
otion-defined form task, example
item square;
C. Biological motion: the single walker,
walking to the right. scrambled figure
and distracter dots are not shown;
D. M otion speed: dots in left car move
faster.
The global motion stimulus consisted of two random dot kinematograms

(size 14.7 x 22.4 deg) containing 1103 white dots (dot size 0.07 deg, limited
life time 130 ms), presented next to one another with a distance between
them (size 3.3 deg). A variable proportion of dots (starting level 100%, scaling
factor 0.33) in each kinematogram oscillated coherently in horizontal direc-
tion (reversal time 330 ms, velocity 6.7 deg/s). Participants had to locate a
horizontal strip (size 14.7 x 7.5 deg) in the middle of one of the random dot
kinematograms, where the coherent dots oscillated in the opposite direction.
Because the proportion of coherent dots was constant throughout the random
dot kinematograms, the strip could not be located by tracing the movement of
single dots. The proportion of coherently moving dots, or the coherence level
determined the difficulty of the task and was used to calculate the coherence 5
threshold. Participants were instructed to help a lost person to find his way in
the snow (presentation = 15 s, answer time 5 s).
The motion-defined form stimuli consisted of objects hidden in a random
dot kinematogram (size 20.6 x 16.0 deg, 5000 dots, dot size 0.13 deg, life
time 200 ms, velocity 3.4 deg/s). Each object could be displayed in three
successive conditions with decreasing level of difficulty (presentation max. 15
s). In all conditions, the dots outside the contour moved coherently in oblique
direction. In the first condition, the dots in the contour of the object moved
coherently downwards. In the second condition, the dots in the contour were
standing still, and in the third condition there were no dots in the contour.
After an object was correctly identified the trial was aborted and the next trial,
with a new object, was started. If the object was correctly named or described
in the first, second or third condition a score of 1, 0.5 or 0 was noted. If the
object was not correctly identified in the third condition the response was
marked as inconclusive, and the item was not used in the computation of the
visual motion perception score. Three subtasks, increasing in difficulty, with
six objects were presented. Objects in task 1 were: circle, star, bear, banana,
heart and fish; task 2: arrow, kangaroo, boat, guitar, ostrich and bag; task 3:
beetle, seat, airplane, seahorse, car and shoe.
The biological motion stimulus consisted of a human-point light walker and
a phase-scrambled point-light figure (both 11 dots, dot size 0.13 deg, dot
lifetime 40 ms, stimulus height approx. 11 deg). The location of the human
point-light walker (left or right of the screen), its position within the field
and its walking direction (left or right) were randomized. Participants had
to indicate the location of the human point-light walker (presentation max.
20 s, walking speed 48 cycl/min). Difficulty was increased by adding noise
106 Chapter 5
dots (starting level 1 dot added, scaling factor 0.075 and 0.150 from second
reversal), and the threshold was the critical number of noise dots added.
The motion speed stimulus consisted of two identical contours of a car (car
length approx. 17 deg) filled with leftwards moving dots (dot density 11 dots/
2
deg , dot size 0.07 deg, lifetime 120 ms). Participants were asked to indicate
the location of the fastest car (presentation time 10 s). A decrease in the
speed difference of the dots in the cars made the task more difficult (starting
speed difference 17.0 deg/s, scaling factor 0.33, 0.25 from fifth reversal) and
the critical speed difference was the score for this task.
Before each task, example stimuli were used to familiarize participants with
task elements and verify that they understood the task.
Developmental age
Because previous studies suggest that non-verbal cognitive ability, and not
[16-20]
verbal cognitive skill, is predictive of perceptual performance we only
studied non-verbal intelligence in addition to visual perception in patients and
in a subset of controls. The data collection of IQ data in the control group was
limited due to time constraints. Although the use of a single intelligence test
is preferable, the broad age range in the patient group and the cognitive con-
sequences of the brain damage made this impossible. In addition, we decided
to use recent intelligence results when available to keep the required effort of
the patients as low as possible.
Non-verbal intelligence data were collected in 60 Dutch controls, using the
Snijders-Oomen non-verbal intelligence test (SON-R 2½ - 7). In the patient
group, data were available for SON-R in 19, Wechsler preschool and primary
scales of intelligence (WPPSI-R) in 24, Wechsler Intelligence Scale for Chil-
dren III (WISC-III) in 5 and Wechsler Intelligence Scale for Children-revised
(WISC-R) in 1 patient. All these tests have normative data for the Dutch
speaking population of Belgium and The Netherlands.
The SON-R differs from the Wechsler tests in that the SON-R can be ad-
ministered without using verbal instructions, which makes the assessment in
children with hearing or language problems possible, and feedback is provided
after each item, which gives the child the possibility to learn. The effect of
these differences on the outcomes is unknown, but because the correlation
[21] [22]
between SON-R IQ and WPSSI-R PIQ is 0.93 and WISC-R PIQ is 0.79,
we considered these tests interchangeable for the developmental age estima-
tion.
The developmental age was defined as the median age equivalent based
on the raw subtest scores of the non-verbal intelligence scale. We used age
equivalent tables published in the manuals of the different IQ tests to convert

each raw subtest score to an age equivalent and then calculated the median
age equivalent, see Table 1 for examples of the WPPS-R results. Because there
was a time lag (M = 1.85, SD = 2.39 months) between the assessment of
non-verbal intelligence and motion perception, the estimated developmental
age was extrapolated from the time of intelligence assessment to the time of
motion perception assessment.
Statistical analysis
We used IBM SPSS Statistics version 20, the data of the 60 Dutch controls and 5
all patients and the Mann-Whitney test to study group performances. Because
we assumed that patients performed equally well or worse than controls we
used the one-tailed significance and a p-value of .05.
To study the relation between the outcomes of the motion perception tasks
and the age parameters chronological age and developmental age we calculated
the Spearman correlation and partial correlation. Outcomes of these analysis
in controls were used to decide whether chronological age or developmental
age should be used to set the normal limits. We ordered the performances of
the controls from worst to best and computed the weighted average of the
th th th th th th
5 , 10 , 25 , 50 and 75 percentile. We set the 5 percentile score as the
th
cut-off value for abnormal performance and the 10 percentile score as the
cut-off value for weak performance.
We then calculated the 95%-confidence interval for the found cut-off values
based on 1000 bootstrap samples, with replacement from the original data-
set. In addition we calculated the 95%-confidence interval for the percentage
[23]
excluded participants with the exact binomial method of Clopper-Pearson.
Results
Motion perception performance in relation to chronological

and developmental age
Data were present on chronological age, PIQ and developmental age in the 60
Dutch controls, 25 boys and 35 girls. In the control group, the mean chrono-
logical age was 5.67 years (SD = 0.82), the mean PIQ 105 (SD = 14), and the
mean developmental age 5.85 years (SD = 1.10). As expected, the Spearman
correlation was high between chronological age and developmental age (rs =
108 Chapter 5
.72, p < .01, df = 59) and absent between chronological age and PIQ (rs =
- .01, ns, df = 59), showing that mean PIQ is similar across age. In eight con-
trols one or two motion perception tasks were not completed due to reduced
compliance. With respect to the biological motion task it was remarkable that
several children mentioned that the phase-scrambled point-light figure looked
like a dancing person when it was introduced.
Table 1. Calculating developmental age in months from non-verbal IQ subtest age equivalents of
the WPPSI-R in 4 different Belgian patients (BE04; BE02; BE11; BE07). * middle two values, values
around the median.
BE04 BE02 BE11 BE07
(Raw test scores)
WPPSI-R subtests Age equivalents in months according manual
Object Assembly (28) 88* (24) 66* (20) 54* (10) 33
Geometric designs (47) 68 (48) 69 (34) 57* (15) 41
Block design (20) 57 (18) 54 (10) 43 (8) 39*
Mazes (19) 75* (17) 67* (8) 42 (9) 41
Picture completion (23) 92 (18) 58 (22) 83 (10) 40*
Animal Pegs (58) 96 (56) 84 (58) 96 (50) 67
Median age equivalent 81 66 56 40
Age IQ test 73.13 73.86 93.96 91.73
Age motion tests 73.13 73.86 96.89 103.72
a
Developmental age 81 66 57.74 45.23
PIQ 102 83 62 -
VIQ 106 83 102 -
a
Developmental age = (Median age equivalent/Age IQ test)*Age motion test.
In the patient group one girl and one boy did not complete any of the motion
perception tasks. The girl complained that she did not see anything moving.
Ten patients completed all four tasks, 24 three, 7 two, and 6 one. Chronologi-
cal age of these remaining 47 patients was 7.32 years (SD = 2.28), mean PIQ
78 (SD = 20, n = 40), and mean developmental age 5.38 years (SD = 1.42).
The correlation between chronological and developmental age was modest (rs
= .41, p < .01), but significantly lower than in the control group (.41 versus
.72, p < .01). The relation between chronological age and PIQ was absent (rs
= - .07, ns).
In Figure 2 the distribution of scores on the motion perception tasks in
patients and controls are plotted. The boxplots and measures for skewness
and kurtosis show that the data of most tasks are not normally distributed,
except for the data of the patient group for the global motion and biological
motion tasks. The distribution for the global motion task in the controls is near-
normal. The non-normal distributions are skewed and sometimes leptokurtic
(clustered about the center and thinner tails, except at the end points where
the tails are thicker than in the normal distribution). The motion coherence
level in patients (Mdn = .45) did not differ significantly from controls (Mdn
= .40; U = 1416.50, z = 1.56, p = .06, effect size r = 0.16). However, the
patients (Mdn = .72) performed significantly worse than the controls (Mdn =
.84) on the motion-defined form task (U = 861.50, z= -1.91, p = .03, effect
size r = -0.19). The performance on the biological motion task did not differ
significantly (Mdnpatients = 8.12, Mdncontrols = 7.87; U = 813.50, z = -0.24,
ns, effect size r = 0.03). Unexpectedly patients performed significantly better
on the motion speed task (Mdnpatients = 4.28 deg/s, Mdncontrols = 5.78; U =
507.50, z = -1.66, p = .05, effect size r = 0.19). The data of the patient group
is more clustered and shows 1 outlier. 5
Figure 2. Motion perception performances in controls and patients. Boxplots: horizontal line is the
median, the box represents the interquartile range (range between 25th and 75th percentile), the
wishers represent the 95%-confidence interval, the circles are data points outside the confidence
interval and the asterisk is an outlier in that group A. Motion coherence task; B. Motion-defined form
task; C. Biological motion; D. Motion speed.
Table 2 and 3 present the relation between motion perception outcomes

and age parameters in controls and patients. The signs of the correlations
indicate that on all motion perception tasks the performance tends to improve
110 Chapter 5
with increasing chronological and developmental age, except for biological

motion in the patient group, for which performance tends to deteriorate with
increasing chronological age.
In the control group (Table 2), all correlations of motion perception perfor-
mance with chronological age and with developmental age were significant.
Similar results for chronological and developmental age in the control group
are not surprising, given their strong intercorrelation. Therefore, we also stud-
ied the unique contribution of each age parameter to motion perception using
the partial correlation. Chronological age explained some unique variance in
the global motion scores (p = .02), whereas developmental age explained
some unique variance in the motion speed scores (p = .03).
Table 2. Controls: relation between motion perception task outcomes, chronological age and de-
velopmental age, as well as unique contribution of chronological age and developmental age to task
outcomes. * p ≤ .05 ** p ≤ .01
Spearman correlation Spearman partial correlation
Chronological Developmental Chronological Developmental
a b
n age age age age
Global motion 57 - .43 ** - .35 ** - .27 * - .07
Motion-defined form 59 .39 * .40 * .16 .19
Biological motion 60 .31 * .27 * .17 .08
Motion Speed 53 - .34 * - .42 ** - .07 - .27 *
a b
controlled for developmental age; controlled for developmental age
In the patient group (Table 3), correlations of motion perception perfor-

mance with chronological age were non-significant. In contrast, the correla-
tions with developmental age were significant and similar in magnitude to
those found in controls. The partial correlations were also mainly significant
for developmental age.
Table 3. Patients: relation between motion perception task outcomes, chronological age and devel-
opmental age, as well as unique contribution of chronological age and developmental age to task
outcomes. * p ≤ .05 ** p ≤ .01
Spearman correlation Spearman partial correlation
Chronological Developmental Chronological Developmental
a b
n age age age age
Global motion 42 - .17 - .42 ** .00 - .39 **
Motion-defined form 38 .01 .32 * - .14 .34 *
Biological motion 28 - .12 .48 ** - .39 * .59 **
Motion Speed 25 - .18 - .33 * - .05 - .29
a b
controlled for developmental age; controlled for developmental age
Normal limits
The results above show that in typically developing children the unique contri-
bution of performance age over chronological age in visual motion perception
performance is limited. Therefore, we had no reasons to deviate from common
practice and calculated the normal limits based on controls’ chronological age.
As a result, we could use data of the entire control group (N = 119), including
the controls without IQ data.
The control groups performed equally on the global motion task (Mdn = .40
vs .37) and the biological motion task (Mdn = 7.87 vs 9.42, effect size r =
0.14). The control group with IQ data (Mdn = .84) performed significantly bet-
ter than the control group without IQ data (Mdn = .77) on the motion-defined
task (U = 2222.50, z = 2.79, p < .01, effect size r = 0.26) and on the motion 5
speed task (Mdn = 5.78 vs. 9.40 deg/s; U = 635.50, z = -2.83, p < .01, effect
size r = 0.30).
In Figure 3 the motion perception scores obtained in the group with and
without IQ information are plotted relative to age. As the chronological age in
the group without IQ data was slightly but significantly lower than in the group
with IQ data (5.29 ± 1.21 years vs. 5.66 ± 0.81 years; t(101.5) = 1.99, p =
.05), there is a slight shift of the groups relative to the X-axis. This explains
the differences in performance between the control groups.
To construct age-specific cut-off criteria, the entire sample was optimally
divided in three age groups with a minimum of 30 participants per group. The
youngest group, group 1, consisted of 19 boys and 12 girls younger than 4.75
years with a mean age of 4.20 years (SD = 0.39; range = 3.50 - 4.67). Group
2 consisted of 17 boys and 26 girls with an age between 4.75 and 5.75 years
with a mean age of 5.26 years (SD = 0.27; range = 4.75 - 5.74). Group 3, the
oldest group, consisted of 18 boys and 27 girls aged 5.75 years and older with
a mean age of 6.56 years (SD = 0.60; range 5.75 - 7.86).
The data of the global motion task was normally distributed in group 1 and
2. Skewness and kurtosis were within normality ranges in group 3, but the
Shapiro-Wilks was significant (p = .05), suggesting a non-normal distribution.
The data of the motion-defined form task was normally distributed for group
1 and 3. The data of the group 2 was skewed and the Shapiro-Wilks was near
significant (p = .06). The data of the motion speed task had a non-normal
distribution in all groups (Shapiro-Wilks ps < .01). In group 1 skewness and
kurtosis were within normality ranges, the data in group 2 was skewed and the
data of group 3 was more skewed, looked more like a lognormal distribution,
but was also leptokurtic. Because of the different shapes of the distributions
we decided not to transform and normalize the data.
112 Chapter 5
Figure 3. Performance of the control group with IQ (n = 60) information and the control group
without IQ information (n = 59) on A. global motion, B. motion-defined form, C. Biological motion,
D. motion speed
For the global motion and motion speed task the 10th and 5th percentile
cut-off values decreased by age. In the youngest group (n1 = 31) the global
motion cut-off coherence levels were 0.78 and 0.80, in the second group (n2
= 39) 0.69 and 0.74, and in the oldest group (n3 = 45) 0.46 and 0.56. 10th
percentile cut-off values for the motion speed differences were 23.80, 20.00
and 12.49 deg/s and 5th percentiles cut-off values were 23.80, 21.53 and
19.87 (n1 = 25; n2 = 34; n3 = 31). For the motion-defined form task the
proportion correct cut-off values increased by age from 0.45 in group 1 (n1
= 31), 0.63 in group 2 (n2 = 43) to 0.74 in group 3 (n3 = 43) for the 10th
th th
percentile and from 0.33, 0.59 to 0.70 for the 5 percentile. The 5 percentile
of the biological motion task coincided with the lowest score obtainable in all
age groups, implying a bottom effect for this task. Therefore, presentation of
cut-off values and clinical evaluation on this task is not useful.
Figure 4. Confidence intervals for cut-off values for percentile 5 (left) and percentile 10 (right).
Filled circle: median score; filled black triangle: cut-off value; grey triangles: confidence limits. The
lines are fitted to illustrate possible trends. A. global motion, B. motion-defined form, C. motion
speed.
Confidence intervals for task cut-off values and for the

percentages of participants scoring below the 5th/10th
percentile
To illustrate possible developmental trends, we fitted trendlines to the median
th th
scores of different age groups, cut-off values for the 5 and 10 percentile,
and their confidence intervals in Figure 4. Because bootstrapping relies on
the observed data, the upper bound (if data is ordered from worst to best
performance) of the confidence intervals (illustrated with open grey triangles)
th
coincided with the worst performance observed in each group for the 5
th
percentile, and some, but not all groups, for the 10 percentile. For the re-
th
maining groups, the outcomes were near the 5 percentile score. The lower
bound of the confidence intervals (illustrated with filled grey triangles) for the
th th
5 percentile was near the outcomes of the 10 percentile, and that for the
th th
10 percentile coincided with or was near the outcomes for the 25 percentile
in all groups for all tasks. As a result, the changes in the distributions over
114 Chapter 5
age, most notable for the motion speed task, are reflected in our plots: the
increasing variance and skewness in the motion speed task resulted in a wider
confidence interval, although group size changed from 25 to 34 and 31.
The exact binomial method of Clopper-Pearson showed that we might label
th
1% to 17-21% of the population as abnormal when using the 5 percentile
cut-off values of the global motion task and 2-4% to 24-26% as weak when
th
using the 10 percentile cut-off values. The results for the motion-defined
form are comparable. For the motion speed task, we might label 1% to 20-
26% as abnormal and 2-3% to 26-31% as weak when using the cut-off values
th th
for the 5 and 10 percentile.
Discussion
This study shows that a child’s level of non-verbal cognitive ability is indicative
of the level of performance on a visual motion perception task. This non-
verbal cognitive level is not the relative level in comparison to peers, but the
absolute level of ability as reflected in the non-verbal cognitive complexity of
the tasks that the child is able to perform. In typically developing children this
absolute level of non-verbal cognitive ability is directly related to the child’s
chronological age, which is a good proxy for it. In children with early brain
damage, on the other hand, the level of non-verbal ability is often dissociated
[16, 24-27]
from chronological age. Our results show that in that case the non-
verbal cognitive level is predictive of visual motion perception performance.
This is in line with previous studies on object and form perception abilities in
[16, 18, 28]
neuropediatric populations, such as cerebral palsy, hypoxic-ischemic
[16]
encephalopathy in premature infants and in birth asphyxia and spina
[19]
bifida. The high correlations between PIQ and object or form perception
scores reported, ranging from 0.33 to 0.85, imply that a lower perception
score than expected from a child’s chronological age is not as such an indica-
tion of a perceptual disability. This implies that, in general, the scores on visual
motion perception tasks reflect a patient’s global non-verbal cognitive level,
in addition to a possible specific visual motion perception disability. Therefore,
to evaluate perceptual ability in these children it is important to use their
global non-verbal cognitive level, as expressed by the developmental age, as
a baseline. Only in this way, it can be avoided to erroneously interpret non-
verbal cognitive impairment manifested in the motion task performance as a
[16]
motion perception impairment. Stiers, De Cock and Vandenbussche (see
[14] [15]
also Stiers et al. , Stiers & Fazzi ) suggest a simple and clinically feasible
approach. In this approach, the scores on a set of non-verbal intelligence

subtest is used to estimate the overall non-verbal performance level of a child.
This performance level, expressed as an age equivalent, is used instead of
chronological age to relate the child’s visual motion score to the scores of the
normative group.
Given that normative data for tests are usually based on participants’ chron-
ological age, the results in our study provide no reason for deviating from this
practice; the gains of collecting IQ information in typically developing children
seem to be minimal. We therefore state that motion perception performances
in different chronological age groups of normally developing children should
be used to set norm values for motion perception tasks.
A number of limitations to our study must be addressed. Four different 5
motion perception tasks were included in our study. Unfortunately, the bio-
logical motion paradigm was too difficult or even invalid for the age range
of the typically developing children. The remarks of several children that the
phase-scrambled point-light figure looked like a dancing person, suggests that
top-down processes like expectations might have influenced the results and
makes the construct validity of the task questionable.
It should be noted that the reliability of the estimated incidence of visual
motion perception impairment will depend on the reliability of the statistical
cut-offs used to define impairment. In the present study, the 95%-CI for the
chosen cut-off scores were rather large, e.g. 1-16% for the 5th percentile
in a group with a sample size of 43 and 1-26% in a group with a sample
size of 25. Using currently applied cut-off values might therefore result in
the overdiagnosis of motion perception abnormalities and weaknesses. Our
bootstrap overdiagnosis results also showed that our cut-off values are not
that precise yet. In addition, the results of the motion speed task showed that
the distribution of data influenced the width of the confidence interval. With
the increase in age the distribution looked more like a lognormal distribution.
To set normal limits more precisely, larger samples of typically developing
children are needed. We advise to study the developmental trends and dis-
tribution in different age groups and age ranges before setting the percentile
normal limit. Especially, in non-linear developmental trends with more steep
development, smaller age ranges per norm group should be considered. Ad-
ditionally, transformation of data might be needed to get more precise normal
limits. Even if lognormally distributed data is present, it might be advanta-
[29]
geous to perform a data transformation.
Finally, before clinical implementation can take place, we not only need accu-
rate norm values, but also should evaluate the sensitivity and specificity of the
116 Chapter 5
tasks and assure ourselves that we assess clinically meaningful weaknesses.

The question is whether children with weak motion perception show more
problems in dynamic daily activities, such as pouring a drink, sport and traffic
participation and social interactions. Although effects of weak motion percep-
tion might be less debilitating in developing children than reported by patient
LM as mentioned in the introduction, studies in children and adults suggest
that the ability to perceive motion might indeed influence daily functioning in
several ways. The middle temporal area (MT) is essential in the perception
of global motion, the segregation of the moving object from its background
[30]
and fixating and following the object. Smooth pursuit and global motion
perception seem essential in controlling the water level while pouring a cup
of tea and tracking a ball and/or predicting the ball’s course before hitting or
[31]
catching it. Studies in adults demonstrate that performances on a motion-
[32] [33]
defined and a 3D speed discrimination task and a global motion task
predict car driving performances. It was also shown that motion perception
[32]
training can improve traffic performances. A study on biological motion
[34]
perception and autism shows that superior temporal sulcus (STS) is not
only selectively activated by biological motion stimuli, but its activity is also
modulated by the perception of other people’s intentions. Children with autism
seem less able to distinguish biological and non-biological motion patterns.
The authors suggested that this might have impaired their development of the
theory-of-mind, the ability to interpret and predict other people’s actions and
[35]
intentions correctly, and social skills. Further imaging research shows that
activity in the brain network involved in biological motion perception predicts
the effectiveness of treatment in children with autism, i.e. higher activity
levels were associated with higher treatment effectiveness. Overall, the ability
to perceive motion seems to have significant impact on daily functioning. Still,
not many studies have investigated motion perception problems in dynamic
daily activities. It remains a topic for important future research especially in
children when they have reached school age. For example, are children that
perform worse on a global motion task also slower or clumsier in activities of
daily life or do they perform worse on (fast) ball sports in school? Or does a
worse motion-defined form or global motion task performance relate to the
ability to cross a street safely? Lastly, do children with brain damage that
perform weak on motion perception tasks also have more problems with social
interactions? Additional studies on motion perception performances and daily
life functioning are necessary to answer these questions.
Acknowledgements
We thank the participating children and their parents, the schools (Eduard
van Beinum, Openbare basisschool Charlois, De Bergse Zonnebloem, Visio-
school); Rijndam Revalidatiecentrum, Rijndam Rehabilitation Centre; Royal
Dutch Visio, Centre of expertise for blind and partially sighted people, all
located in Rotterdam, the Netherlands for their cooperation in the data col-
lection.
Funding: This study was supported by the K.U.Leuven Research Fund

(K.U.Leuven Onderzoeksfonds) grants nr. OT/01/43, PDM/01/156 and
PDM/03/251 and Royal Dutch Visio. 5
118 Chapter 5
References
1. Zihl, J., D. Von Cramon, and N. Mai, Selective disturbance of movement vision
after bilateral brain damage. Brain, 1983. 106: p. 313-340.
blind” patient, assessed with limited-lifetime random dot stimuli. The Journal of
Neuroscience, 1991. 11(2): p. 454-61.
3. Cowey, A. and L.M. Vaina, Blindness to form from motion despite intact static
form perception and motion detection. Neuropsychologia, 2000. 38(5): p. 566-
78.
4. McLeod, P., Preserved and Impaired Detection of Structure From Motion by a
‘Motion-blind” Patient. Visual Cognition, 1996. 3(4): p. 363 - 392.
5. Vaina, L.M., et al., Intact “biological motion” and “structure from motion” percep-
tion in a patient with impaired motion mechanisms: a case study. Visual Neurosci-
ence, 1990. 5(4): p. 353-69.
10. Benassi, M., et al., Motion Perception and Form Discrimination in Extremely Pre-
term School-Aged Children. Child Dev, 2017.
11. Murias, K., et al., A review of cognitive outcomes in children following perinatal
stroke. Dev Neuropsychol, 2014. 39(2): p. 131-57.
12. Wouters, H., A. Fonteyne, and P. Stiers, Specific memory impairment in a multiple
disabled male with fragile X syndrome and temporal lobe epilepsy. Developmental
Medicine & Child Neurology, 2006. 48(5): p. 378-382.
13. Stiers, P. and E. Vandenbussche, The dissociation of perception and cognition in
children with early brain damage. Brain and Development, 2004. 26(2): p. 81-92.
14. Stiers, P., et al., The variety of visual perceptual impairments in pre-school
children with perinatal brain damage. Brain and Development, 2001. 23(5): p.
333-48.
15. Stiers, P. and E. Fazzi, Psychometric evaluation of higher visual disorders: strate-
gies for clinical settings, in Visual Impairment in Children Due to Damage to the
Brain. Clinics in Developmental Medicine, G.N. Dutton and M. Bax, Editors. 2010,
Mac Keith Press: London. p. 149-161.
16. Stiers, P., P. De Cock, and E. Vandenbussche, Separating visual perception and
non-verbal intelligence in children with early brain injury. Brain and Development,
1999. 21(6): p. 397-406.
17. Stiers, P., P. De Cock, and E. Vandenbussche, Impaired visual perceptual perfor-
mance on an object recognition task in children with cerebral visual impairment.
Neuropediatrics, 1998. 29(2): p. 80-8.
18. Ito, J., et al., Assessment of visuoperceptual disturbance in children with spastic
diplegia using measurements of the lateral ventricles on cerebral MRI. Develop-
mental Medicine & Child Neurology, 1996. 38(6): p. 496-502.
19. Ito, J., et al., Neuroradiological assessment of visuoperceptual disturbance in
children with spina bifida and hydrocephalus. Developmental Medicine & Child
Neurology, 1997. 39(6): p. 385-92.
20. Stiers, P., et al., Atypical neuropsychological profile in a boy with 22q11.2 Deletion
Syndrome. Child Neuropsychol, 2005. 11(1): p. 87-108.
21. Moore, C., S.L. O’Keefe, and D. Lawhon, Concurrent validity of the Snijders-
Oomen Nonverbal Intelligence Test 2 1/2-7--Revised with the Wechsler Preschool
and Primary Scale of Intelligence--Revised. Psychological Reports, 1998. 82(2).
22. Nieuwenhuys, M., A comparative study between Son-R, WISC-R and Raven-SPM.
[Een vergelijkingsonderzoek SON-R, WISC-R en Raven-SPM], in Ontwikkeling-
spsychologie/Poli Kinderpsychiatrie. 1991, University of Amsterdam 5
23. Newcombe, R.G., Two-sided confidence intervals for the single proportion: com-
parison of seven methods. Statistics in Medicine, 1998. 17: p. 857-872.
24. Fedrizzi, E., et al., The cognitive development of children born preterm and af-
fected by spastic diplegia. Brain and Development, 1993. 15(6): p. 428-32.
25. Bava, S., A.O. Ballantyne, and D.A. Trauner, Disparity of verbal and performance
IQ following early bilateral brain damage. Cognitive and Behavioral Neurology,
2005. 18(3): p. 163-70.
26. Ballantyne, A.O., K.M. Scarvie, and D.A. Trauner, Verbal and Performance Iq Pat-
terns in Children after Perinatal Stroke. Developmental Neuropsychology, 1994.
10(1): p. 39-50.
27. Abercrombie, M.L., et al., Visual, Perceptual and Visuomotor Impairment in Physi-
cally Handicapped Children. Perceptual and Motor Skills, 1964. 18: p. 561-625.
28. Koeda, T. and K. Takeshita, Visuo-perceptual impairment and cerebral lesions
in spastic diplegia with preterm birth. Brain and Development, 1992. 14(4): p.
239-44.
29. Schoonjans, F., D. De Bacquer, and P. Schmid, Estimation of population percen-
tiles. Epidemiology, 2011. 22(5): p. 750-1.
30. Born, R.T., et al., Segregation of object and background motion in visual area MT:
effects of microstimulation on eye movements. Neuron, 2000. 26(3): p. 725-34.
31. Land, M.F., Eye movements and the control of actions in everyday life. Prog Retin
Eye Res, 2006. 25(3): p. 296-324.
32. Wilkins, L., et al., Motion perception and driving: predicting performance through
testing and shortening braking reaction times through training. Invest Ophthalmol
Vis Sci, 2013. 54(13): p. 8364-74.
33. Wood, J.M., Age and visual impairment decrease driving performance as mea-
sured on a closed-road circuit. Hum Factors, 2002. 44(3): p. 482-94.
34. Pelphrey, K.A. and E.J. Carter, Brain mechanisms for social perception: lessons
from autism and typical development. Ann N Y Acad Sci, 2008. 1145: p. 283-99.
35. Yang, D., et al., Brain responses to biological motion predict treatment outcome
in young children with autism. Transl Psychiatry, 2016. 6(11): p. e948.
Chapter
Chapter 6
6
Clinical assessment of visual

motion perception in children
with brain damage and the risk of
overdiagnosis
van der Zee Y., Stiers P., Lagae L., et al. Clinical assessment of
visual motion perception in children with brain damage and the risk
of overdiagnosis. Submitted.
122 Chapter 6
Abstract
Aim
In this study, we examined 1) the presence of weaknesses in various visual
motion perception aspects in children with brain damage; 2) the clinical im-
pact of chronological age and developmental age as reference level when
interpreting motion perception performances.
Method
Performances of 46 children with indications of brain damage (Mage = 7y4m,
SD = 2y4m) on three visual motion perception aspects (global motion, motion
speed, motion-defined form) were evaluated. We used developmental and
chronological age as entry of a preliminary reference table, and compared
the proportions of weak performers with probabilities expected in the general
population.
Results
When using developmental age as reference level, the proportion of weak
performers on at least one of the three tasks was significantly higher than
expected in the general population (19/46, p = .02). In 15/19 patients only
one aspect of motion perception was affected. When using chronological age
as reference level, the proportion of patients that performed weak on at least
one task increased significantly from 19/46 to 24/46 (p = .03).
Interpretation
There is an increased risk of isolated motion perception problems in children
with brain damage, independent of their performance IQ. The use of chrono-
logical age as reference level increases the risk of overdiagnosis significantly.
Keywords: Motion perception; Global motion; Motion defined form; Motion

speed; Performance age; PIQ
Introduction
Studies in children with cerebral palsy or periventricular brain damage suggest

[1-3]
that various aspects of visual motion perception can be impaired. Cortical
areas that are critical to visual motion perception include areas V3a, MT/V5+
(encompasses V5, MST and other alleged motion areas), and the superior
[4] [5]
temporal sulcus (STS). V5+ is typically activated in global motion and
[6]
motion speed perception. The perception of motion-defined forms is as-
[7, 8]
sociated with activity in ventral as well as dorsal visual areas. Because
different studies addressed different aspects of motion perception in isolation,
it is currently not known whether various aspects of visual motion perception
are impaired in individual children with brain damage.
An important factor that should be taken into account when interpreting
motion perception performances is non-verbal functioning. Children with
congenital or acquired brain damage showed lower performal and verbal IQ
6
[9]
scores. A study on the relation between performance IQ and motion per-
[10]
ception outcomes in children with brain damage showed that non-verbal
cognitive intelligence partly explained visual motion perception performance
as measured with neuropsychological tasks. Thus, such task outcomes reflect
a patient’s global non-verbal cognitive level, in addition to a possible specific
visual motion perception disability. Currently, a limited number of studies in
motion perception controlled for intellectual disability and/or developmental
delay by matching individual patients to individual controls or matching patients
[11-13]
and controls on group level. In neuropsychological assessment, these
methods are not suitable. To uncover specific motion perception problems, i.e.
disentangle general effects of the established non-verbal cognitive impairment
[14]
from motion perception problems, Stiers et al have been suggesting an ap-
plicable method: the use of the developmental age, the median age equivalent
of (non)verbal intelligence subtests, as entry of the reference table.
The use of chronological age as reference level in neuropsychological as-
sessments might lead to a profile with several weaknesses and increase the
[14]
number of false positives, resulting in overdiagnosis. The extent of the risk
of overdiagnosis is currently unknown.
In this study, we evaluated whether children with brain damage have iso-
lated or multiple motion perception weaknesses by testing three aspects of
visual motion perception: global motion, motion speed, motion-defined form.
If these children would have an increased risk of motion perception problems,
we expected the proportion of weak performers to be significantly higher than
the probability in the general population, using their developmental age as
124 Chapter 6
reference level. We also studied the developmental-chronological age effect,

i.e. whether significantly more children are classified as weak performers
when chronological age is used as reference level.
Method
Participants
The patient group consisted of 46 children with indications of brain damage,
brain malformation, or clinical indication of visual perceptual impairment.
They were recruited through rehabilitation centres in the Rotterdam area (Ri-
jndam Rehabilitation Centre and Royal Dutch Visio) and the Leuven University
Hospital, Belgium. The chronological age ranged from 4y1m to 14y7m (M =
7y4m, SD = 2y4m). Performance IQ ranged from 45 to 119 (M =75, SD =
21), developmental age, the extrapolated median age equivalent on nonverbal
[14]
intelligence subtests, ranged from 2y5m to 8y2m (M = 5y4m, SD = 1y5m).
Twenty children had been born prematurely (gestational age <37 weeks). The
aetiology of brain damage was brain malformations in three children, hypoxic-
ischemic encephalopathy in twenty-one cases (18 periventricular leukomala-
cia, 3 intraventricular haemorrhage), perinatal asphyxia in five, intracranial
haemorrhage in one, hydrocephalus in one and acquired brain injury in six (4
trauma, 1 meningitis, 1 tumour). Of the nine patients in whom no or normal
imaging results were present, three had a genetic disorder (Velo-Cardio-Facial
syndrome; Beckwith Wiedemann syndrome; 46XY + m), five had neurological
signs such as cerebral palsy, one had visual problems not explained by ocular
abnormalities and one was dysmature probably due to prenatal drug expo-
sure. Five patients had ocular abnormalities other than refractive errors or
oculomotor dysfunctions, like nystagmus, saccadic dysfunction, convergence
abnormality and horizontal oculomotor apraxia. In twenty-two children, visual
acuity and/or visual field abnormalities were found. Eight had low vision (vi-
sual acuity between 0.1 and 0.3) and could therefore be considered visually
impaired and eight children had a subnormal visual acuity for their age (0.5
- 0.8). In seven children a slow or late response was found in one side of the
visual field, in one child a late response was found in the lower visual field,
two children had a concentric visual field loss, but one side was more affected
then the other and one child had a scotoma in the right visual field. None of
the patients had abnormalities to such a degree that it would interfere with
perceiving details of the motion stimuli.

Center (MEC-2006-056) and the Leuven University Hospital. For all partici-
pants informed consent was obtained from their parents or guardians.
Procedures
The Dutch patient group (n = 17) was tested at the children’s primary schools.
The Belgian patient group (n = 29) was studied at the Leuven University
Hospital. In the order.
random Dutch group, motion perception tasks were presented in the
order: motion-defined form, global motion and motion speed. In the Belgian
Task administration was done by trained senior psychology students or
group tasks were administered in random order.
Task administration was done by trained senior psychology students or neu-
ropsychologists. Tasks
Participants were
were presented
placed on aof15-inch
in front CRT monitor
the screen attached a 40 cm.
at approximately
laptop. Participants were placed in front of the screen at approximately 40 cm.
Figure
Figure 1. 1. Illustration
Illustration of motion
of motion perception
perception tasks. In thetasks.
real taskIn the real
borders are not
A. global
defined motion
by lines. task motion
A. global with target area
task with onarea
target theon the B.
left; left; B. motion-
square;
defined C.task,
form motion speed:
example dots inC.right
item square; motioncar move
speed: dotsfaster.
in right car move
faster.
Motion Perception
Motion tasks
Perception tasks
All stimuli (see Figure 1 for examples of the stimuli) consisted of white dots
on a black background, with a resolution of 640 x 480 and refresh rate 25
frames/s. In the global motion task and the motion speed task psychophysical
thresholds were estimated by calculating the mean of the values of the last
4 of 8 reversals, using a 2up-1down staircase procedure. In these tasks, a
correct answer was followed by a beep.
126 Chapter 6
The global motion stimulus consisted of two random dot kinematograms

(size 14.7 x 22.4 deg) containing 1103 white dots (dot size 0.07 deg, limited
life time 130 ms), presented next to one another with a distance between
them (size 3.3 deg). A variable proportion of dots (starting level 100%, scaling
factor 0.33) in each kinematogram oscillated coherently in horizontal direc-
tion (reversal time 330 ms, velocity 6.7 deg/s). Participants had to locate a
horizontal strip (size 14.7 x 7.5 deg) in the middle of one of the random dot
kinematograms, where the coherent dots oscillated in the opposite direction.
Because the proportion of coherent dots was constant throughout the random
dot kinematograms, the strip could not be located by tracing the movement of
single dots. The proportion of coherently moving dots, or the coherence level,
determined the difficulty of the task and was used to calculate the coherence
threshold. Participants were instructed to help a lost person to find his way in
the snow (presentation = 15 s, answer time 5 s).
The motion-defined form stimuli consisted of objects hidden in a random
dot kinematogram (size 20.6 x 16.0 deg, 5000 dots, dot size 0.13 deg, life
time 200 ms, velocity 3.4 deg/s). Each object could be displayed in three
successive conditions with decreasing level of difficulty (presentation max. 15
s). In all conditions, the dots outside the contour moved coherently in oblique
direction. In the first condition, the dots in the contour of the object moved
coherently downwards. In the second condition, the dots in the contour were
standing still, and in the third condition there were no dots in the contour.
After an object was correctly identified the trial was aborted and the next trial,
with a new object, was started. If the object was correctly named or described
in the first, second or third condition a score of 1, 0.5 or 0 was noted. If the
object was not correctly identified in the third condition the response was
marked as inconclusive, and the item was not used in the computation of the
visual motion perception score. Three subtasks, increasing in difficulty, with
six objects were presented.
The motion speed stimulus consisted of two identical contours of a car (car
length approx. 17 deg) filled with leftwards moving dots (dot density 11 dots/
2
deg , dot size 0.07 deg, lifetime 120 ms). Participants were asked to indicate
the location of the fastest car (presentation time 10 s). A decrease in the
speed difference of the dots in the cars made the task more difficult (starting
speed difference 17.0 deg/s, scaling factor 0.33, 0.25 from fifth reversal) and
the critical speed difference was the score for this task.
Before each task, example stimuli were used to familiarize participants with
task elements and verify that they understood the task.
Scoring task outcomes

[10]
We used preliminary reference data from Van der Zee et al to evaluate
individual patient performances, using their developmental age (if lower than
chronological age) and chronological age as entry for the reference table
(Table 1). If the developmental age was lower than 3y6m, task outcomes
were compared to results of the youngest reference group. If the age was
higher than 7y11m, task outcomes were compared to the results of the oldest
reference group.
th
Scores below the 10 percentile were defined as weak performances.
[10]
Table 1. Reference table, based on a study in 119 typically developing children. Percentiles for
th
global motion, motion speed and motion defined form task in different age groups. 10 percentile
(p10) outcomes were used as cut-off values for weak performance.
Global motion Motion speed Motion defined form
Coherence level Speed difference (deg/s) Proportion correct
age in years age in years age in years
3y6m 4y9m 5y9m- 3y6m 4y9m 5y9m- 3y6m 4y9m 5y9m-
6
-4y8m -5y9m 7y10m -4y8m -5y9m 7y10m -4y8m -5y9m 7y10m
n 31 39 45 25 34 31 31 43 43
best 0.19 0.18 0.10 1.85 2.43 0.78 0.93 0.97 1.00
p75 0.34 0.37 0.27 7.14 4.63 2.89 0.78 0.89 0.94
p50 0.46 0.43 0.32 13.74 6.31 4.40 0.64 0.82 0.89
p25 0.69 0.55 0.42 22.53 13.41 7.19 0.50 0.76 0.81
p10 0.78 0.69 0.46 23.80 20.00 12.49 0.45 0.63 0.74
p05 0.80 0.74 0.56 23.80 21.53 19.87 0.33 0.59 0.70
worst 0.83 0.74 0.64 23.80 22.53 20.96 0.28 0.54 0.64
95%-CI p10 2-26 3-25 4-24 * 3-31 * 3-27 * 2-26 2-26 4-25 * 4-25 *
* 95%-CI indicates the precision of the percentiles (range of general population that could be ex-
cluded using the cut-off values). For this sample, it was calculated for the nearest percentile above
p10 giving a whole number of participants excluded.
The binomial test was used to establish whether the proportion weak per-
formers was significantly higher than the probability in the general population
th
given the use of the 10 percentile cut-off in each task and the number of
tasks evaluated: i.e. higher than .27 (change at no weaknesses is .73) for
three tasks and .10 for a single task.
The McNemar test was used to study the developmental-chronological age
effect. A one-sided alpha ≤ .05 was considered significant.
128 Chapter 6
Results
Eight patients completed 1 task, seventeen 2 tasks and twenty-one 3 tasks.

Figure 2 shows the visual motion perception scores of the patient group rela-
tive to the sores in the reference sample. Nineteen out of forty-six patients
were classified as weak performers on at least one of three tasks. This was
significantly higher than expected in the general population (19/46 vs .27; p
= .02) and points towards an increased risk for motion perception problems
in the patient group, independent of their performance IQ. Fifteen patients
h
scored below the 10 percentile on a single task. Four patients performed
weakly on two out of three tasks, which is also higher than expected (4/46 vs
.03; p = .03).
All 25 patients that completed the motion speed task, which included only
one Dutch patient, performed normally. We classified ten patients as weak
performers on the global motion task and thirteen patients on the motion-
defined form task, which was significantly higher than expected on a single
task (10/42 vs .10 and 13/38 vs .10; ps < .01).
Mean developmental age of the nineteen patients with weak performances
on any of the tasks was slightly lower than that of patients without weak
performances (5y4m ± 1y10 years vs 5y6m ± 1y1m; t = 0.32, df = 26.9,
ns), because it included four children with a developmental age under 3 years.
Although a developmental age below 3 years did not always result in a weak
task performance (one patient performed normally on the global motion task),
excluding these patients reduced the proportion weak performers from 10/42
to 7/38 for the global motion task and from 13/38 to 10/35 for the motion-
defined form task. The latter was still significantly higher than expected (p <
.01).
If chronological age was used as reference level, the proportion weak per-
formers on at least one task increased significantly (19/46 to 24/46; p =
.03). One patient performed weak on all three tasks, the proportion of weak
performers on two tasks did increase from 4/46 to 9/46 (p = .03), but the
proportion of weak performers on one task did not change remarkably (15/46
vs 14/46; ns). The increase in weak performers was significant on global
motion (10/42 vs 18/42; p < .01), but not on motion-defined form (13/38 vs
16/38; p = .13).
Figure 2. Patients’ performances in relation to the trendlines for p05, p10 and p50 of the reference
Figure 2. Patients' performances in relation to the trendlines for p motion
group using developmental age as entry for reference table. A. global motion task; B. 05
, p10 speed
and p50
task; C.group
motionusing developmental
defined form task. age as entry for reference table. A. global motion task; B.
speed task; C. motion defined form task.
A
130 Chapter 6
Additionally, we explored the characteristics of the weak performers. There

was no clear pattern of increased risk for visual motion perception weak-
nesses associated with any of the aetiological categories of brain damage.
Weak performers were found in following aetiological categories: hypoxic-
ischemic encephalopathy (6/21); acquired brain damage (3/6); genetic (2/3);
unknown (3/6); other (4/10). There was a significant association between
weak performance on the visual motion perception tasks and the presence
of low vision (φ = 0.33, p = .03) or the presence of remarkable peripheral
visual field outcomes (φ = 0.37, p = .01). It should be noted that these visual
conditions were not sufficient to account for the reduced performance on the
visual perception tasks. The lowest visual acuity value measured was 0.17 and
one patient with this acuity performed normally on the motion defined form
task and motion speed task, while another patient with this acuity performed
normally on the global motion task.
Discussion
In this study on motion perception performances in 46 children with indica-

tions of brain damage, we showed an increased risk for isolated weaknesses
in visual motion perception, when we controlled for their non-verbal cognitive
level by using their median age equivalent of nonverbal intelligence subtests
[14]
as entry of the reference table, as proposed by Stiers et al. Incidences of
weaknesses were increased for the global motion task (10/42 = 0.24) and
the motion-defined form task (13/38 = 0.34), but not for the motion speed
task. Importantly, only a small proportion of the patients with impaired mo-
tion perception (4/19 = 0.21) were impaired on both tasks, indicating that
different neural networks are damaged in this patient group. We confirmed our
expectation that the use of chronological age as reference level significantly
increases the risk of overdiagnosis. Weak performances were not limited to
hypoxic-ischemic encephalopathy, but were also found in other aetiological
categories of brain damage.
The found incidences are lower than those reported in other studies of
neuropediatric populations with comparable cut-off criteria. This difference is
most likely due to the more rigorous control for non-verbal performance level
[15]
in the present study. Mackay et al, for instance, reported an incidence of
8/19 (42%) of impaired global motion scores in very low birth-weight children
with a verbal cognitive ability within the normal range, and age-matched con-
[1]
trols. Similarly, in the study by Jacobson et al, 21/43 (49%) children born
preterm and with mild periventricular brain damage scored below 2 standard
deviations from the mean in age-matched controls. However, their average
performance IQ was almost 20 points lower than that of the controls. It is
likely that these incidences, at least in part, also reflect reduced non-verbal
cognitive ability.
The use of developmental age as reference level not only reduced the num-
ber of diagnosed problems in groups and individual children, but it will also
help neuropsychologists to uncover specific problems and weighing the effect
of different neuropsychological factors on the child’s (dis)abilities. Determin-
ing and knowing a child’s relative strengths is very important, especially in
case of training or support.
It should be noted that the reliability of the preliminary reference cut-offs
is still limited. To set normal limits more precisely, larger samples of typically
developing children are needed. Another limitation of our study is that the
heterogeneity of the patient population does not allow to delineate specific
6
aetiological conditions of risk for visual motion perception disability. Also, evi-
dence for the clinical relevance, i.e. the relation between outcomes of these
motion perception tasks and problems in daily life, is still missing. Currently,
clinical relevance is based on publication of a single case study, that of pa-
[16]
tient LM. More extensive research on carefully selected patient samples is
needed. At this moment, motion perception tasks should only be applied as
additional observational tools in children with brain damage that have typical
problems, for example difficulties in traffic participation.
It is not clear why no weaknesses were observed on the motion speed task.
It might be a result of a bias in compliance due to the fixed order of presenta-
tion in the Dutch patient group, with motion speed last: only one patient
completed the motion speed task. To avoid this bias in future studies, tasks
should be administered in random order. Another possible explanation might
be that this was the only task that allows for a low-level comparison of stimu-
lus features, whereas the other tasks all require a higher level of integration
of the moving dots to find the correct answer. Unlike speed discrimination,
motion-defined form and global motion require the child to find or identify an
object from a limited amount of visual information. This relies on the function-
ing of more complex networks, the engagement of attention, search and hy-
pothesis testing operations that are associated with frontal-parietal networks.
[17, 18]
The integrated functioning of posterior visual and anterior executive
areas thrives on long-range connection fibres, which may be affected by early
[19]
brain damage or brain malformation. Additionally, performances on the
motion speed task might be less dependent on the integrity of primary visual
132 Chapter 6
cortex (V1) due to the higher dot speeds used, and might mainly depend on
the direct route from the retina to the superior colliculus and pulvinar to the
[20]
prestriate cortex. The common involvement of V1 in visual acuity, visual
[21] [20]
field and motion processing of the global motion and motion-defined
form task, might also explain why we find significant correlations between
these measures. We suggest, that at least a low-speed task (< 6 deg/s) which
activates V1 before V5, and a high-speed task (> 15 deg/s) which activates
the colliculo-prestriate cortical route, should be developed and studied for
[20]
different motion aspects, in order to study the integrity of different neural
networks and the relation with visual acuity and visual field outcomes.
Acknowledgements
This study was supported by the K.U.Leuven Research Fund (K.U.Leuven

Onderzoeksfonds) grants nr. OT/01/43, PDM/01/156 and PDM/03/251 and
Royal Dutch Visio. We also thank the participating children and their parents
from Belgium and the Netherlands, the participating schools and rehabilitation
centres, all located in Rotterdam, the Netherlands, for their cooperation in the
data collection.
References
3. Guzzetta, A., et al., Motion perception in preterm children: role of prematurity and
brain damage. Neuroreport, 2009.
2001. 30(1): p. 61-72.
6. Sunaert, S., et al., Attention to speed of motion, speed discrimination, and task
difficulty: an fMRI study. Neuroimage, 2000. 11(6 Pt 1): p. 612-23.
7. Marcar, V.L., et al., An fMRI study of the cerebral macro network involved in
‘cue invariant’ form perception and how it is influenced by stimulus complexity.
Neuroimage, 2004. 23(3): p. 947-55.
8. Klaver, P., et al., Dorsal stream development in motion and structure-from-motion
6
perception. Neuroimage, 2008. 39(4): p. 1815-23.
9. Bava, S., A.O. Ballantyne, and D.A. Trauner, Disparity of verbal and performance
IQ following early bilateral brain damage. Cognitive and Behavioral Neurology,
2005. 18(3): p. 163-70.
10. van der Zee, Y., et al., Chronological age versus developmental age in evaluating
patients’ performances on motion perception tests. Accepted for publication in
Neuropsychological Trends, 2018.
14. Stiers, P., et al., The variety of visual perceptual impairments in pre-school
children with perinatal brain damage. Brain and Development, 2001. 23(5): p.
333-48.
16. Zihl, J., D. Von Cramon, and N. Mai, Selective disturbance of movement vision
after bilateral brain damage. Brain, 1983. 106: p. 313-340.
17. Pollmann, S., et al., A fronto-posterior network involved in visual dimension
changes. Journal of Cognitive Neuroscience, 2000. 12(3): p. 480-94.
18. Corbetta, M. and G.L. Shulman, Control of goal-directed and stimulus-driven at-
tention in the brain. Nature Reviews Neuroscience, 2002. 3(3): p. 201-215.
19. Ortibus, E., et al., Integrity of the inferior longitudinal fasciculus and impaired
object recognition in children: a diffusion tensor imaging study. Developmental
Medicine & Child Neurology, 2012. 54(1): p. 38-43.
134 Chapter 6
21. Duncan, R.O. and G.M. Boynton, Cortical magnification within human primary
visual cortex correlates with acuity thresholds. Neuron, 2003. 38(4): p. 659-71.
Chapter
Chapter 7
7
Object recognition and dorsal

stream vulnerabilities in children
with early brain damage
van der Zee, Y., Stiers, P., & Evenhuis, H. Object recognition and
dorsal stream vulnerabilities in children with early brain damage.
Submitted.
136 Chapter 7
Abstract
Aim
Dorsal stream functions are considered vulnerable in children with early brain
damage. Considering the recognition of objects in suboptimal representations
a dorsal stream dysfunction, we examined whether children with early brain
damage and impaired recognition have either general or selective dorsal
stream dysfunctions.
Method
Performances on motion perception, visual attention and visuoconstruction
were evaluated in patients with (n = 18) and without (n = 11) object recogni-
tion abnormalities. Increased risk of dorsal stream dysfunctions was deter-
mined by the comparison of the proportions of abnormal/weak performers
with the probabilities expected in the general population, given the cut-offs
and number of tasks used.
Results
Compared to patients with normal object recognition, patients with object
recognition problems performed significantly more often weak on motion per-
ception and visual attention (ps = .03) but did not differ on visuoconstructive
skills. Six of the eighteen patients with object recognition problems performed
weak on at least two additional dorsal stream functions, which we considered
a general dorsal stream dysfunction.
Interpretation
Children with object recognition problems are at risk for other dorsal stream
dysfunctions, but dysfunctions seem rather specific than general. Multiple
functions/aspects should be assessed in neuropsychological assessment of
children at risk.
Dorsal stream functioning in at-risk children 137
Introduction
[1]
Studies on visual perception in children with developmental disorders and
[2, 3]
early brain damage suggest that the dorsal stream of the cerebral vi-
sual system is more vulnerable than the ventral stream. The dorsal stream
[4, 5] [6]
is associated with motion perception, visuomotor integration, and
[7]
visual attention. Different brain areas are thought to be crucial for these
functions: V5/MT+ for global motion, motion speed and motion-defined form
[8-10]
perception; the superior temporal sulcus (STS) area for biological motion
[11] [12]
perception and target detection; the temporoparietal junction (TPJ) for
[13] [13]
attention shift, the intraparietal sulcus (IPS) for visuospatial attention
[12]
and response prepartion; the inferior parietal lobule (IPL) for visuomo-
[14] [7]
tor planning, and the superior parietal lobule (SPL) for attentional bias,
[14]
action, visuomotor control. The ventral stream, which projects into the
inferotemporal cortex, is associated with form perception, object recogni-
[1, 6] [1]
tion and face recognition. Although object recognition is primarily
considered a ventral stream function, existing evidence indicates that it is not 7
[15] [15, 16] [16]
a ventral stream function alone: IPS, IPL and SPL play a role in
object recognition in suboptimal representations. Damage in the parietal lobe
is associated with impaired recognition in the following suboptimal representa-
[3, 17] [18]
tions: incomplete drawings, overlapping drawings, objects presented
[18] [18]
from an unconventional view, and drawings degraded by noise.
Studies on dorsal stream functions in children with early brain damage or
developmental disorders indicate that different aspects of object recogni-
[19] [8-10]
tion and motion perception can be selectively impaired. Because in
most studies single functions or aspects were studied, it is unknown how often
multiple dorsal stream functions are affected. One study in prematurely born
children with complications such as periventricular white matter disease and
retinopathy of prematurity, showed that 20-55% of the patients had clinically
[20]
significant impairments in different dorsal stream functions. However, it
was not reported whether individual patients were impaired on multiple tasks,
and object recognition in suboptimal representations was not addressed in this
study.
The aim of the present study was to determine whether other dorsal stream
dysfunctions are commonly present in children with an identified dorsal stream
dysfunction. In this study, the L94 was used to detect children with a dorsal
stream dysfunction in a group of children with (indications of) early brain
damage. The L94 is a valid diagnostic test battery for object recognition in
suboptimal representations. Abnormal performances on the L94 are associated
138 Chapter 7
[21]
with parietal lobe damage. We hypothesized that patients with object rec-
ognition abnormalities perform significantly more often weak on other dorsal
stream aspects, i.e. motion perception, visual attention, and visuoconstructive
skills, than children with normal object recognition. We also examined how
often a general dorsal stream dysfunction was present, i.e. how often at least
2 out of three additional dorsal stream functions were weak.
Methods
Participants
The patient group consisted of 48 children at risk for object recognition prob-
lems, because of brain damage, indications of brain damage and/or reports of
suspicion of visual perceptual impairments in the medical file. Patients were
recruited through rehabilitation centres in the Rotterdam area (Rijndam Reha-
bilitation Centre and Royal Dutch Visio, Centre of Expertise for blind and par-
tially sighted people, n = 19) and through the Laboratory of Neuropsychology
at the University Hospital in Leuven, Belgium (n = 29). Their chronological age
ranged from 4y1m to 14y7m (M = 7y3m, SD = 2y4m). Their developmental
[22]
age, the median age equivalent of nonverbal intelligence subtests, ranged
from 2y5m to 7y8m (M = 5y3m, SD = 1y5m). Patient characteristics are
presented in Table 1. Eight patients had low vision, a visual acuity between
0.1 and 0.3, but all patients should be able to perceive the detailed stimuli.
Centre MEC-2006-056) and the Catholic University of Leuven. Informed con-
sent was obtained for all participants through their parents or guardians.
Table 1. Presence of neurodevelopmental and ophthalmologic conditions in patients with confirmed

or suspected brain damage
Patient group
(n = 48)
Neurodevelopmental conditions n %
Aetiology
Asphyxia 5 10
Hypoxic-ischemic encephalopathy (HIE)
Periventricular leukomalacia (PVL) 18 38
Intraventricular haemorrhage (IVH) 3 6
PVL + IVH 1 2
Malformation 3 6
Hydrocephalus 1 2
Intracranial haemorrhage (ICH) 1 2

or suspected brain damage (continued)
Patient group
(n = 48)
Intracranial haemorrhage + cytomegalovirus infection 1 2

Acquired brain damage (8 months - 2.5 years)
Tumour 1 2
Trauma 4 8
Meningitis 1 2
Genetic 3 6
Unclear 6 12
Neonatal Condition
Prematurity (Gestational age < 37 weeks) 20 41
Performance IQ (PIQ)
Normal IQ (> 84) 12 25
Borderline (71-84) 11 23
Mild retardation (50-70) 12 25
Moderate retardation (< 50) 6 13
Unknown 7 15 7
Motor disorder 31 65
Spastic cerebral palsy
Hemiplegia 7 15
Diplegia 8 16
Quadriplegia 3 6
Undefined 1 2
Nonspastic cerebral palsy
Athetoid 2 4
Ataxic 1 2
Mixed cerebral palsy 3 6
Bipyramidal syndrome 4 8
Developmental delay 3 6
(Neuro-)ophthalmologic conditions
Refractive error 10 21
Anisohyperopia 2 4
Myopia 1 2
Hyperopia 3 6
Hyperopia gravior (≥ + 6D) 2 4
Pseudoaphakia 2 4
Retinopathy of prematurity
Stage I or II 2 4
Optic disc abnormality 7 15
Pale appearance 4 8
140 Chapter 7

or suspected brain damage (continued)
Patient group
(n = 48)
Smaller than normal 1 2

Optic nerve atrophia (posttraumatic) 2 4
Strabismus 16 33
Manifest 10 21
Intermittens 4 8
Latent 2 4
Oculomotor dysfunction 8 16
Nystagmus
Manifest 3 6
Latent 1 2
Undefined 1 2
Saccadic dysfunction 2 4
Convergence abnormality 1 2
Horizontal oculomotor apraxia 1 2
Visual field defect 13 27
Scotoma 1 2
Mixed (hemi and altitude) 2 4
Hemianopsia 7 15
Altitude defect 1 2
Concentric, one side more affected 2 4
Other ophthalmologic conditions 5 10
Bilateral cataract 2 4
Posterior embryotoxon 1 2
Septo-optic dysplasia (SOD) 1 2
Choroidal colomboma + periveral fundus abnormality + intact optic 1 2
nerve
Procedures
The following dorsal stream functions were studied in arbitrary order: object
recognition in suboptimal representations; motion perception; visual atten-
tion; visuoconstructive abilities.
To assess object recognition in suboptimal representations we used five
[19]
computerised subtasks of the L94 that require recognition of line draw-
ings of everyday objects: visual matching (VISM); drawings occluded by
noise (NOISE); overlapping line drawings (OVERL); unconventional object
views (VIEW); De Vos (DE VOS). In the De Vos task, objects are embedded
in context or they are partially drawn, only contours are presented, a typical
part of the object is omitted, or they are presented in an unconventional view.

We excluded the items rifle, bench and alarm clock from the analysis of VIEW,
because Dutch controls tended to name these objects differently, for example
by their general category and often did not change their answer with changing
views. This was considered no problem, because inconclusive items, items
that are not named correctly in the control condition of the task, are excluded
in the scoring procedure of the L94. Tasks, test and scoring procedures have
[19] th
been described by Stiers et al. Scores below the 5 percentile were defined
as abnormal performances.
Motion perception was assessed with three different computerised motion
perception tasks: global motion, finding ‘the road, while its snowing’ in a
random dot kinematogram; motion-defined form, naming objects hidden in
a random dot kinematogram; and motion speed, telling which of the two car
is moving faster by looking at the horizontally moving dots. Tasks, test and
[23]
scoring procedures and cut-off values are described in Van der Zee et al.
Visual attention was assessed by a computerised visual search task. Visual
search time, i.e. response time minus reaction time, was our primary outcome 7
measure for this task. Details of the task, test procedure and cut-off values
are reported in Appendix 1.
[24]
To assess visuoconstructive skills, we administered the Beery VMI and
[25]
the subtest mosaics of the SON-R 2 ½ -7 as prescribed in the manuals.
For all function tasks, including the visuoconstructive tasks, we used devel-
opmental age, instead of chronological age, as entry to the norm tables. If de-
velopmental age was out of range of the norm tables, we used the nearest age
th
group. Scores below the 10 percentile were defined as weak performances.
Testing was done by trained senior psychology students and neuropsycholo-
gists. All computerised tasks were run on a laptop connected to a 15-inch CRT
monitor. Participants were placed in front of the screen at approximately 40
cm.
Analysis
Participants were included in the analysis if at least three out of five L94 tasks
had been completed. Because the probability of finding one or more abnor-
malities increases with the number of task evaluated, from .05 for one task to
.23 (change at no abnormalities is .77) for five tasks, we checked whether this
patient group was indeed a group at risk for dorsal stream dysfunction, i.e.
we tested with the binomial test whether the proportion abnormal perform-
ers was significantly higher than the probability in the general population. To
control for the variation in number of tasks completed, we calculated the mean
142 Chapter 7
expected proportion abnormal performers in the general population given the

number of tasks evaluated in individual patients.
To study whether children with object recognition problems had an increased
risk of a general dorsal stream dysfunction, we first divided the patient group
th
in a group with at least one object recognition abnormality (score below 5
th
percentile) and a group with normal object recognition (no score below 5
percentile). For this analysis, we only included patients who completed at
least one task per function for the other dorsal stream functions. We used the
Fisher’s exact test to analyse whether the group with object recognition ab-
normalities had a higher proportion of weak additional dorsal stream functions
th
(score below 10 percentile) than patients with normal object recognition.
We defined that the patient group with an abnormal object recognition had
rather general than specific dorsal stream dysfunctions, if more than 50% of
this group had a weak performance on at least one task in two out of three
other dorsal stream functions (motion perception, visual attention, visuocon-
structive skills).
Results
L94: object recognition

At least three out of the five L94 tasks could be evaluated in 46 of 48 patients:
5 tasks in 35, 4 in 7, and 3 in 4 patients. No tasks were systematically missing.
Given the number of tasks evaluated, at least one abnormal task performance
was expected in .21 of the general population. In Figure 1 we compared the
observed with the expected proportions. Twenty-nine out of 46 patients (.64)
performed abnormally on at least one L94 task, which is significantly higher
than expected in the general population (p < .01). The proportions of abnor-
mal performers with 1, 2 or 3 abnormal L94 outcomes were all significantly
higher than expected, indicating that this patient group is indeed at risk for
dorsal stream dysfunctions.
The proportion of abnormal performers was not significantly higher for the
subtask VISM (4/41 vs .05; ns), but was significantly higher for the other four
subtasks: 6/45 (p = .02) for OVERL; 8/45 (p < .01) for NOISE; 13/44 (p <
.01) for VIEW; 15/40 for DE VOS (p < .01).
130 | Chapter 7 Dorsal stream functioning in at-risk children 143
Figure 1. Expected
Figure 1. and observed
Expected and observed proportion
proportion abnormal
abnormal performances
performances on
on the
the L94
L94
7
Dorsal stream
Dorsal stream functioning
functioning in patients
in patients with and without with
object and without
recognition object
problems
recognition problems
Out of remaining 46 patients, 44 completed at least one motion perception task, 41 at
Out
least of
oneremaining 46 patients,
visuoconstructive task, and44
33 completed at least
the visual attention oneThis
task. motion perception
resulted in 29
task, 41inat
patients leastboth
whom oneatvisuoconstructive
least three L94 taskstask, and
and all 33 other
three the visual
dorsal attention
stream task.
This resulted
functions could in
be 29 patients
evaluated. Inin whom
Figure both
2 we at least
report three L94
the observed tasks and
proportions all three
of weak
other dorsal
functions in thestream functions
total patient group, could be with
the group evaluated. In Figure
normal L94 2 (n
outcomes we=report
11) andthe
observed
the group with abnormal L94 outcomes (n = 18). The median number of evaluated group
proportions of weak functions in the total patient group, the tasks
with normal
of other dorsal L94 outcomes
stream functions (n
was=seven
11) out
andofthe group
totally eightwith abnormal
(sub)tasks (rangeL94 out-
5-8).
comes (n we
Therefore, = 18). The that
expected median number of
the proportion of evaluated tasks
patients with of other
at least dorsal
two weak stream
functions,
functions
independentwas seven
of L94 out of
outcomes wastotally eight
.12. The (sub)tasks
proportion (range
of patients with 5-8).
at leastTherefore,
2
we expected
additional weakthat the proportion
functions of patients
was significantly with
higher than at least(ptwo
expected weak
= .04), functions,
as a result of
independent of L94
the relatively high outcomes
proportion was .12.
of patients withThe proportion
exactly of patients
3 additional with at
weak functions least
in the
2 additional
patient weak
group with functions
L94 was (p
abnormalities significantly higher than
< .01). The observed expected
proportions (pleast
for at = .04),
1
as a result of the relatively high proportion of patients with exactly 3 additional
weak function differed significantly between the patient groups (3/11 vs 13/18; p < .02),
weak functions
whereas in the
the observed patient group
proportions with
for 1 and L94 abnormalities
2 additional (p <
weak functions did.01). The ob-
not differ
served proportions
significantly for at(p
between groups least 1 weak
= .15) function
as result differed
of the small significantly
sample sizes. Thesebetween
results
the patient
indicate groups with
that patients (3/11 vs 13/18;
abnormal objectprecognition
< .02), whereas the representations
in suboptimal observed propor-
tions
have afor 1 and 2 additional
significantly weak
increased risk of a functions did stream
general dorsal not differ significantly between
dysfunction.
groups (p = .15) as result of the small sample sizes. These results indicate
that patients with abnormal object recognition in suboptimal representations
have a significantly increased risk of a general dorsal stream dysfunction.
144 Chapter 7 Assessing visual motion perception problems| 131
Figure 2. Expected and observed proportion weak dorsal functions in addition to L94 outcomes
In Table 2, we present group characteristics, proportion of weak performers per function

In Table 2, we present group characteristics, proportion of weak performers
and (sub)task and median task outcomes for both groups. The group with object
per function and (sub)task and median task outcomes for both groups. The
recognition problems had a significantly higher proportion of weak performers on motion
group with object recognition problems had a significantly higher proportion
perception (p = .03) than the group with unimpaired object recognition. They performed
of weak performers on motion perception (p = .03) than the group with un-
significantly more often weak on the global motion task (p = .02) and the coherence
impaired object recognition. They performed significantly more often weak
level was significantly higher than that in patients without object recognition problems
on the global motion task (p = .02) and the coherence level was significantly
(coherence level .53 vs .42; p = .04). Although the proportion of weak performers did
higher than that in patients without object recognition problems (coherence
not differ significantly on the motion-defined form task, patients with object recognition
level .53 vs .42; p = .04). Although the proportion of weak performers did
problems were significantly less able to recognise the motion-defined forms (proportion
not differ significantly on the motion-defined form task, patients with object
correct .72 vs .92, p< .01). The proportion of weak performers for the visual attention
recognition problems were significantly less able to recognise the motion-
task was also
defined significantly
forms higher
(proportion in the group
correct with.92,
.72 vs object
p recognition
< .01). Theproblems (p =
proportion of
.03). Although the observed proportions of weak performers did not differ significantly for
weak performers for the visual attention task was also significantly higher
the visual search subtasks, the difference in weak performers was near-significant for the
in the group with object recognition problems (p = .03). Although the ob-
condition with four distracters (p = .09). Additionally, patients with object recognition
served proportions of weak performers did not differ significantly for the visual
problems were significantly slower on each subtask and made significantly more mistakes
search subtasks, the difference in weak performers was near-significant for
than the group with unimpaired object recognition (ps <.01). The differences between
the condition with four distracters (p = .09). Additionally, patients with object
the groups could not be explained by differences in chronological age, developmental
recognition problems were significantly slower on each subtask and made
age, or motor response
significantly time (reaction
more mistakes timegroup
than the visualwith
search task). No significant
unimpaired differences
object recognition
were
(ps <found forThe
.01). the visuoconstructive skills. the groups could not be explained by
differences between
differences in chronological age, developmental age, or motor response time
(reaction time visual search task). No significant differences were found for
the visuoconstructive skills.
Table 2. Age parameters, proportion of weak performers per function and tasks, and task outcomes
for patient with normal and abnormal performances on L94.
Patients with 3 additional function evaluations
a
Normal L94 ≥ 1 abnormal L94 task p-value
n = 11 n = 18
Median age
Chronological 6;2 6;2 ns
(range) (4;1 – 8;7) (4;1 – 10;2)
Performance 5;6 4;8 ns
(range) (3;1 – 7;8) (2;5 – 6;8)
Motion perception
# weak 1/11 9/18 .03
Global motion task
# weak 0/10 7/16 .02
median coherence level .42 .53 .04
(range) (0.32-0.58) (0.12-0.83)
Motion speed task
# weak 0/9 0/5 ns
median speed difference deg/s 4.83 4.28 ns
(range) (1.70 - 9.66) (1.27 – 23.24)
Motion-defined form task 7
# weak 1/10 5/16 ns
median proportion correct .92 .71 < .01
(range) (0.73 -1.00) (0.00 – 0.97)
Visual attention
# weak 1/11 9/18 .03
Search task 4 distracters
# weak 1/11 7/18 ns
median search time sec 0.73 1.21 .01
(range) (0.00 -1.97) (0.56 – 3.45)
# weak 0/11 3/18 ns
median search time sec 1.13 1.74 < .01
(range) (0.72 - 1.71) (0.71 – 6.72)
# weak 0/11 3/18 ns
median search time sec 2.11 3.90 < .01
(range) (0.99 - 3.46) (1.17 - 11.55)
mistakes
# weak 0/11 9/18 < .01
median number 0.00 1.00 < .01
(range) (0.00 - 0.00) (0.00 -11.00)
reaction time task
# weak 0/11 1/18 ns
median response time 1.32 1.49 ns
(range) (0.90 - 2.55) (0.94 - 2.83)
146 Chapter 7
Table 2. Age parameters, proportion of weak performers per function and tasks, and task outcomes
for patient with normal and abnormal performances on L94. (continued)
Patients with 3 additional function evaluations
a
Normal L94 ≥ 1 abnormal L94 task p-value
n = 11 n = 18
Visuoconstructive skills
# weak 2/11 3/18 ns
VMI
# weak 0/6 2/16 ns
median standard score 96.50 92.00 ns
(range) (83 - 106) (73 - 142)
Mosaics (SON-R)
# weak 2/11 2/18
median standard score 10.00 10.50 ns
(range) (3 -12) (5– 15)
a
The significance of differences in median scores were tested with Mann-Whitney test and the sig-
nificance of differences in proportions were tested with the Fisher’s Exact Test (2x2-table, one-sided
significance)
Discussion
In this study, we first demonstrated that children with early brain damage are
at risk for dorsal stream problems, such as problems with object recognition in
[21]
suboptimal representations, by comparing their performances on the L94
with published norm values. While controlling for their developmental age, 29
of 46 children with early brain damage performed significantly abnormal on
one or multiple object recognition subtasks. This was significantly higher than
th
expected in the general population, given the use of 5 percentile cut-off val-
ues and number of tasks evaluated. We then studied whether general dorsal
stream problems were common in children with impaired object recognition
and compared their performances to that of children with early brain damage
with unimpaired object recognition. We defined a general dorsal stream dys-
th
function as a weak performance (score below 10 percentile) on at least two
additional dorsal stream functions, such as motion perception, visual atten-
tion, visuocontstructive skills. The results showed that in 6 of 18 patients with
impaired object recognition a general dorsal stream dysfunction was present,
another 7 patients performed weakly on one additional dorsal stream function.
Dorsal stream problems were uncommon in patients without object recogni-
tion problems: they were only found in 3 of 11 patients with unimpaired object
recognition. Therefore, the risk of a general dorsal stream dysfunction can be
considered increased in children with object recognition problems, but dorsal
stream dysfunctions are rather specific than general. Motion perception and
visual attention, but not visuoconstructive skills, were specifically affected.
Patients with object recognition problems performed worse on the global
motion and motion-defined form task and were significantly slower on the
visual search tasks. They also made more mistakes, which makes a speed-
accuracy trade-off less likely. The available data are insufficient to study causal
or interactional relationships between functions. One possibility is that weak
visual attention leads to weak motion and object perception scores, but per-
formances on the assessed tasks could also be weak as a result of difficulties
in object discrimination, impulse control and cognitive flexibility. Therefore, to
test hierarchical models, not only a larger sample size but also assessment of
additional indicators is required.
The presence of object recognition and motion perception problems in com-
bination with the normal visuoconstructive performances on the Beery VMI
support the more fundamental idea that developmental age estimations based
on PIQ subtests can be used to control for effects of motor impairments in
[24]
addition to intellectual impairments. Because the Beery VMI can provide 7
age equivalents, it might be suggested that the outcome of the Beery VMI
could be used to estimate developmental age and control for motor impair-
ments by using these age equivalents as entry of the norm tables. Although
[24]
performance IQ and the outcome of the Beery VMI are significantly related,
we consider the developmental age estimation based on PIQ subtests more
reliable, because the estimation is based on multiple subtest results instead of
a single test outcome. In current study, we found two abnormal performers on
the Beery VMI. The use of the age equivalents of the Beery VMI would make
this impossible.
Further, a more detailed analysis of inconclusive items, i.e. items that were
not named correctly in the control conditions of the L94 and the motion-defined
form task and were excluded from the calculation of perception scores, might
help explain performance patterns in children at risk. Inconclusive items can
indicate differences in experience or object knowledge, but also the pres-
ence of other problems such as problems in language development, naming,
especially in relation to words (items) with lower word frequencies, memory
and attention, or a combination of these problems. Differences in the number
of inconclusive items between a group with unimpaired and impaired object
recognition could provide indications for the reason why abnormal performers
were unable to name or describe objects in the control items.
We conclude that in children with early brain damage, dorsal stream functions
and their aspects seem specifically and not generally affected, and that more
148 Chapter 7
extensive research is required for a better understanding of causal relation-

ships and underlying mechanisms. For now, we recommend that in specialised
clinical practice, multiple functions and their different aspects are assessed in
a neuropsychological assessment of at-risk children, using developmental age
as reference level, i.e. entry of the norm table.
Acknowledgement
This study was supported by the K.U.Leuven Research Fund (K.U.Leuven

Onderzoeksfonds) grants nr. OT/01/43, PDM/01/156 and PDM/03/251 and
Royal Dutch Visio. We also thank the participating children and their parents
from Belgium and the Netherlands, the participating schools and rehabilita-
tion centres for their cooperation in the data collection. We are grateful to
the Department of Psychologic and Pedagogic Sciences of Leuven Catholic
University for the use of this visual search task, developed by two of their
master students.
References
3. Fazzi, E., et al., Visual-perceptual impairment in children with periventricular
leukomalacia. Brain Dev, 2004. 26(8): p. 506-12.
2001. 30(1): p. 61-72.
6. James, T.W., et al., Ventral occipital lesions impair object recognition but not
object-directed grasping: an fMRI study. Brain, 2003. 126(Pt 11): p. 2463-75.
7. Pollmann, S., et al., Separating distractor rejection and target detection in poste-
rior parietal cortex--an event-related fMRI study of visual marking. Neuroimage,
2003. 18(2): p. 310-23.
9. Reiss, J.E., J.E. Hoffman, and B. Landau, Motion processing specialization in Wil- 7
ogy & Strabismus, 2007. 44(6): p. 363-371.
11. Bonda, E., et al., Specific involvement of human parietal systems and the
amygdala in the perception of biological motion. Journal of Neuroscience, 1996.
16(11): p. 3737-44.
12. Pollmann, S., et al., A fronto-posterior network involved in visual dimension
changes. Journal of Cognitive Neuroscience, 2000. 12(3): p. 480-94.
tention in the brain. Nat Rev Neurosci, 2002. 3(3): p. 201-15.
14. Glover, S., Separate visual representations in the planning and control of action.
Behav Brain Sci, 2004. 27(1): p. 3-24.
15. Vuilleumier, P., et al., Multiple levels of visual object constancy revealed by event-
related fMRI of repetition priming. Nat Neurosci, 2002. 5(5): p. 491-9.
16. Eger, E., et al., Mechanisms of top-down facilitation in perception of visual objects
studied by FMRI. Cereb Cortex, 2007. 17(9): p. 2123-33.
17. Warrington, E.K. and M. James, Disorders of visual perception in patients with
localised cerebral lesions. Neuropsychologia, 1967. 5(3): p. 253-266.
18. Warrington, E.K. and A.M. Taylor, The contribution of the right parietal lobe to
object recognition. Cortex, 1973. 9(2): p. 152-64.
19. Stiers, P., et al., The variety of visual perceptual impairments in pre-school chil-
dren with perinatal brain damage. Brain Dev, 2001. 23(5): p. 333-48.
150 Chapter 7
21. Ortibus, E., et al., Assessment of cerebral visual impairment with the L94 visual
perceptual battery: clinical value and correlation with MRI findings. Dev Med Child
Neurol, 2009. 51(3): p. 209-17.
22. Stiers, P., P. De Cock, and E. Vandenbussche, Separating visual perception and
non-verbal intelligence in children with early brain injury. Brain and Development,
1999. 21(6): p. 397-406.
23. van der Zee, Y., et al., Chronological age versus developmental age in evaluating
patients’ performances on motion perception tests. Accepted for publication in
Neuropsychological Trends, 2018.
24. Beery, K.E., N.A. Buktenica, and N.A. Beery, Beery VMI: Beery-Buktenica De-
velopmental Test of Visual-Motor Integration 5th Edition. 2004, San Antonio:
Pearson Assessment.
25. Tellegen, P.J., et al., Snijders-Oomen nonverbal intelligence test : SON-R 2 1/2 -
7: manual and research report. Vol. Special. 1998, Lisse: Swets & Zeitlinger. 256.
Chapter
Chapter 7:
7: Appendix
Appendix 1
1
Visual search: task characteristics,

procedures and 10th percentile
cut-off values
Appendix 1: Visual search task 153
General procedure
All controls were tested at the children’s primary schools. Task administration
was done by trained senior psychology students or neuropsychologists.
Visual Tasks
search task | 139
were presented on a 15-inch CRT monitor attached a laptop. Participants were

General procedure
placed in front of the screen at approximately 40 cm.
All controls were tested at the children’s primary schools. Task administration was done
by trained senior psychology students or neuropsychologists. Tasks were presented on a
15-inch CRT monitor attached a laptop. Participants were placed in front of the screen at
Visual search tasks
approximately 40 cm.
The computerised visual

Visual search
search taskstasks, developed by master thesis students
at the department
Theof Psychologic
computerised and
visual Pedagogic
search Sciences
tasks, developed of Leuven
by master Catholic
thesis students at the
1 department of Psychologic and Pedagogic
(size Sciences of Leuven at
Catholic University, 1
University, consisted of greyscale pictures 5 x 5 degrees 40 cm dis-
tance) on a greyconsisted of greyscale pictures (size 5 x 5 degrees at 40 cm distance) on a grey
background (area size approximately 37x29 degrees, see
background (area size approximately 37x29 degrees, see example in Figure 1).
example in Figure 1).
7A1
Figure
Figure 1. 1. Example
Example of visual
of the the visual search
search tasknine
task with with nine
distracters and 1 target (bike).
distracters and 1 target (bike).
Before testing was started, all pictures were presented on the screen to familiarise the
Before testing participants
was started, allpictures.
with the picturesThe were presented
participants ontothe
were asked screen
name to
the individual
familiarise the participants

pictures, after with
which the pictures. The
the red-bordered targetparticipants wereofasked
picture in the centre to was
the screen
introduced.
name the individual The participant
pictures, was instructed
after which to point out thetarget
the red-bordered picture picture
identical to
inthe central
picture as fast and accurate as possible and to put his/her hands on the table in front of
the centre of the screen was introduced. The participant was instructed to
the screen before each trial. The moment the participants touched a picture on the
point out the picture identical to the central picture as fast and accurate as
screen, the test administrator pressed the spacebar and the trial was ended. To verify
possible and to put his/her hands on the table in front of the screen before
that the participants understood the task, three practise-trials with a target stimulus and
each trial. The moment the participants touched a picture on the screen, the
two distracters were presented.
test administrator pressed
Testing the with
was started spacebar
a simpleand the trial
reaction-time was
task withended.
five trials.To verify
In the reaction-time
that the participants understood
task only the task,
the central picture three
and the targetpractise-trials with Because
picture were presented. a target
the location
stimulus and twoofdistracters werewould
the target stimulus presented.
pop-out to the participants, the reaction time was
Testing was started with a simple reaction-time task with five trials. In
the reaction-time task only the central picture and the target picture were
presented. Because the location of the target stimulus would pop-out to the
participants, the reaction time was considered identical to the motor response
time. The motor response time not only included the time needed by the
participants to point out the target stimulus, but also the time needed by the
administrator to press the spacebar.
In the next stage, three visual search tasks were presented with four, nine
and nineteen distracters. Each task consisted of 10 trials. If needed, the par-
ticipant was encouraged to keep looking. To control for effects of fatigue and
task experience, testing was ended with the above reaction time task.
All reaction times were saved and the administrator noted all false alarms
(mistakes: child pointed out a wrong picture). Reaction times for false alarms
were excluded. Median reaction times were calculated for both reaction time
tasks and for each visual search task. Because the presence of distracters in
the search tasks made pop-out less likely, a serial search process (scanning
individual pictures) was assumed to be needed to detect the targets. In a
serial search process, the reaction time is the sum of the motor response
time and the visual search time. Visual search time was our primary outcome
measure, so median reaction times of the reaction time tasks were distracted
from median reaction times of the search tasks.
th
10 percentile cut-offs for the visual search tasks
The control group consisted of 60 typically developing children (25 boys, 35
girls) without any signs of neurological or visual impairment and with a normal
visual acuity. They were recruited through primary schools in Rotterdam, The
Netherlands. Their chronological age ranged from 4y3m to 7y4m (M = 5y7m
SD = 9m). We divided the group in three age groups, equal to those for the
motion perception tasks. Cut-off values can be found in Table 1.
th
Table 1. 10 percentile scores for visual search outcomes in different age groups.
5 items 10 items 20 items Reaction task
Search time Search time Search time Response Total number
n (sec) (sec) (sec) time (sec) of errors
4y3m -4y7m 7 1.38 2.75 5.91 2.54 2
4y9m -5y8m 23 1.23 2.20 5.47 2.03 1
5y10m - 7y4m 26 1.09 2.66 4.16 1.70 1
Appendix 1: Visual search task 155
References
1. Ledeganck, Luyten M & Vandebussche E. Visueel zoekgedrag bij kinderen met
cerebrale of oculaire visuele inperking: een empirisch onderzoek (Visual search of
children with cerebral or ocular visual impairment: an empyrical study). Catholic
University Leuven, Belgium, Dept. Psychologic and Pedagogic Sciences, 2002.
7A1
Chapter
Chapter 8
8
Remote eye tracking assesses

age dependence processing of
coherent motion in typically
developing children
Pel, J. J. M., van der Zee, Y. J., Boot, F. H., Evenhuis, H. M., &
van der Steen, J. (2013). Remote eye tracking assesses age
dependence processing of coherent motion in typically-developing
children. Journal of Medical Engineering & Technology, 37(2), 109-
115. doi:doi:10.3109/03091902.2012.752043
158 Chapter 8
Abstract
The aim of this study was to quantify processing of different types of coher-
ent motion in terms of ocular motor response times in a group of normally
developing children (age 0-12+ years old) using remote eye tracking. Motion
coherence was applied in three different types of Random Dot Kinematograms
(RDKs): vertical (RDK1) and diagonal (RDK2) motion and expansion (RDK3).
Orienting eye movements were quantified using the Reaction Time to the
first Fixation (RTF). The children were divided into two groups: the ‘young-
est group’ between 0-3+ years and the ‘oldest group’ between 4-12+ years
old. The results showed that RTF was significantly prolonged in the ‘youngest
group’ compared to the ‘oldest group’ for each RDK. In the ‘oldest group’, RTF
was significantly affected by the type of RDK shown. The presented results
suggest that based on ocular motor responses age dependence of processing
different types of coherent motion may be revealed.
Key words: Eye movements; Random Dot Kinematograms; coherent motion;

reaction time; eye tracking
Motion perception and eye tracking 159
Introduction
Processing of motion is one of the most fundamental aspects of vision. Visual

motion sensitivity originates from retinal ganglion cells that convey their in-
formation via the magnocellular pathways to the lateral geniculate nucleus
(LGN). From here, the stream of motion signals leads via the primary visual
cortex (V1) to the posterior and anterior parietal cortex, including V3a, V5+
[1-3]
(encompassing V5, MST and other alleged motion areas) and/or hMT+
and extends towards the anterior parietal and the posterior frontal cortex.
This path-way that is used for motion processing, spatial cognition and visual-
[4]
motor planning is also known as the dorsal network. Motion processing
has been extensively investigated at the functional level using Random-Dot-
Kinematograms (RDKs). RDKs mostly consist of a black background combined
with white moving dots displayed on a monitor, either to trigger optokinetic
nystagmus (OKN) or to test coherent (or global) motion. RDKs with full field
coherently moving dot’s mostly trigger OKN. RDKs with two or distinct target
areas have been commonly used in behavioural discrimination tasks in very
young children e.g. preferential looking (PL). Here, the stimulus consists of
a target area in which the dots move in opposite direction compared to the
majority of the dots in the back-ground.
[5,6]
It has been shown that the sen- 8
[7]
sitivity to motion is already present from birth or shortly after a few weeks
[8]
of age and that different visual pathways are involved in processing OKN
[9]
compared to PL during development. Therefore, measurement of coherent
motion sensitivity in relation to age may provide valuable information about
maturation of the visual pathways in healthy children and in children with
visual processing dysfunctions.
Quantification of processing coherent motion presented in an RDK may be
done by a button press, a verbal response or by observing eye movements.
Also, psycho-physical thresholds can be assessed, e.g. by using a stair-case
procedure during a true/false method in which the number of coherent moving
white dots in the target area is gradually reduced. There is, however, a large
variability in RDKs used in clinical and scientific studies, i.e. stimulus type,
direction of coherent motion, dot size and speed. This makes it difficult to
compare studies available in literature on individual thresholds and reaction
times across different RDK tasks. Additionally, in most RDKs, the dots in the
target area move in opposite direction to the dots in the background. This
induces a motion discontinuity at the intersection of both areas. Here, the rela-
tive speed of passing dots in opposite direction is twice the average dot speed.
This creates a strong transient edge effect, which is absent in optic flow stimuli
160 Chapter 8
like expansion. In so-called motion-defined form stimuli, these edges are used
[10]
to make different forms visible. However, it is still unknown to what extent
these edge effects contribute to coherent motion detection. Finally, most RDK
tasks require subjects to understand the specific test instructions and to be
able to give verbal or motor responses. In children under the age of 3-4 years
or children with intellectual disabilities, these tasks cannot be performed ad-
equately. Moreover, the outcome measures depend on the skills of an observer
who assesses orienting responses as a measure for correct motion detection.
To overcome (part of) these problems, we developed a method to quantify
visual orienting responses to RDKs in children using remote eye tracking tech-
[11]
nology. Visual stimuli are presented on an eye tracker monitor. The stimuli
that are shown test basic eye movements (saccades and pursuit), lower order
visual functions (visual acuity and contrast) and higher order visual functions
(form and motion coherence and competitive dots and non-competitive dots).
Upon presentation, orienting behaviour in response to these stimuli is mea-
sured. This approach requires no specific verbal instructions prior to the test or
active cooperation in terms of pointing or pressing a button. In addition, it allows
objective quantification of orienting responses in terms of reaction time and fixa-
[11]
tion areas. The first aim of this study was to compare processing of coherent
motion in terms of reaction times to three different types of RDKs in children
between 0-13 years of age: coherent motion in vertical (RDK1) and diagonal
(RDK2) motion and expansion (RDK3). Our second aim was to investigate if the
presence of motion edges in RDK1 and RDK2 affected orienting behaviour.
Materials and Methods
Study population
We approached 430 healthy children between 0-13 years old in the region
Rijnmond, Rotterdam, The Netherlands. The children aged 0 till 4 years (0-
3+), the so-called ‘youngest group’ attended a regular day-care centre and
the children aged 4 till 13 (4-12+) years, the so-called ‘oldest group’ attended
a regular primary school. Parents were informed about the study by letter
and 188 (~44%) written consents were received. Children had normal or
corrected-to-normal vision. We included for this study 104 females, 6.2 (SD =
3.5) years and 84 males, 5.7 (SD = 3.6) years. The experimental procedures
were approved by the Medical Ethical Committee of Erasmus University Medi-
cal Centre, Rotterdam, The Netherlands (METC-2006-055). The study adhered
to the Declaration of Helsinki for research involving human subjects.
Measurement setup and procedure

The setup consisted of a 17-inch monitor with an integrated 50 Hz infrared
eye-tracking system (Tobii 1750, Tobii Corporation, Sweden). The measure-
ments were conducted in a quiet room. Children sat on their parents lap or on
a comfortable chair at ~60 cm in front of the eye tracker. The system’s latency
was 30ms. First a standardized 5-point calibration procedure of both eyes was
performed. Next a sequence of short movies was shown. The objective was
to randomly present short movies of visual stimuli for testing higher visual
functions, e.g. form and motion coherence. Each stimulus contained a specific
area with a higher salience defined as the target area. In the present study,
we analysed three different types of coherent motion stimuli, see Figure 1.
In Table 1, an overview is presented of the main characteristics of each
stimulus. During one trial, the stimulus RDK1 (vertical motion) was shown
two times (the target area on the left and on the right monitor side) and
the stimuli RDK2 (diagonal motion) and RDK3 (expansion motion) were each
shown 4 times (the target area in all 4 monitor corners), for a duration of 4 s.
All dots had a limited lifetime of 0.4 s. We started with including the ‘youngest
children’ and prepared two different sequences (sequence I and sequence II)
of movies, each with a duration of ~10 min. Note that in this initial phase of
our research, sequence I only contained RDK1 and RDK2 and sequence II only 8
contained RDK1 and RDK3. In total, we included 78 children in the ‘youngest
group’. Sequence I was shown to 56 children, sequence II was shown to 42
children. To a subset of 17 children, both sequences were shown for a test-
retest. Thus, RDK1 was shown to all 78 children, RDK2 to 56 children and
RDK3 to 42 children. Next, we included the ‘oldest children’, and we prepared
one sequence that contained all three RDK’s, with a duration of ~15 min. In
total, we included 110 children in this group. Orienting eye movements were
stored on the hard disk and manually analysed off-line using custom Matlab
programs (Mathworks Inc., Natick, MA, USA).
Table 1. An overview of the main characteristics of the three types of RDKs shown to healthy sub-
jects of 0-12+ years old; see also Figure 1 for a schematic illustration of each RDK.
Coherent Motion
Dot RDK1 RDK2 RDK3
direction vertical diagonal Expansion
velocity (degree/s) 11.8 11.8 11.8
size (degree) 0.25 0.25 0.25
2
density (dots/degree ) 2.6 1.3 2.6
RDK = Random Dot Kinematogram

162 Chapter 8
Figure 1. The top panel illustrates the

vertical coherent motion (RDK1). White
dots in a 5 degree width vertical bar on
the left side of the monitor (the target
area indicated with the dashed lines)
moved in upward direction, while the
dots in the remaining areas moved
downwards. The middle panel illustrates
the diagonal coherent motion (RDK2).
In the target area, the top left corner,
the white dots moved in coherent diago-
nal direction, while in the remaining ar-
eas, the white dots moved in opposite
diagonal direction. The bottom panel
illustrates coherent expansion motion
(RDK3). White dots moved from the cen-
tre of the target area, the top left corner)
to the borders of the monitor. All dots
had a limited life time of 0.4 s.
Figure 1. The top panel illustrates the vertical coherent moti

Data analysis and statistics

Data analysis started with the calculation of the visual angles between gaze
point and the edge of a target area. For RDK1 this edge was defined as two ver-
tical lines, positioned at three degrees to the left and to the right of the target
area midline. For RDK2, this edge was defined as a diagonal line, positioned
at six degrees from the target area midpoint. For both RDKs, this resulted in
a target with a margin at the edge of 0.5 degrees. For RDK3, this was an 8
degrees circle around the expansion point in the centre of the target area. For
each RDK presentation, we selected the time values that gaze crossed the pre-
set border. The reaction time to an RDK was calculated as the 15th percentile
of the two-time values obtained for each RDK. This reaction time was denoted
as the Response Time to Fixation (RTF), defined as the time between showing
a stimulus and landing of gaze within the predefined target areas. The focus
of gaze analysis was to assess reflexive orienting eye movements to the target
area as a best estimate of visual processing time. If gaze was already in the
target area of an RDK at presentation, eye movements were not included for
analysis. In addition, an eye movement response was excluded for analysis if
(1) no gaze data was available within the first 500 ms after presentation of the
stimulus, (2) a reaction time was deter-mined within 120ms after presentation
of the stimulus, (3) more than three saccades were made to reach the target 8
area, or (4) the duration of fixation prior to an eye movement was more than
1500 ms and (5) the duration after reaching the target area was less than
200 ms. Finally, author JP checked by visual inspection whether the first or
second saccade fixated the edge of motion discontinuities (using a margin
of about 1 degree) caused by dots moving in opposite directions, see Figure
1, top and middle panel for a schematic illustration. The percentage of edge
fixations was calculated by dividing the total number of edge fixations divided
by the total number of correctly analysed RDKs for the complete group. Two
separate age groups were defined: the ‘youngest group’ of children were the
preschoolers between 0-3+ years old and the ‘oldest group’ of children went
to primary school and were aged between 4-12+ years old. A Kolmogorov-
Smirnov test was done to test parameter values for a normal distribution. It
revealed that a normal distribution for the RTF values per RDK existed in the
‘oldest group’. A Mann-Whitney U-test was applied to test between groups
(age group). In the ‘oldest group’, a one-way repeated-measures analysis of
variance (ANOVA) was used to test within subjects effects (RDKs). Bonferroni
posthoc testing was used for pairwise comparison of RTF-values within these
subjects. All statistical analyses were performed in SPSS-17 (SPSS, Chicago,
US) and significance levels of p < .05 were considered statistically significant.
164 Chapter 8
Results
Figure 2 top panel illustrates, superimposed, orienting gaze starting at the

onset of the presentation of the two types of RDK1, i.e. the target area on
the left side and on the right side of the eye tracker monitor. After detection
of the coherent motion (either on the left or right side), the subject made an
eye movement into the target area. Figure 2, bottom panel, illustrates, again
superimposed, orienting gaze starting at the onset of presentation of each of
four types of RDK2 stimuli. After detection of the coherent motion, this subject
fixated the motion edge three times and one time the centre of the target area
was fixated (the upper right target area).
148 | Chapter 8
Figure 2. The top panel illustrates superim-

posed the orienting gaze to the two target areas
(one on the left side and one on the right side)
of RDK1. After detection of the vertical coher-
ent motion, the subject made an eye movement
into the target area. The bottom panel illus-
trates superimposed the orienting gaze starting
at the onset of presentation of the four types
of RDK2. After detection of the coherent diago-
nal motion, three of the eye movements of this
subject resulted in transient edge fixation, while
one time, a saccade was made into the target
area (the top right monitor corner).
Figure 2. The top panel illustrates superimposed the orienting gaze to the two target areas (one
on the left side and one on the right side) of RDK1. After detection of the vertical coherent motion,
the subject made an eye movement into the target area. The bottom panel illustrates
superimposed the orienting gaze starting at the onset of presentation of the four types of RDK2.
Figure 3 shows for RDK3 the distance between the gaze point and the mid-
After detection of the coherent diagonal motion, three of the eye movements of this subject
resulted in transient edge fixation, while one time, a saccade was made into the target area (the
top right monitor corner).
point of expansion and the preset eight degree borderline (the dashed line).
Figure 3 shows for RDK3 the distance between the gaze point and the mid-point of
From the time values at which gaze crossed the 8 degrees border, the Reaction
expansion and the preset eight degree borderline (the dashed line). From the time values
Time to Fixation was calculated. Note that the duration of presentation is 4 s.
at which gaze crossed the 8 degrees border, the Reaction Time to Fixation was
calculated. Note that the duration of presentation is 4s. We plotted the first 2.0s for
We plotted the first 2.0 s for illustration purpose only.
illustration purpose only.
Figure 4 shows the RTF-values of the children to respectively RDK1, RDK2

and RDK3 against age. In each panel, a power line was fitted to the data
to illustrate the decrease of RTF-values with increasing age (dashed lines).
Orienting responses to the RDKs showed a development over age, see Figure
4. Large variability in reaction times was found in children between 0-3+ years
Motion perception and eye tracking | 149
Figure 3. The top panel shows the visual an-

gles between gaze and centre of the target area
against time of 4 coherent expansion motion
stimuli presented in RDK3. The bottom panel
shows the cumulative plot of the time values
that gaze crossed an 8 degrees border (dashed
line) of the target area. An exponential curve
with time constant tau was fitted to this cumula-
tive plot to quantify the Reaction Times to Fixa-
tion (RTF).
Figure 3. The top panel shows the visual angles between gaze and centre of the target area
against time of 4 coherent expansion motion stimuli presented in RDK3. The bottom panel shows
the cumulative plot of the time values that gaze crossed an 8 degrees border (dashed line) of the
target area. An exponential curve with time constant tau was fitted to this cumulative plot to
old. The fastest reaction times with small variability were found in children
quantify the Reaction Times to Fixation (RTF).
between 8-12+ years.

Figure 4 shows the RTF-values of the children to respectively RDK1, RDK2 and RDK3
In Table 2, an over-view is presented of the orienting responses to the three
against age. In each panel, a power line was fitted to the data to illustrate the decrease 8
of RTF-values with increasing age (dashed lines). Orienting responses to the RDKs
RDKs. The results are presented in two subgroups: children between 0-3+
showed a development over age, see Figure 4. Large variability in reaction times was
years old (n = 78) and children between 4-12+ years old (n = 110). The success
found in children between 0-3+ years old. The fastest reaction times with small
rate in assessing RTF-values in the ‘youngest group’ was highest when using
variability were found in children between 8-12+ years.
RDK1 and RDK3 (vertical and expansion motion) and in the ‘oldest group’ this
rate was almost equally high for all three RDK’s. The Kolmogorov-Smirnov test
showed normal distribution for the RTF values per RDK in the ‘oldest group’. In
addition, only 9 children in the ‘youngest group’ delivered an RTF value for all
RDK’s, whereas this figure was 92 in the children of the ‘oldest group’. Based
on the Mann-Whitney U-test, RTF values were significantly prolonged in the
‘youngest group’ compared to the ‘oldest group’ for each RDK. For the ‘oldest
group’, the Mauchly’s test, as part of the one-way repeated-measures ANOVA,
2
indicated that the assumption of sphericity was not violated, Х (2) = 2.46, p
= .293 (epsilon=.64). The results showed that RTF was significantly affected
by the type of RDK shown, F(2,182) = 14.66, p < .001. Pairwise comparison
within subjects revealed significant differences between RDK1-RDK2 (mean
difference -80 ms (SD 15 ms); p < .001) and between RDK1-RDK3 (mean
difference -70 ms (SD 15 ms), p < 0.001). Finally, we found that 7% of
all subjects fixated edges of motion discontinuities in RDK1. In RDK2 this
percentage was increased to 64%.
166 Chapter 8
150 | Chapter 8
Figure 4. In the top, middle and bottom panel,

RTF-values of the control group to respectively
RDK1, RDK2 and RDK3 are plotted against age.
In each panel, an age-related power line was
fitted to the data (dashed lines).
Figure 4. In the top, middle and bottom panel, RTF-values of the control group to respectively
RDK1, RDK2 and RDK3 are plotted against age. In each panel, an age-related power line was fitted
to the data (dashed lines).
Table 2. An overview of the objectively assessed orienting responses to three types of coherent
In Tabletasks
motion 2, an results
over-view is presented
presented of the
in two orienting responses
subgroups: to the0-3+
children aged threeyears
RDKs.(n
The
= 78) and children
aged
results4-12+ years old
are presented in (n
two= subgroups:
110). children between 0-3+ years old (n = 78) and
children between 4-12+ years old (n = 110). The success rate inCoherent
assessingmotion values
RTF-values in
the
Age‘youngest
group group’ was highest when using RDK1 and
Parameter RDK3 (vertical and
RDK1 expansion
RDK2 RDK3
motion) and in theSuccesses
0-3+ years ‘oldest group’
% this rate was almost equally
72 high for all three
53 RDK’s. 81
(number)
The Kolmogorov-Smirnov (56/78)
test showed normal distribution (28/53)
for the RTF values per RDK in (34/42)
the ‘oldest group’. Median RTF-values

In addition, ms
only 9 children 810
in the ‘youngest 940 an RTF
group’ delivered 820
(25-75 percentiles) (695-980) (830-1160) (600-1200)
4-12+ years Successes % 93 94 96
(number) (102/110) (103/110) (106/110)
Median RTF-values ms 505 560 575
(25-75 percentiles) (450-580) (480-705) (490-690)
0-12+ years Edge fixations % 7 64 -
RTF = Reaction Time to Fixation; RDK = Random Dot Kinematogram

Discussion
The aim of the present study was to quantify the processing of motion of
three different RDKs in normally developing children by means of orient-
ing responses. In the ‘oldest group’ the RDKs with vertical coherent motion
induced significantly faster ocular motor responses, up to 70 ms, than the
RDKs with diagonal and expansion motion. One explanation might be that the
diagonal motion stimulus contained only half the number of dots displayed
on the monitor making this stimulus less pronounced compared to the coher-
ent vertical and even the expansion motion. Differences in RTF-values might
also relate to differences in target areas used, since each target area was
different in size and screen location. We tested the influence of target area
size on RTF by increasing and decreasing each area by 25%. We found that
the RTF-values differed on average not more than 30 ms (i.e. a maximum of
less than two sample values given a sample frequency of 50 Hz), indicating
that area size only partially could explain the difference found in RTF-values.
Previously, it was shown that cells in hV5+, located in the inferior temporal
[12-16]
sulcus (ITS), are highly sensitive to directional motion. Processing of
optic flow stimuli like rotation or expansion, predominantly activate cells in
MST.
[17-21]
The presented results suggest that, based on the reaction times 8
to fixation, age dependence exists in processing different types of coherent
motion.
Performance of motion processing is mostly based on psychophysical thresh-
old methods or visual event-related potentials (VERPs). Based on thresholds,
it was found in healthy children aged 4-11+ that motion processing reached
[22]
adult levels at the age of 10-11+ years. Based on VERPs, processing of
expansion motion showed a similar time course, i.e. reaching of adult levels at
[23]
the age of 8-10+ years. The present data not only support these previous
findings, but in addition show quantitative differences in processing different
types of coherent motion. As stated, orienting eye and/or head responses are
reflexively induced towards visual stimuli when its information is processed by
the brain. At this level of processing visual information, processing of (visual)
perception to name the shape of the form or to indicate the direction of a mo-
tion, like in psychophysical threshold methods, might not even be completed
yet. In a future study, it would be interesting to test the RTF-values as a
function of decreasing coherent threshold levels. It might be that timelines of
maturation become different when levels of coherent motion decrease.
In the present study, we remotely assessed gaze while the RDKs were pre-
sented on the monitor of the eye tracker. Gaze responses can be analysed
168 Chapter 8
in two different ways to quantify the performance of a preferential looking

based RDK. Firstly, the outcome may be a simple classification of a target
area as seen or unseen based on eye movements, instead of a verbal or motor
response. When the RDKs are extended with stepwise reduction of the number
of coherent moving dots in combination with a staircase procedure, objective
individual thresholds can even be assessed. We are currently developing these
RDKs in our lab to assess thresholds for coherent form, motion and contrast
levels based on eye movements. Secondly, as presented in this paper, the
outcome can be a reaction time. In previously published papers, we showed
[24]
development over age of orienting responses to cartoons. Small variations
in response times were found in younger as well as older children (on average
RTF-values of ~200 ms), while data presented in the present paper show large
variability in response times in children between 0-3+ years old and relatively
high RTF-values. This may relate to the fact that the cartoons are high contrast
and colourful images whereas the coherent motion stimuli are low contrast and
complex of shape. Recent theories argue that the brain constructs an internal
[25-27]
saliency map based on features such as colour, form and movement.
A salience map represents the most conspicuous area in a visual scene. It is
suggested that a saliency map guides visual attention and triggers orienting
behavioural responses. Presumably, the cartoon stimulus has a much higher
saliency compared to the coherent motion stimuli, explaining the relatively
fast orienting responses to cartoons. The more delayed reaction times to the
coherent motion stimuli may reflect the difficulty of processing these types of
stimuli in the tested children.
The RDK with coherent diagonal motion had the largest area in which dots
moved unidirectional and only a small portion of dots that moved in the op-
posite direction. For this stimulus, we found a high number of edge fixations
compared to both other stimuli. We suggest that the one edge of motion
discontinuity in this RDK with coherent diagonal motion presumably repre-
sents the most conspicuous area of this stimulus. This edge might trigger
comparable visual attention in young as well as elderly children, resulting in
edge fixations in 64% of cases. In the RDK with coherent vertical motion, two
edges of motion discontinuity exist. Here, the most conspicuous area may not
be a line, but a large vertical block of rectangular shape. Presumably, visual
attention is now drawn to the centre of this block rather than to its edges,
resulting in only 7% of edge fixations.
We believe that quantification of typical orienting behaviour in childhood
represents a reference of neurological development. Future studies may as-
sess abnormal development and diagnosis of visual processing dysfunctions in
a case control experiment. The children that have our special attention include
(very) preterm born children with a high risk of having visual processing dys-
functions and children diagnosed with a loss of visual function without damage
to the ocular structures. The latter group is diagnosed with cerebral visual
impairment (CVI), a very broad diagnosis of exclusion and based on ana-
[28]
tomical landmarks. With the current eye tracker method, the wide range
of visual impairments that might occur in these children may be quantified in
terms of functional deficits. Testing the effectiveness of processing different
types of e.g. contrasts, colours and form and motion coherences enables the
construction of a unique visual profile per child. Such profile could support
specific rehabilitation or visual stimulation programs in children with CVI.
Conclusions
In the present paper we showed that orienting responses to three types of

coherent motion had a development over age. Large variability in reaction
times was found in children between 0-3+ years old. The fastest reaction
times with small variability were found in children between 8-12+ years, sug-
gesting adult-like responses around this age. Measurement of coherent motion 8
sensitivity in relation to age may provide valuable information of processing
motion in patients with visual processing dysfunctions and typically developing
children.
Acknowledgement
The authors would like to thank the children and their parents for their coop-
eration.
Declaration of Interest
This research was supported by a grant from the Novum Foundation; a non-
profit organisation providing financial support to (research) projects that
improve the quality of life of individuals with a visual impairment (www.
stichtingnovum.org).
170 Chapter 8
References
2. Rodman HR, Gross CG, Albright TD. Afferent basis of visual response properties in
area MT of the macaque. II. Effects of superior colliculus removal. J Neuroscience
1990;10:1154-64.
3. Merigan WH, Maunsell JH. How parallel are the primate visual pathways? Annual
Review of Neuroscience 1993;16:369-402.
4. Ffytche DH, Skidmore BD, Zeki S. Motion-from-hue activates area V5 of human
visual cortex. Proc Biol Sci 1995;260:353-8.
5. Goodale MA, Milner AD. Separate visual pathways for perception and action.
Trends in Neurosciences 1992;15:20-5.
6. Williams DW, Sekuler R. Coherent global motion percepts from stochastic local
motions. Vision Research 1984;24:55-62.
7. Wattam-Bell J. Coherence thresholds for discrimination of motion direction in
infants. Vision Research 1994;34:877-83.
8. Kremenitzer JP, Vaughan HG, Kurtzberg D, Dowling K. Smooth pursuit eye move-
ments in the newborn infant. Child Development 1979;50:442-8.
9. Wattam-Bell J. The development of visual motion processing. In: Vital-Durand F,
Atkinson J, Braddick OJ, editors. Infant Vision. Oxford: Oxford University Press;
1996. p 79-94.
10. Mason AJ, Braddick OJ, Wattam-Bell J. Motion coherence threshold in infants
– different tasks identify at least two distinct motion systems. Vision Research
2003;43:1149-57.
11. Johnson SP, Mason U. Perception of kinetic illusory contours by two-month-old
infants. Child Development 2002;73:22-34.
12. Pel JJ, Manders JC, van der Steen J. Assessment of visual orienting behaviour in
young children using remote eye tracking: methodology and reliability. J Neurosci
Methods 2010;189:252-6.
13. Huk AC, Ress D, Heeger DJ. Neuronal basis of the motion aftereffect reconsidered.
Neuron 2001;32:161-72.
14. Rees G, Friston K, Koch C. A direct quantitative relationship between the functional
properties of human and macaque V5. Nature Neuroscience 2000;3:716-23.
15. Stiers P, Peeters R, Lagae L, van Hecke P, Sunaert S. Mapping multiple visual
areas in the human brain with a short fMRI sequence. Neuroimage 2006;29:74-
89.
16. Tootell RB, Reppas JB, Dale AM, Look RB, Sereno MI, Malach R, Brady TJ, Rosen
BR. Visual motion aftereffect in human cortical area MT revealed by functional
magnetic resonance imaging. Nature 1995;375:139-41.
17. Watson JD, Myers R, Frackowiak RS, Hajnal JV, Woods RP, Mazziotta JC, Shipp
S, Zeki S. Area V5 of the human brain: evidence from a combined study using
positron emission tomography and magnetic resonance imaging. Cerebral Cortex
1993;3:79-94.
18. Duffy CJ, Wurtz RH. Sensitivity of MST neurons to optic flow stimuli. I. A con-
tinuum of response selectivity to large-field stimuli. Journal of Neurophysiology
1991a;65:1329-45.
19. Duffy CJ, Wurtz RH. Sensitivity of MST neurons to optic flow stimuli. II. Mecha-
nisms of response selectivity revealed by small-field stimuli. Journal of Neuro-
physiology 1991b;65:1346-59.
20. Dukelow SP, DeSouza JF, Culham JC, van den Berg AV, Menon RS, T. Vilis T.
Distinguishing subregions of the human MT+ complex using visual fields and
pursuit eye movements. Journal of Neurophysiology 2001;86:1991-2000.
21. Smith AT, Wall MB, Williams AL, Singh KD. Sensitivity to optic flow in human corti-
cal areas MT and MST. The European Journal of Neuroscience 2006;23:561-9.
22. Wall MB, Lingnau A, Ashida H, Smith AT. Selective visual responses to expansion
and rotation in the human MT complex revealed by functional magnetic resonance
imaging adaptation. The European Journal of Neuroscience 2008;27:2747-57.
23. Gunn A, Cory E, Atkinson J, Braddick O, Wattam-Bell J, Guzzetta A, Cioni G.
Dorsal and ventral stream sensitivity in normal development and hemiplegia.
Neuroreport 2002;13:843-7.
24. Braddick O, Atkinson J, Wattam-Bell J. Normal and anomalous development of
visual motion processing: Motion coherence and ‘dorsal-stream vulnerability.
Neuropsychologia 2003;41:1769-84.
25. Pel JJ, van der Does JM, Boot FH, de Faber TJ, van der Steen-Kant SP, Willemsen
SP, van der Steen J. Effects of visual processing and con-genital nystagmus on
visually guided ocular motor behaviour. Developmental Medicine & Child Neurol-
ogy 2011;53:344-9.
26. Itti L, Koch C. A saliency-based search mechanism for overt and covert shifts of
visual attention. Vision Research 2000;40:1489-506.
8
27. Koch C, Ullman S. Shifts in Selective Visual Attention: Towards the Underlying
Neural Circuitry. Human Neurobiology 1985;4:219-27.
28. Treue S. Visual attention: the where, what, how and why of saliency. Curr Opin
Neurobiol 2003;13:428-32.
29. Boot FH, Pel JJ, van der Steen J, Evenhuis HM. Cerebral Visual Impairment: Which
perceptive visual dysfunctions can be expected in children with brain damage? A
systematic review. Research in Developmental Disabilities 2010;31:1149-59.
Chapter
Chapter 9
9
Gestalt perception in children

with visual impairments: item-
specific performance and looking
behaviour
van der Zee, Y., Kooiker, M., Talamante Ojeda, M., & Pel J. Gestalt
perception in children with visual impairments: item-specific
performance and looking behavior. Accepted for publication in
Developmental Neuropsychology
174 Chapter 9
Abstract
Visual closure is the ability to visualize a complete whole when an incom-

plete picture is presented. The aim of the present study was to investigate
the Kaufman Gestalt closure task in children with cerebral and ocular visual
impairments. Looking behaviour was assessed by an eye tracker system to
quantify the number and duration of fixations. We found that children with
visual impairments due to cerebral damage show weaker Gestalt perception
and had different looking patterns than children with ocular or without visual
impairments. Children with brain damage performed significantly worse on the
animate items than the group without brain damage.
Gestalt perception and eye tracking 175
Introduction
Neuropsychology aims to understand how behaviour and cognition are in-

fluenced by brain functioning. Especially the effect of mental processes on
behaviour, such as attention, memory, and perception, are assessed in stan-
dardized neuropsychological tests. During such tests the processing of visual
information plays an important role, because neuropsychological tests often
consist of and employ visual representations. Dependent on the task at hand,
a participant can be asked to discriminate details, to name forms and objects
and to interpret relations within and between pictures. Thus, besides under-
standing a given instruction and being able to give a verbal or motor response,
participants should be able to create a meaningful representation of the world
from visual elements. An underlying premise is that simple visual elements
are integrated into a whole, also known as a Gestalt, and that the given rep-
resentations are visually organized through the same principles, or laws, of
[1, 2]
perceptual grouping. Examples of Max Wertheimer’s classical grouping
[2, 3]
principals are proximity and visual closure. The proximity principle means
that elements that are closer to each other are more likely to be perceptually
grouped than more distant elements. The visual closure principle is the ability
to visualize a complete whole when presented with incomplete information or
a partial picture, thus elements that form a closed figure tend to be grouped
[3]
together. It was shown that visual closure is essential in the identification
[4] [5] 9
of incomplete drawings, of global letters and of hierarchically organized
[6]
figures.
In adults without neurological impairment, the bilateral temporo-parietal
junction (TPJ), the bilateral anterior cingulate cortex (ACC), and the pre-
cuneus (PC), primarily on the left side, became activated in a recognition
[6]
task. Presumably, perceptual grouping depends on the integrity of a large
brain network and brain damage may result in poorer performance. Indeed,
[4]
damage in the right occipital lobe and the medial parietal lobe, i.e. the pre-
[5]
cuneus (PC) resulted in weaker recognition performance. In adults, Gestalt
perception, visual illusions or hierarchical letters are used to test aspects of
the principle grouping. In children, the subtest Gestalt Closure of the Kaufman
Assessment Battery for Children second edition (KABC-II) is frequently used.
This test is an object recognition task with incomplete representations and
is a subtest of the Simultaneous Processing scale. Despite the importance
of these aspects of visual information processing, not many studies have
addressed Gestalt perception performance in visually impaired children. A
population-based cohort study of very preterm children showed that at an age
176 Chapter 9
of 5 years, children born premature had lower scores than full-term controls
[7]
on multiple scales, including the simultaneous processing scale. Exclusion
of participants with cerebral palsy (9%), an indication of brain damage, visual
impairments (1% of the population had low vision, a visual acuity <3/10) and
[7]
hearing impairment (<1%) had no influence on the results. These data
suggested that prematurity combined with brain damage did not explain dif-
ferences in performances. Yet, the number of children with impairments was
too small to draw definite conclusions.
As stated, perceptual grouping of simple visual elements (e.g., lines and
arcs) into simple forms and shapes helps to create meaningful representations
of the world. In this process, visual attention plays a very important role. The
frontoparietal network is essential for both bottom-up attention (externally
driven and involving stimuli that are salient because of their inherent proper-
ties relative to the background) and top-down attention (internally driven on
[8]
prior knowledge, willful plans and current goals). Based on event-related
potential patterns, it was shown that visual objects organized according to
Gestalt principles were prioritized over competing non-Gestalt stimuli in
[9]
a bottom-up and automatic fashion. However, these electrophysiological
responses (~150 ms after stimulus onset) only relate to the first steps of
perceptual grouping of simple visual elements, and are not directly as-
sociated with the process of perceptual grouping that results in perception.
Alternatively, visual attention can be quantified from infancy onwards using
visually-guided eye movement responses. Differences in interactions between
eye movement responses and salience in target areas have been found in
children at risk of cerebral and ocular visual impairments compared to typi-
[10]
cally developing children. Specifically, in children at risk of cerebral visual
impairments (CVI), higher reaction times to large and high contrast stimuli
[11]
were found in combination with poorer fixation stability. These children
might have problems with bottom-up regulated visual attention orienting, not
only in visually-guided response tasks but also in visual recognition tasks, like
Gestalt Closure.
The aim of the present study was to investigate the role of visual attention
orienting in the Kaufman Gestalt closure task in children with cerebral and
ocular visual impairments. Digitized Gestalt items were shown one-by-one
on a monitor with an integrated eye tracker. Each participant was asked to
name the complete image that each item represented. For analysis purposes,
we grouped the items in animate (i.e. humans and animals) and inanimate
objects (i.e. everything else). For each item, eye movement responses were
collected to calculate the number of gaze shifts and fixation durations. We
hypothesized that children with visual impairments score lower on Gestalt

Closure performance than age-matched American controls and that gaze fixa-
tion duration on visual elements of the Gestalt (area of interest) in children at
risk of CVI is longer than in children with OVI.
Methods
Participants
For this study, a database was used that consisted of 122 clients of a rehabili-
tation centre for visually impaired and blind people. The experimental proce-
dures were approved by the local Medical Ethical Committee (MEC 2012-097)
and adhered to the tenets of the declaration of Helsinki (2013). The parents of
all children gave written informed consent. Of the total cohort of 122 patients,
72 patients underwent the Gestalt Closure test. The remaining 51 patients had
either a developmental level below 3 years old, or the task was not performed
due to a time limitation. The chronological age of the selected patients (42
male, 30 female) ranged from 4y6m to 13y8m (M = 9y8m, SD = 2y5m). Four
patients had syndromes: one CHARGE-syndrome, one Down syndrome, one
Noonan syndrome, one Sotos syndrome. Forty-four (60%) had nystagmus as
main diagnosis or as a symptom of another diagnosis, like albinism. Based
on the information in the medical records about ocular abnormalities, brain
9
damage, presence of paresis and visual diagnosis, the children were divided
in three different groups: ocular visual impairments (OVI), ocular and cerebral
visual impairment (OCVI) or cerebral visual impairments (CVI).
Group 1: Ocular visual impairment (OVI)

Of the 72 patients, 38 (23 males, 15 females) were included in group 1.
Twenty-three patients had iris or retinal abnormalities: iris coloboma (n =
1); aniridia combined with lens luxation and glaucoma (n = 1); uveitis and
retinopathy of prematurity (ROP; n = 1); ocular or oculocutaneous albinism
(n = 10); cone/rod dysfunction (n = 6); CMV retinitis (n = 1); retinitis pig-
mentosa (n = 1); retinoschisis (n = 1). Ten patients had lens abnormalities:
hypermetropia (n = 2); high myopia (n = 2); cataract (n = 6). Four patients
had congenital nystagmus and the cause of visual problems was unknown in
two patients. A total of twenty-three patients had nystagmus, twelve patients
had strabismus and twenty-one wore glasses. Visual acuity ranged from 0.05
to 0.80 (Snellen equivalent). Thirty-one patients had a visual acuity equal to
or lower than 0.3 (cut-off for low vision), of whom eight could be considered
178 Chapter 9
legally blind (visual acuity ≤0.1). Of the remaining seven patients six had a
subnormal visual acuity (< 0.8) and one had a normal visual acuity (≥ 0.8).
Behavioural problems were mentioned in nine patients: 3 had no diagnosis, one
had an autism spectrum disorder (ASD), two had attention deficit hyperactiv-
ity disorder (ADHD), one had ASD and ADHD, one had ADHD and oppositional
defiant disorder (ODD), one had an attachment disorder. In eleven patients
no estimation of the cognitive level was present. Of the remaining patients
three had a mild cognitive impairment (50 < TIQ < 70), four functioned at a
borderline cognitive level (70 < TIQ < 89), six at a below average level (70 <
TIQ < 89) and seven at an average level (TIQ ≥ 90). The chronological age of
the OVI group ranged from 4y8m to 13y7m (M = 9y9m, SD = 2y5m).
Group 2: Ocular and cerebral visual impairment (OCVI)

Fourteen patients (7 males, 7 females) were included in group 2. Eight had
ocular abnormalities and brain damage. Six had no brain damage but were
included based on the clinical signs and diagnosis of CVI, in the addition to
the presence of ocular abnormalities. In Table 1 the etiological categories for
the brain damage and the distribution of cognitive level of the CVI and OCVI
group are presented. Present ocular abnormalities were: nystagmus (n = 7);
myopia (n = 1); ocular albinism (n = 2); optic nerve atrophy (n = 1); cataract
(n = 1) and retinitis pigmentosa (n = 2). Seven children had strabismus and
eleven wore glasses. Visual acuity ranged from 0.10 to 0.60. Ten were visually
impaired (visual acuity: ≤ 0.3), one could be considered legally blind (visual
acuity: ≤ 0.1). The other four had a subnormal visual acuity. The (estimated)
level of cognitive functioning ranged from a mild intellectual disability (50
< TIQ < 70) to an average level (90 < TIQ < 110). Behavioural problems
were mentioned in four patients: one had no diagnosis, three had ADHD.
The chronological age of the OCVI group ranged from 5y11m to 13y7m (M =
9y6m, SD = 2y7m).
Group 3: Cerebral visual impairment (CVI)

Twenty patients (12 males, 8 females) were included in group 3. Some patients
had multiple causes. Two patients had no objectified brain damage but were
included based on the clinical diagnosis of CVI, one patient had a develop-
mental coordination disorder (DCD) and in two patients a paresis was present,
suggesting the presence of some type of brain dysfunction or damage. Of the
patients with dysgenesis of the brain one patient had septo-optic dysplasia,
a condition characterized by the underdevelopment of the optic nerve and
the absence of the septum pellucidum (midbrain structure); one had Arnold-
Table 1. Aetiology of brain damage and estimated cognitive levels in children in the groups OCVI (n
= 14) and CVI (n = 20).
OCVI CVI
n (%) n (%)
Aetiology of brain damage
No imaging data 6 (43) 5 (24)
Perinatal 8 (57) 12 (57)
Asphyxia 3 (21) 3 (15)
a
Hypoxic-ischemic encephalopathy (HIE) 2 (14) 6 (30)
Dysgenesis/agenesis 2 (14) 2 (10)
Postnatal brain damage
b
Acquired brain damage - 4 (20)
Peri- and/or postnatal
Hydrocephalus - 1 (5)
Mild atrophy 1 (7) -
Estimated cognitive level
Average (TIQ ≥ 90) 4 (29) 4 (20)
Below average (70 < TIQ < 89) 2 (14) 6 (30)
Borderline (70 < TIQ < 89) 4 (29) 3 (15)
Mild impairment (50 < TIQ < 70) 2 (14) 4 (20)
Moderate impairment (TIQ < 50) - 2 (10)
No data 2 (14) 1 (5)
a
Hypoxic-ischemic encephalopathy (HIE) includes different categories for perinatal brain damage,
i.e. Periventricular leukomalacia (PVL), Intraventricular haemorrhage (IVH), Cerebral Vascular Ac-
b
cident (CVA) or combinations of these categories; Acquired brain damage consists of different
categories, i.e. CVA, meningitis and tumour. 9
Chiari malformation, a condition of characterized by the malformation of the

cerebellum and/or brainstem. One patient had iris coloboma, as a symptom
of the CHARGE syndrome combined with hydrocephalus, and one patient had
a cone dysfunction.
Eight patients had nystagmus, fourteen had strabismus and eight wore
glasses. Visual acuity ranged from 0.05 to 1.00. Ten were visually impaired
(visual acuity ≤0.3), of whom three could be considered legally blind (visual
acuity ≤ 0.1). Of the remaining ten patients six had a subnormal visual acuity
(visual acuity < 0.8). Behavioural problems were mentioned in five patients:
four had no diagnosis, one of them had ASD. The (estimated) level of cogni-
tive functioning ranged from a moderate intellectual disability (TIQ < 50) to
an average level (90 < TIQ < 110) and cognitive level was unknown in one
patient. The chronological age of the CVI group ranged from 4y6m to 13y10m
(M = 9y8m, SD = 2y5m).
180 Chapter 9
The mean age and the distribution of visual acuity levels of the three different
groups did not differ significantly. The distributions of the estimated cognitive
2
level did significantly differ (χ (2) = 11.29, p < .01) between groups: post-hoc
analyses showed a higher mean cognitive level of the OVI group than of the
two other groups.
Procedures
The subtest Gestalt Closure of the Kaufman Assessment Battery for children
second edition (KABC-II) consisted of one example item and 37 tests items
[12]
of increasing difficulty. The items contained incomplete black figures on
a white background and represent 12 animate objects (7 animals including
a dinosaur and 5 human faces or activities), and 25 inanimate objects. The
normal clinical procedure would be to start with the example item followed by
the age appropriate items. Here, we showed all items on a 24-inch monitor
with an integrated infrared eye-tracking system (Tobii T60-XL, Tobii Corpora-
tion, Sweden). In-between each item, a neutral display was shown with a
fixation point to let the child refocus on the centre of the monitor. The system
measured gaze position of each eye separately using cornea reflection and
compensated for free head movements. Each child sat at approximately 60
cm distance from the monitor. The visual angle towards the monitor was ap-
proximately 30x24 degrees (1280 x 1024 pixels). Eye movement responses
were recorded with a sampling rate of 60 Hz and the latency of the system was
approximately 25 ms. The experiments were conducted in a quiet room with
ambient light conditions. Prior to starting the Gestalt Closure test a standard-
ized 5-point calibration procedure of both eyes was performed. Item presenta-
tion time was dependent on the child’s verbal response, i.e. the examiner
pushed the spacebar after the first answer to each item and thereby ended the
item presentation. The examiner also noted whether the answer was correct.
After 4 consecutive errors the assessment was stopped. All data were stored
off-line.
Data analysis
Performance data
Firstly, we scored items the standard way, i.e. recognition was either correct
or incorrect. The items after discontinuation were assumed to be too difficult
and therefore scored as incorrect. Raw scores, i.e. the total correct items up to
four consecutive incorrect answers, were noted and the American norm values
[12]
for the age group 4-14 years old were used to classify the performances,
based on scaled scores (M = 10; SD = 3): scaled scores 1-5 (expected: 9%)
= weak; 6-7 = below average (expected: 16%); 8-11 = average (expected:
50%); 12-13 above average (expected: 16%); 14-19 = superior (expected:
9%). Secondly, a simple item analysis was done by calculating the percentage
of patients that correctly recognized each item. This way, we were able to
check whether items increased in difficulty as reported in the original popula-
tion. For the analysis of the subgroups, i.e. the animate and inanimate items,
we calculated the percentage of correct items with respect to the total number
of items presented in that subgroup.
Eye tracking data

Gaze data was analyzed off-line using Tobii Studio 3.2 analysis software (To-
bii, Danderyd, Stockholm, Sweden). Patients with at least 40% gaze data
measured during the assessment were selected for further analysis to ensure
reliability of the eye movement responses. For each stimulus, an area of inter-
est (AOI) was defined that corresponded with the size and the location of
the Gestalt item. The eye movement patterns relative to these AOIs were
obtained to calculate 1) the mean and total time spent within the AOI and 2)
the number of times gaze went in and out of the AOI. A lower mean and/or
total time within the AOI might be the result of efficient information process-
ing, i.e. fast recognition, or caused by the presence of nystagmus, attentional
9
problems or a lack of effort.
We used IBM SPSS Statistics version 20 for statistical analyses. Non-paramet-
ric tests were used, since data were not normally distributed, even not after
data transformation. To compare groups with ordinal (number of items shown;
scaled scores) and skewed data (percentage correct of presented items),
Kruskal Wallis and Mann-Whitney U-tests were used. To test for differences
in performance while controlling for differences in cognitive level, we used
the partial Spearman correlation test. Binomial tests were used the compare
the expected percentage of weak, average and superior performers with the
actual percentages found in the patient groups. Finally, an explorative cor-
relation analysis was done between weak overall performance, visual acuity,
nystagmus and eye tracker results. p-Values ≤ .05 were considered significant.
182 Chapter 9
Results
The collection of the performance scores of one patient with albinism in the
group OVI (nOVI =37, 22 males, 15 females) failed, resulting in a total group of
seventy-one tested patients (Mage= 9y9m, SD = 2y5m; 41 males, 30 females).
Overall performance
Scaled scores
Analysis of the scaled scores showed that the distributions of scores signifi-
2
cantly differed across groups (χ (2) = 8.65, p =.01). The group CVI performed
significantly worse (Mdn = 2) than the group OVI (Mdn = 9; U = 197.5, p <
.01). Compared to the normal distribution of the scaled scores in the norm
population, half of the patients in the group CVI performed extremely weak.
A scaled score of 5 is expected in 9% of the population, yet, this scaled score
of 5 or lower was found in 9/37 (24%) of the OVI patients, 5/14 (36%) of the
OCVI patients and 12/20 (60%) CVI patients (ps < .01). Despite these weak
performances, still 12/37 (32%) of the OVI patients, (7/14) (50% of the OCVI
patients and only 12/37 (32%) of the CVI patients performed on an average
level (p = .05 and p < .01). The percentage of patients that performed at a
superior level was not higher than expected. The Spearman rho correlations
between the patient groups and 1) the scaled item scores (rs = -.31) and
2) the proportion weak performers (rs = .32) were significant (ps = .02).
Controlling for the cognitive level, group differences in scaled scores were
only near significant (rs,cognitive level = -.25, p = .07), but group differences
in the proportion of weak performers remained significant (rs,cognitive level =
.26, p = .05). This suggests that the performance of the group CVI remains
significantly weaker after correction for differences in cognitive level.
Item scores
The item scores showed a monotone decrease with increasing number of
presented items, suggesting that test difficulty increased per presented item
(Figure 1). Notably, the percentage correct animate items, especially the ani-
mal items, deviated from this general pattern. We found that the animal items
were more difficult to recognise or to name than the inanimate items. The
most difficult items were the dinosaur (item 7; 44% correct) and the elephant
(item 18; 39% correct).
general pattern. We found that the animal items were more difficult to recognise or to
name than the inanimate items. The most difficult items were the dinosaur (item 7; 44%
correct) and the elephant (item 18; 39% correct). Gestalt perception and eye tracking 183
Figure1.1.Percentage
Figure Percentageof of correct
correct answers
answers per item-group:
per item-group: inanimate
inanimate objectsobjects and animate
and animate objects, objects,
human(activity)
human (activity) and
and animals.
animals.
Analysis
Analysisof of
thethe
percentage correct
percentage answers
correct of the
answers of presented itemsitems
the presented showed that
showed
2
that significant
significant differences
differences on performances
on performances werepresent
were indeed indeed(χpresent
2
(χ (2)
(2) = 7.28, =
p=.03), see
7.28, p = .03),
also Figure see
2. The alsoCVI
group Figure
(Mdn2.=The
59.94)group CVI (Mdn
performed = 59.94)worse
significantly performed
than the group
significantly worse than
OVI (Mdn = 73.33; the group
U = 215.0, OVIThe
p=.01). (Mdn = 73.33;
difference U = 215.0,
between p = . CVI
the groups 01).and OCVI
The difference between the groups CVI and OCVI (Mdn = 67.95) and OCVI
(Mdn = 67.95) and OCVI and OVI were not significant. Significant negative correlations9
and OVI were not significant. Significant negative correlations (rs = -.34, p
(rs = -.34, p =.01; rs.cognitve level = -.29, p = .03) were found between the percentage
= .01; rs.cognitve level = -.29, p = .03) were found between the percentage
correctly recognized items and the patient groups, indicating that cognitive level does not
correctly recognized items and the patient groups, indicating that cognitive
explain the group difference found.
level does not explain the group difference found.
Gestalt perception and eye tracking | 167
Figure 2. Distribution of percent-

age correct answers given the
number presented items in the
different patient groups.
Figure 2. Distribution of percentage correct answers given the number presented items in the
different patient groups.
Animate items
Analysis of the number of animate items presented showed that the number of presented
animal items differed across groups (χ2(2) = 6.82, p =.03), suggesting differences in
184 Chapter 9
Animate items
Analysis of the number of animate items presented showed that the number
2
of presented animal items differed across groups (χ (2) = 6.82, p = .03),
suggesting differences in performances. In the groups OVI and the OCVI, only
1 patient was presented less than 7 items (the total number of animal items)
(OVI: 2 items; OCVI: 5 items). In the group CVI, 7 patients were presented
less than 7 items; less than 3 items were presented to 5 patients, less than
5 items to 1 patient and less than 6 items were presented to 3 patients. The
number of presented items was significantly different between the groups OVI
and the CVI (U = 290.0, p < .01). Analysis of the percentage correct answers
of the presented items showed that significant differences in performances
2
were indeed present (χ (2) = 13.45, p < .01), see also Figure 3. The groups
CVI (Mdn = 50.00) and OCVI (Mdn = 64.29) performed significantly worse
than the group OVI (Mdn = 85.71; U = 171.5, p < .01 and U = 158.00, p =
.03). The difference between the groups CVI and OCVI was not significant.
A significant negative correlation (rs = -.43, p <.01) was found between the
percentage correctly recognized animals and the patient groups. This indi-
cates that children with brain damage tend to perform worse than children
without brain damage (OVI) on the animal items. Importantly, this correlation
remained significant when controlling for differences in cognitive level (rs.
level 9=
168 | Chapter
cognitve -.38, p < .01).
Figure 3. Distribution of percent-

age correct answers given the
number presented animal items
(max 7 items) in the different pa-
tient groups.
Figure 3. Distribution of percentage correct answers given the number presented animal items
(max 7 items) in the different patient groups.
Inanimate items
Inanimate items
To test whether the same pattern of group differences was also present in
To test whether the same pattern of group differences was also present in recognising
recognising the we
the inanimate items, inanimate
first selecteditems, weitems
7 inanimate firstwith
selected 7 inanimate
approximately the same items with ap-
proximately the
performance score same
as the performance
7 animal score
items. We selected itemas the16,7 19,
2, 10, animal items.
20, 23 and 24. We selected
The median difference in performance between these two item sets was -1% (range -6%
to +4%). Thus, compared to the animate items, these selected inanimate items were a
bit easier to recognize. Again, the number of presented inanimate items varied from 1 to
7 in the groups OVI an CVI and 2 to 7 in the group OCVI. The distribution of the number
of presented inanimate items significantly differed across the groups (χ2(2) = 11.45, p
item 2, 10, 16, 19, 20, 23 and 24. The median difference in performance
between these two item sets was -1% (range -6% to +4%). Thus, compared
to the animate items, these selected inanimate items were a bit easier to
recognize. Again, the number of presented inanimate items varied from 1 to 7
in the groups OVI an CVI and 2 to 7 in the group OCVI. The distribution of the
number of presented inanimate items significantly differed across the groups
2
(χ (2) = 11.45, p < .01). In the group CVI, a significantly smaller number of
items was presented (Mdn = 5) than in the group OVI (Mdn = 7; U = 212.5,
p < .01). The percentage correct answers, however, to the inanimate items
was not significantly different between the groups (MdnOVI = 85.71, MdnOCVI
= 71.43 MdnCVI = 60.00; p = .13). Correlations between performance and
groups were not significant (rs = -.21, rs.cognitve level = -.19). Thus, children
with CVI tend to perform in a similar way as the children with OCVI and OVI
on inanimate items.
Gaze patterns
Group characteristics
Of the seventy-two patients forty-five patients had at least 40% gaze data. The
group OVI consisted of 20 patients (12 males, 8 females). The chronological
age ranged from 5y0m to 13y7m (M = 9y10m, SD = 2y7m). Thirteen patients
had iris or retinal abnormalities: iris coloboma (n = 1); ocular or oculocutane-
9
ous albinism (n = 9); cone/rod dysfunction (n = 2); retinitis pigmentosa (n
= 1). Three patient had lens abnormalities: high myopia (n = 1), Stickler
Syndrome); cataract (n = 2). Three patients had congenital nystagmus and
the cause of visual problems was unknown in one patient. The group included
4 patients with a weak performance on the Gestalt test.
The group OCVI consisted of 11 patients (5 males, 6 females). The chrono-
logical age ranged from 6y5m to 13y7m (M = 9y7m, SD = 2y8m). Aetiology
of brain damage was: asphyxia (n = 3), dysgenesis (n = 1), periventricular
leukomalacia (n = 1), intraventricular haemorrhage (n = 1), atrophy (n = 1).
Imaging data was not available for four patients. Present ocular abnormalities
were: nystagmus (n = 5); ocular albinism (n = 2); optic nerve atrophy (n =
1); and retinitis pigmentosa (n = 2), unknown (n = 1). The group included 3
patients with a weak performance on the Gestalt test.
The group CVI consisted of 14 patients (10 males, 4 females). The chrono-
logical age ranged from 4y6m to 13y10m (M = 10y2m, SD = 2y6m). Aetiology
of brain damage was: asphyxia (n = 3), Arnold-Chiari malformation (n =
1), periventricular leukomalacia (n = 2), cerebral vascular incident (n = 3),
186 Chapter 9
tumour (n = 1). No imaging data was present in four patients. This group
included 8 patients with a weak performance on the Gestalt test.
The mean age and the estimated cognitive level of the three different groups
did not differ significantly, see Table 2. Group differences were present in the
2 2
distributions of the visual acuity (χ (2) = 7.27, p = .03) and nystagmus (χ (2)
= 12.22, p < .01). Post-hoc analyses showed a higher visual acuity and fewer
patients with nystagmus in the group CVI than of the two other groups (ps
≤.02). Visual acuity in the groups OVI and OCVI were considered comparable.
Table 2. Group characteristics of children with at least 40% gaze data in the groups OVI (n = 20),
OCVI (n = 11) and CVI (n = 14).
OVI OCVI CVI
Characteristic n (%) n (%) n (%) p-value
Glasses 9 (45) 8 (73) 6 (43)
Strabismus 7 (35) 4 (36) 10 (71) .09, ns
Nystagmus 15 (75) 11 (100) 5 (36) <.01
Visual acuity .03
≥ 0.8 1 (5) 4 (29)
0.5-0.8 4 (20) 3 (21)
0.4-0.5 3 (27) 3 (21)
≤ 0.3 15 (75) 8 (73) 4 (29)
Cognitive level .07, ns
Average (TIQ ≥ 90) 7 (35) 4 (36) 1 (7)
Below average (70 < TIQ < 89) 2 (10) 1 (9) 4 (29)
Borderline (70 < TIQ < 89) 3 (15) 3 (27) 3 (21)
Mild impairment (50 < TIQ < 70) 2 (10) 1 (9) 3 (21)
Moderate impairment (TIQ < 50) 2 (14)
No data 6 (30) 2 (18) 1 (7)
p-values < .05 indicate statistically significant group differences
Item characteristics
In one CVI patient, no eye tracking data was available during the presentation
of the 7 animal and 7 selected inanimate items. In the remaining group of
44 patients, eye tracking data during the presentation of the animal items
was present for at least 4 items in 20/20 OVI patient, 11/11 OCVI patients
and 11/13 CVI patients. To control for individual differences in the number
of items, we used mean individual outcomes for group comparisons. Overall,
no significant group differences on the number of visits and the median visit
duration of each visit within the area of interest (AOI) were found, but we
2
did find a significant difference in the total visit duration (χ (2) = 6.38, p =
.04), see Table 3. This indicates that children with CVI tend to need more time
to analyse the Gestalt to formulate an answer. Analysis of individual animal

items showed that children with CVI tend to be a bit slower on 6 of 7 items,
but differences were not significant. Of the 7 selected inanimate items, eye
tracking data of at least 4 items were available in 20/20 OVI patients, 10/11
OCVI patients and 12/13 CVI patients. Here, no significant group differences
were found, see Table 4.
Finally, explorative correlation analyses were done between weak overall
performance, visual acuity, nystagmus and eye tracker results of the ani-
mal items within the patient groups. The correlations between weak overall
performance, visual acuity and nystagmus were not significant, except the
correlation between weak overall performance and visual acuity in the group
CVI (rs = -.63, p = .02). This suggest that weak performers within the group
CVI have lower visual acuities. Within the group OVI we found that the weak
performers tend to stay shorter within the AOI (rs = -.52, p = .02), children
with lower visual acuity visit the AOI more often (rs = -.59, p < .01), children
with higher visual acuity seem to stay longer within the AOI (rs = .43, p = .06)
and children with nystagmus need in total more time to analyse and respond
(rs = .47, p = .03). Within the group CVI we found that weak performers tend
to visit the AOI more often (rs = .63, p = .03), and need in total more time to
analyse and give a response (rs = .72, p < .01). This indicates that weak score
of these children was not the result of a lack of effort.
Table 3. Outcomes of eye tracking data (number of visits of area of interest, median visit duration 9
and total visit duration) on animal items in the groups OVI (n = 20), OCVI (n = 11) and CVI (n = 13).
OVI OCVI CVI p-value
Mean
# visits AOI 7 (1-33) 10 (3-15) 6 (2-28) ns
Median visit duration 0.7 (0.1-3.4) 0.7 (0.1-3.3) 1.3 (0.1-4.7) ns
Total visit duration 3.7 (1.0-6.7) 3.8 (0.5-12.0) 8.4 (0.2-14.4) .04
Butterfly
n 20 11 11
# visits AOI 4 (1-15) 2 (1-12) 3 (1-13) ns
Total visit duration 2.4 (0.1-3.5) 2.2 (0.1-5.2) 2.7 (0.8-13.1) ns
Dog
n 20 11 13
# visits AOI 3 (1-19) 2 (1-11) 2 (1-22) ns
188 Chapter 9
Table 3. Outcomes of eye tracking data (number of visits of area of interest, median visit duration
and total visit duration) on animal items in the groups OVI (n = 20), OCVI (n = 11) and CVI (n =
13). (continued)
Dinosaur
n 20 11 12
# visits AOI 8 (1-47) 7 (1-25) 5 (1-63) ns
Median visit duration 0.7 (0.1-6.2) 0.8 (0.1-11.8) 2.4 (0.3-7.2) .08
Fish
n 20 10 10
# visits AOI 3.0 (1-90) 5.5 (1-24) 7.0 (1-44) ns
Turtle
n 20 10 12
# visits AOI 10 (1-13) 10 (4-34) 8.5 (1-27) ns
Pig
n 19 10 10
# visits AOI 7.0 (1-29) 4.0 (1-17) 4.0 (1-12) ns
Elephant
n 20 10 9
# visits AOI 7 (2-44) 15 (2-35) 5 (1-94) ns
AOI = area of interest. p-values indicate statistically significant differences in eye tracking param-
eters between patient groups.
Table 4. Outcomes of eye tracking data (number of visits of area of interest, median visit duration
and total visit duration) of the seven selected object items in the groups OVI (n = 20), OCVI (n =
11) and CVI (n = 12).
Mean
# visits AOI 6 (1-34) 5 (2-13) 6 (2-19) ns
AOI = area of interest. p-values indicate statistically significant differences in eye tracking param-
eters between patient groups.
Discussion
Neuropsychological assessment in children often includes subtests that exam-

ine the ability to integrate visual elements into a (meaningful) whole, i.e. per-
ceive gestalts. This is the first study to address Gestalt perception in visually
impaired children. This study about performances and looking behaviour on a
Gestalt Closure test (KABC-II) in groups of children with visual impairments,
i.e. children with brain damage or clinical signs of cerebral visual impairment
(CVI), children with ocular abnormalities and brain damage (OCVI) and chil-
dren ocular abnormalities (OVI), shows that children with visual impairments
more often performed weak than age-matched controls of the American norm
population, and that children with brain damage performed worse than children
without brain damage. The proportion of weak performers in the groups CVI
(60%) and OCVI (36%) was significantly higher than that in the group OVI
(24%). Differences remained significant even after controlling for differences
in cognitive level. Eye movement patterns revealed that weak performance on
animate items in the CVI group was associated with longer fixations and more
re-fixations.
Literature on Gestalt performance in children, including fragmented picture
recognition, hierarchical figures and embedded figures, covers many condi-
[13] [14, 15]
tions, such as autism spectrum disorder (ASD), schizophrenia and
[16]
intellectual disability syndromes. These reports may provide insights into
9
mechanisms and networks underlying our current results. For example, our
finding that children with a (partly) cerebral cause for their visual impairment
more often perform weak on Gestalt Closure is related to findings in children
with brain damage in general, who were found to have worse visuomotor
Gestalt performance (i.e. Bender Gestalt test) than children without brain
[17]
damage. More specific, PVL is an important underlying cause of CVI in
children, and children with PVL showed poor results on a Closure subtest of a
[18]
visual perception battery.
Neuroimaging studies showed that performance in visual closure tasks is
associated with activity in object-related processing areas, including the lat-
[19, 20]
eral occipital cortex and inferior temporal areas. For example, patients’
impaired perceptual closure performance was associated with a diminished
closure-related ERP negativity, of which the neural source has been attributed
[21]
to visual association areas. Furthermore, two studies on Gestalt perception
in schizophrenia patients showed a relation of reduced neural connectivity in
visual cortical areas and impaired perceptual closure. Although no studies
about Gestalt performance in visually impaired children could be found, one
190 Chapter 9
study showed that early visual deprivation in patients with congenital cataract
aged 9-23 years was related to impairments in holistic (face) processing.
[22]
Our current results in a more general population of children with visual
impairments confirms that in children who have experienced some sort of
visual deprivation early in life, holistic visual processing in general seems to
be affected.
Notably, children with brain damage (groups CVI and OCVI) performed
significantly worse on the animate items than the group without brain damage
(OVI). Differences between groups on the inanimate items were smaller and
non-significant, suggesting that the overall difference can mainly be attributed
to the weak performance on the animal items. The gaze data collected with
the eye tracker reveals that the CVI children with weak performances fixate
these Gestalt items longer and even re-fixate them more often. This suggests
that this difference in performance may not attributed to a lack of effort.
Further analysis on the 7 animal items, items earlier in the test than the 7
selected object items but with a comparable difficulty level, shows that weak
performers within the group OVI spend less time within the area of interest
(AOI) per visit. There was no relation with visual acuity or nystagmus, which
suggests there is another reason why they perform weak. Maybe children with
OVI have learned other strategies, for example to guess more often to save
time and energy or to perform as fast as children without a visual impairment.
In order to study this hypothesis a control group should be added, and reac-
tion times and performances should be compared.
It has been shown that processing of either animate or inanimate pictures
can be selectively impaired in patients with damage to anterior visual areas.
[23-25]
More recently, fMRI studies found clusters of voxel population vectors
that were associated with animate and inanimate categories located anterior
[26, 27]
to retinotopic visual areas. However, in these studies, images of hun-
dreds of well-defined objects were presented to the subjects. In the present
study, visual processing of each item involved perceptually grouping of closely
projected elements to visualize a complete object, even when presented
incomplete information. To our surprise, we found a significantly worse perfor-
mance for naming the animate items in children with confirmed or suspected
brain damage. We could not find any support for our finding in literature. At
this point, we do not know if the differences we found between animate and
inanimate object recognition in children with cerebral visual impairments are
related to possible damage to the associated processing areas in the brain,
[26]
e.g. the occipital face area (OFA) and the extra-striate body area (EBA).
The available imaging data lack the detailed information that is required to
obtain insight in these specific regions.
This study has some limitations that need to be addressed. Firstly, the
Kaufman Gestalt closure task contains a fixed number of 37 items. We applied
the classic stopping rule, i.e. we stopped testing, when a subject gave 4 con-
secutive incorrect answers. As a result, the total number of tested items was
different between the subjects. For the between group comparisons per item,
we were able to control for this by calculating the correct items with respect
to the total number of items presented within each subject. The comparison of
the performance scores between the animate and inanimate items, however,
inevitably led to two different subgroups of patients. Secondly, in a majority of
the patients, only the global cognitive functioning was scored using total IQ.
We are aware of the fact that more exact data on verbal and non-verbal IQ is
preferable to control for differences in developmental level between groups.
Thirdly, the performance of the children in the present study was compared
with American norm group. We did investigate any cultural differences. Some
of the items were typical American illustrations, such as a land map of the US.
These items were not judged too strictly – and were not selected for further
item analysis. Fourthly, we did not check after the Gestalt closure test whether
children could name standard drawings of the same objects. Especially for the
animate items, this would have been a valuable control condition. Lastly, the
size of the area of interest dependent on the size of the Gestalt closure item.
9
As a result, some AOI’s were larger than others but not more than ~20%.
Presumably, this had some influence on the number of re-fixations, especially
for the small Gestalt closure items. In addition, we attempted to investigate
whether a subject fixated the different elements within the Gestalt closure
item to create an overview or that this was obtained with one central fixation.
Due to the variability in gaze behaviour, we were not able to find a consistent
pattern. In a future study, we would like to address this question using a
fixed set of Gestalt closure items of equal sizes and a higher resolution eye
tracker. Lastly, the quality of each eye tracking study depends on the quality
of the recorded gaze patterns. In our group, 45 of the 72 patients (63%) had
at least 40% valid gaze data. In the present study, the investigator was able
to keep track of the gaze tracking quality. Still, we applied this rather strict
criterion because we needed good continues gaze tracking data during item
presentation to be able to calculate the presented gaze parameters.
In the present study we presented data on Gestalt perception performance
in combination with looking behaviour in visually impaired children. Further
research is necessary to address causality of the results: did a different way
192 Chapter 9
of looking result in worse perceptual performance, or did poor perception lead

to a different looking pattern? We would like to stress that the analysis of eye
movement patterns during perception tasks adds value for the visual testing
of ‘non-verbal children’ and may provide a link to later visual training.
Conclusions
This study showed that children with visual impairments more often performed
weak on the Kaufman Gestalt closure task than chronologically age-matched
controls, and that children with brain damage performed worse than children
without brain damage. Notably, children with brain damage (groups CVI and
OCVI) performed significantly worse on the animate items than the group
without brain damage (OVI). Based on the looking patterns obtained with the
eye tracker, we can conclude that weak Gestalt performance of the children in
the CVI group is not the result of intellectual disability or a lack of effort, but
instead seems a pure perceptual dysfunction.
References
1. Jäkel, F., et al., An overview of quantitative approaches in Gestalt perception.
Vision Research, 2016. 126: p. 3-8.
2. Han, S. and G.W. Humphreys, Interactions between perceptual organization
based on Gestalt laws and those based on hierarchical processing. Perception and
Psychophysics, 1999. 61(7): p. 1287-1298.
3. Wagemans, J., et al., A century of Gestalt psychology in visual perception: I.
Perceptual grouping and figure-ground organization. Psychological Bulletin, 2012.
138(6): p. 1172-1217.
4. Russell, E.W., M.E. Hendrickson, and E. VanEaton, Verbal and figural Gestalt
Completion Tests with lateralized occipital area brain damage. Journal of Clinical
Psychology, 1988. 44(2): p. 217-225.
5. Pflugshaupt, T., et al., Bottom-up Visual Integration in the Medial Parietal Lobe.
Cerebral Cortex, 2016. 26(3): p. 943-949.
6. Huberle, E. and H.O. Karnath, The role of temporo-parietal junction (TPJ) in global
Gestalt perception. Brain Structure and Function, 2012. 217(3): p. 735-746.
7. Larroque, B., et al., Neurodevelopmental disabilities and special care of 5-year-old
children born before 33 weeks of gestation (the EPIPAGE study): a longitudinal
cohort study. Lancet, 2008. 371(9615): p. 813-820.
8. Katsuki, F. and C. Constantinidis, Bottom-up and top-down attention: different
processes and overlapping neural systems. Neuroscientist, 2014. 20(5): p. 509-
521.
9. Marini, F. and C.A. Marzi, Gestalt Perceptual Organization of Visual Stimuli Cap-
tures Attention Automatically: Electrophysiological Evidence. Frontiers in Human
Neuroscience, 2016. 10: p. 446.
10. Kooiker, M.J., et al., A Method to Quantify Visual Information Processing in Chil- 9
dren Using Eye Tracking. Journal of Visualized Experiments, 2016(113).
11. Kooiker, M., et al., Combining Visual Sensory Functions and Visuospatial Orienting
Functions in Children with Visual Pathology: A Longitudinal Study. Accepted for
publication in Brain & Development, 2018.
12. Kaufman, A. and N. Kaufman, Kaufman Assessment Battery for Children: technical
manual. second edition ed. 2004, Circle Pines, MN: American Guidance Service.
13. Bölte, S., et al., Gestalt perception and local-global processing in high-functioning
autism. Journal of Autism and Developmental Disorders, 2007. 37(8): p. 1493-
1504.
14. Uhlhaas, P.J., et al., Perceptual grouping in disorganized schizophrenia. Psychiatry
Research, 2006. 145(2-3): p. 105-117.
15. van de Ven, V., et al., Reduced intrinsic visual cortical connectivity is associated
with impaired perceptual closure in schizophrenia. Neuroimage Clinical, 2017. 15:
p. 45-52.
16. Hodapp, R.M., et al., K-ABC profiles in children with fragile X syndrome, Down
syndrome, and nonspecific mental retardation. American Journal of Mental Retar-
dation, 1992. 97(1): p. 39-46.
17. Koppitz, E.M., Diagnosing brain damage in young children with the Bender Gestalt
test. Journal of Consulting Psychology, 1962. 26: p. 541-546.
194 Chapter 9
18. Fazzi, E., et al., Visual-perceptual impairment in children with periventricular

leukomalacia. Brain and Development, 2004. 26(8): p. 506-512.
19. Ploran, E.J., et al., Evidence accumulation and the moment of recognition: dis-
sociating perceptual recognition processes using fMRI. Journal of Neuroscience,
2007. 27(44): p. 11912-11924.
20. Doniger, G.M., et al., Activation timecourse of ventral visual stream object-rec-
ognition areas: high density electrical mapping of perceptual closure processes.
Journal of Cognitive Neuroscience, 2000. 12(4): p. 615-621.
21. Doniger, G.M., et al., Impaired visual object recognition and dorsal/ventral stream
interaction in schizophrenia. Archives of General Psychiatry, 2002. 59(11): p.
1011-1020.
22. Le Grand, R., et al., Impairment in holistic face processing following early visual
deprivation. Psychological Science, 2004. 15(11): p. 762-768.
23. Hillis, A.E. and A. Caramazza, Mechanisms for accessing lexical representations
for output: evidence from a category-specific semantic deficit. Brain and Lan-
guage, 1991. 40(1): p. 106-144.
24. Caramazza, A. and J.R. Shelton, Domain-specific knowledge systems in the brain
the animate-inanimate distinction. Journal of Cognitive Neuroscience, 1998.
10(1): p. 1-34.
25. Warrington, E.K. and T. Shallice, Category specific semantic impairments. Brain,
1984. 107 ( Pt 3): p. 829-854.
26. Naselaris, T., D.E. Stansbury, and J.L. Gallant, Cortical representation of animate
and inanimate objects in complex natural scenes. Journal of physiology, 2012.
106(5-6): p. 239-249.
27. Kriegeskorte, N., et al., Matching categorical object representations in inferior
temporal cortex of man and monkey. Neuron, 2008. 60(6): p. 1126-1141.
Chapter
Chapter 10
10
General discussion
General discussion 197
The overall aim of this study, supported by Royal Dutch Visio, Centre of Ex-
pertise for blind and partially sighted people, was to provide new information
to advance efficient evidence-based diagnostic procedures for cerebral visual
impairment in children, in order to improve detection and neuropsychological
diagnosis in clinical practice.
We first investigated the potential of the crowding ratio for early detection
of cerebral visual impairment on different levels of the care chain. Next, we
specifically focussed on aspects of dorsal stream dysfunction (motion per-
ception, object recognition, visual attention, visuoconstructive functioning),
investigating the applicability and shortcomings of available quantitative neu-
ropsychological tests for their diagnosis and detection. We further performed
first epidemiological explorations of dorsal stream dysfunctioning in children.
Because neuropsychological testing is only reliably feasible from the age of
4 years onwards, we included children aged 4 to 7 years into the study, both
children at risk for cerebral visual impairment because of (suspected) brain
damage and a control group of typically developing school children.
For younger children and children with disabilities, a joint collaboration with
the Neuroscience department was initiated for the development of a new
diagnostic method based on the assessment of (reflexive) eye movements
towards visual stimuli using remote eye tracking. Here, we performed explor-
atory data analysis on normal gaze patterns to motion coherent stimuli and to
items that require the ability to visually integrate parts of the representation
of objects into a whole (Gestalt perception) by quantification of reaction times
and fixation qualities.
10
Principal findings
• The crowding ratio can be used as an easily applicable detection method

with sufficient sensitivity and specificity to distinguish children at risk for
cerebral visual dysfunctions (children with brain pathology) from children
with ocular pathology only, in groups that are referred for ophthalmological
or specialised low vision assessment.
• However, the presence of crowding cannot be used to directly screen for
cerebral visual dysfunction, because our analysis indicates that abnormal
crowding is an independent impairment, which can co-occur with percep-
tual problems, especially in children with brain damage.
198 Chapter 10
• For screening in the unselected children’s population, as in youth health-

care and paediatrics, the currently applied crowded visual acuity remains
the best diagnostic measure.
• Normal limits for motion perception tasks can be based on chronological
age in typically developing children, but in children with (suspected) brain
damage, results should be evaluated against performance or developmen-
tal age. Based on chronological age, motion perception problems are likely
to be overestimated in this group.
• Because a bottom effect was found for the biological motion task, this task
is considered too difficult for children with a developmental age of 4 to 7
years. The remarks of several children that the phase-scrambled point-light
figure looked like a dancing person, suggests that top-down processes like
expectations might have influenced the results and makes the construct
validity of the task questionable.
• In children with (suspected) brain damage, motion perception problems
are specifically found for global motion and motion-defined form, but less
[1]
often than has been reported for very low birth-weight children and
prematurely born children born preterm and with mild periventricular brain
[2]
damage. Motion speed problems are scarce in this group: we hardly
found any abnormalities. This is partially due to missing data as a result of
problems with the long total test time.
• Object recognition problems are associated with weak motion perception
and visual attention, but not with weak visuoconstructive functioning.
• Dorsal stream problems (motion perception, object recognition, visual
attention, visuoconstructive functioning) are rather specific and hetero-
geneous than general. We conclude that, if in individual children a dorsal
stream function is found to be impaired, it cannot be predicted whether and
which other dorsal stream aspects may be affected, too.
• Quantitative assessment of orienting responses towards visual stimuli using
[3]
remote eye tracking is finding its acceptance in Dutch low vision centres.
The age dependent norm references presented in this study enables the
quick and easy objective identification of motion perception problems in
children at all age groups.
• The performances on Gestalt perception are weak in children with visual
impairments. Specifically, children with brain damage had poor abilities
in recognising the animate items compared to the children without brain
damage. The eye tracking results add valuable data to grade the perfor-
mance in terms of effort and task efficiency.
Improvement of clinical practice

To what extent can findings of this study already be implemented in clinical
practice of expert low vision services and the chain of care suppliers before
them? Our research mainly focussed on the diagnostic process within spe-
cialised low vision services, assuming that children at risk for cerebral visual
impairment are referred. Among Dutch professional disciplines involved in
the care supply chain (youth healthcare specialists, paediatricians, paediatric
neurologists, rehabilitation specialists, (paediatric) ophthalmologists, general
practitioners, intellectual disability physicians, orthoptists, optometrists, opti-
cians), different guidelines for the detection of visual problems are in use.
Most of these focus on the detection of ocular abnormalities, impairment of
[4-10]
visual acuity, strabismus and amblyopia (Table in Appendix), but cerebral
visual impairments are mentioned in the guidelines for youth healthcare spe-
[6] [9]
cialists and ophthalmologists.
In the youth healthcare guideline, prematurely born children, children with
cerebral palsy, children with hydrocephalus or perinatal asphyxia and chil-
[6]
dren with intellectual disabilities are considered the primary risk groups.
However, the awareness of cerebral visual impairments is still relatively low
among professionals involved in early detection and referral, as has already
been pointed out in 2007 by a multidisciplinary Dutch expert working party
[11]
on early detection of cerebral visual impairment. It took until 2016 before
the development of a Dutch multidisciplinary guideline, specifically aimed at
cerebral visual impairment in children, was initiated. This has now (September
2018) resulted in a well-designed and complete draft guideline, nearly ready
for publication, including all disciplines that are involved in the care chain from 10
early detection to specialist low vision services.
Table 1 shows, that not only the heterogeneous aetiology and manifestations
of cerebral visual dysfunctions and the lack of consciousness of professionals,
but also other factors, such as a lack of valid screening instruments and diag-
nostic tests, limited resources, and current eligibility criteria for specialist low
vision services, may impede timely detection, consistent referral, and early or
long-term intervention or habilitation.
200 Chapter 10
Table 1. Impeding factors for detection, habilitation and intervention of cerebral visual impairment
General practitioner, (Child) ophthalmologist, Special services for visually
neonatologist, orthoptist, optometrist, impaired people or low
paediatrician, child optician vision services
neurologist, rehabilitation
specialist, intellectual
disability physician
o Unconsciousness of cerebral o Limited resources to assess o Lack of objective eligibility
visual impairment, especially all children at risk, specifically criteria for neuropsychological
of impairments other than low children with developmental assessment
vision delay or intellectual disabilities o Not all visual functions in
o Lack of easily applicable o Unconsciousness of possible children can be assessed with
detection instrument impairment of visual functions quantitative tests
o Assumption that assessment other than visual acuity and o Dutch eligibility criteria
of visual functions in children visual fields for long-term intervention
with an intellectual disability or o No referral of children with and habilitation and for
developmental delay is (sub)normal vision for accessibility to and refunding
- too difficult evaluation of visual perception of aids and appliances
- too big a burden for the child insufficiently include
- not useful in all children children with cerebral visual
impairment and (sub)normal
visual acuity
Recommendations for the care supply chain

So, it is of paramount importance that physicians in public youth healthcare
services, general paediatrics, general practice and intellectual disability clinics
timely identify children at risk for visual function problems and refer them for
ophthalmological evaluation. Based on our findings, described in Chapter 2,
we recommend that screening of visual acuity by youth healthcare profession-
als remains essentially the same as now. At school age, many children with
moderate and severe disabilities are not able to cooperate with regular visual
acuity testing methods. For such children, the guideline for screening of visual
[7]
functioning by youth health physicians recommends the use of LH symbols.
If LH symbols are not applicable, assessment with the Teller Acuity Cards is
suggested for young children with no or low verbal abilities but can be applied
in school children with intellectual disability, too.
[11]
Members of the 2007 expert working party, mentioned above, youth
health physicians working in special schools for children with disabilities, pae-
diatricians specialised in congenital and genetic conditions, and intellectual
disability physicians, stressed the need of an easily applicable test to detect
potential cerebral visual impairment. Therefore, in our opinion, addition of
crowding to visual screening in special schools and intellectual disability clinics
- where the risk groups are found - is worth considering, although its applica-
tion is restricted to children with mild intellectual disabilities. An evaluation of
its effectivity would be an interesting research project. If effective, this may
decrease the burden on ophthalmology clinics and low vison care.
In the Netherlands, young children in at risk groups who are checked by

paediatricians, such as children with severe intellectual and motor disabilities,
may not visit regular maternal and child healthcare centres and do not partici-
pate in regular preschool vision screening. The same working party proposed
a safety net construction for children in at risk groups, with routine orthoptic/
ophthalmological referral by the vaccinating child healthcare professional or
[11]
paediatrician at age 11 months and 4 years.
In the Netherlands, eligibility requirements for referral to specialist low vision
services, as formulated in the guideline of the Dutch Ophthalmological Soci-
[9]
ety, are followed for reimbursement decisions according to the Exceptional
Medical Expenses Act or by regular healthcare insurances. This guideline has
been updated in 2011. Whereas the former version focussed predominantly on
[12]
visual acuity and visual fields, in the updated guideline specific attention
is asked for cerebral visual impairment in the introduction of chapter 7 on
[9]
visual impairment in children. It is remarked that cerebral visual impair-
ment is insufficiently recognised in clinical practice, among others by late or
no referral by ophthalmologists to low vision rehabilitation centres. However,
no other referral criteria are given than the outline of risk groups with ‘men-
tal retardation’, low birth weight, and cerebral paresis. This lack of clear-cut
criteria is confusing and may harm the development of children with cerebral
visual impairments and a normal visual acuity, by denying them essential aids
or long-term expert care.
Therefore, we explicitly recommend addition of crowding assessment to
standard paediatric ophthalmologic and orthoptic assessments of children with
intellectual disabilities that are able to match or name symbols, because it 10
would partially solve the existing eligibility problem. Addition of crowding as-
sessment to routine diagnostics may help decide which children at risk should
be referred for expert assessment of low vision. According to the current eli-
gibility criterion for Dutch low vision services (visual acuity 0.5 or lower), only
13 out of our 26 study participants in which crowding was assessed, would
have been referred by the ophthalmologist to a low vision service, whereas a
[13]
crowding ratio of 2 or higher was found in 19 out of 26 patients. Although
these data are still limited, they indicate that referral might improve if based
on the crowding ratio.
Professionals in ophthalmologic care know and accept the concept of crowd-
ing, whereas binocular assessment of crowded acuity is common routine. To
implement and quantify crowding, an assessment of single optotype acuity
has to be added, which would take a few additional minutes. In principle, a
crowding ratio can also be assessed with other optotypes than the Cambridge
202 Chapter 10
Crowding Cards. One should be aware that, because of differences in inter-

symbol spacing and step size (for example 1/3 or 1/2 octave) that define the
letter size of consecutive lines, the norm value for crowding may be different
between tests. Nevertheless, a ratio of 2 or more always remains an alarm
sign.
Recommendations for eligibility criteria for low vision services

Not all children with cerebral visual impairments will be identified by crowding
assessment, whereas it will take time until crowding assessment has been
completely implemented in ophthalmologic routine. So, apart from our rec-
ommendation to add crowding to standard paediatric orthoptic assessments,
addition of ‘confirmed brain damage’ or ‘explicit risk of brain damage’ to the
eligibility criteria will be a large step forward. Indeed, since the update of the
[9]
ophthalmology guideline in 2011, children with (a risk of) brain damage are
now regularly referred by ophthalmologists; this should become standard in
case of doubt about the child’s visual perceptual functioning.
Epidemiological research into prevalences as well as consequences of cere-
bral visual dysfunctions in children at risk is urgently needed to support this
claim.
Recommendations for expert assessment of cerebral visual

functions in low vision centres
Crowding
In low vision services, the decision which children to select for neuropsycho-
logical assessment of cerebral visual functioning would be facilitated by adding
crowding assessment to routine orthoptic intake assessments. In recent years
this has already become more common.
Professionals should be aware that results are likely affected by test char-
acteristics, like symbol spacing (fixed or proportional) and test distance, and
patient characteristics, i.e. presence of absence of brain damage, low vision
[13-15] [15]
or nystagmus. Results of the study of Huurneman et al suggest that
testing at a long distance (5 m) with a test with a fixed symbol spacing results
in higher crowding in children with low vision and nystagmus. Children with low
vision but no nystagmus performed comparable to children with normal vision.
In our study with a chart with a proportional symbol spacing (inter symbol
[13]
spacing 50%) children with ocular abnormalities, including nystagmus,
performed comparable to children without abnormalities, whereas crowding
was higher in children with brain damage. For children with cerebral visual
impairment, testing at a distance may be more difficult than for children with
ocular abnormalities, leading to worse outcomes. This could be investigated,
too, in the research project in youth healthcare, suggested above.
[16]
Crowding problems may also have implications for reading. If all crowd-
ing ratios are size independent, as in amblyopics, critical spacing assessed
with the crowding cards can be used to estimate critical spacing for reading
[16]
(spacing between letters). Adjustment of the spacing in text according to
the individual’s critical spacing could then optimise reading rate. This could
explain why magnification is generally considered to help, because spacing as
well as letter size increase.
Neuropsychological assessment
Based on Chapter 6 and 7 we recommend that multiple aspects and func-
tions should be routinely included in neuropsychological assessment of dorsal
stream dysfunction in children at risk, because problems were found to be
mostly isolated and heterogeneous. In low vision services, use of the child´s
performance age in the interpretation of test outcomes, as recommended
and used in Chapter 5-7, is increasing, but not yet standard. We showed
in Chapter 6 how the use of chronological age leads to overdiagnosis of
[17]
motion perception dysfunction. Testing of performal IQ or expert estima-
tion of performance age should be routinely performed by psychologists and
behavioural scientists in low vision services. Efficiency of different diagnostic
strategies should be systematically evaluated.
Generally, the neuropsychologic tests applied in this study appeared valid
and applicable, but several aspects can be improved. On a group level, they 10
discriminate between children at risk and typically developing children. How-
ever, normal limits remain a bottleneck for all tasks, requiring evaluation in
larger samples. This does not necessarily hamper their current clinical ap-
plication as observational instruments: confidence intervals of normal limits
may give an indication whether a child belongs to the weakest 5 or 10 percent
performers, but more work has to be done to obtain valid reference values.
For subtests of the L94 in which children have to name objects (VIEW, De
Vos, NOISE), the influence of children’s verbal development and word fre-
quency used in their socioeconomic environment on concurrent validity has
to be investigated further, too. Current norms of the nearly 25 years old L94
should urgently be updated in new study groups.
In Dutch low vision services, quantitative neuropsychological tests of mo-
tion perception and visual search or selective visual attention, as applied
and evaluated in our study, are not yet part of routine neuropsychological
204 Chapter 10
diagnostics. At best, unspecified observational assessment of visual following

can be part of orthoptic assessments, whereas motion perception dysfunction
may be considered if the child has problems in traffic situations. In such cases
of suggestive problems in daily life, it may be relevant to include specific tests
of motion perception in the assessment.
Finally, we recommend some technical improvements to the tests. A touch
screen would facilitate the application of computerised object recognition,
visual search and motion perception tasks, because that would eliminate the
necessity of an observer response. Several adaptations would improve the
visual search task: 1. a button as a starting point because the children had to
be urged to put their hands on the table before each trial, 2. video registration
to observe typical behaviour, and 3. addition of a short response inhibition
test.
Applicability of available and new diagnostic tests

For large-scale epidemiological research of cerebral visual impairment
(prevalence, risk groups, disability, natural course), valid, precise and not too
time-consuming measures are needed, with well-evaluated reference values.
Currently, only samples with limited sizes and selected pathology have been
evaluated. To obtain large enough study populations, international collabora-
tion of research groups will be necessary. To that end, international consensus
should be reached on a limited battery of diagnostic tests to further evaluate.
In the discussion paragraph of our systematic review (Chapter 4), we have
made a first, argued selection on a basis of technical specifications as well as
psychometric properties.
We also recommend cross-sectional as well as longitudinal research into the
connection of outcomes of diagnostic tests and visual problems experienced in
daily life. The crowding ratio can be used for epidemiological research into the
prevalence of crowding, its natural development during early childhood, and if
persistent, its relationship with brain damage and amblyopia, as a basis for its
validity as an easily applicable screening method.
After further studies of its validity, the remote eye tracker method might
enable objective assessment of aspects of visual functioning in very young and
intellectually disabled children and might provide additive information that
helps to interpret classical behavioural data.
Quantitative assessment of orienting responses towards visual stimuli using
remote eye tracking may improve detection and diagnosis of visual function
[18-21]
problems, specifically in very young and difficult-to-test children. . From
2010 onwards, studies were published on the quantification of (reflexive)
orienting responses in terms of Reaction Time to Fixation (RTF), fixation

[22, 23]
durations and fixation accuracies resulting from visually-guided ori-
enting responses. Here, eye movements were introduced as a read out of
brain functioning, i.e. bottom-up driven processes. Application within a large
group of healthy children between 1-12 year of age not only showed age-
dependence of orienting responses but more importantly, differences in RTF’s
[24]
between different visual stimuli. It was suggested that RTF represents the
efficiency of processing specific visual features, such as contrast, colour, form
coherence and, most importantly, motion coherence. Indeed, in Chapter 8
we showed that with this approach, age-related processing of three different
types of coherent motion could be identified in a group of typically develop-
[25]
ing children aged 1-12 years. Thus, identification of prolonged reaction
times and fixation pattern abnormalities are potential screening markers. Its
[26]
further development, including longitudinal evaluations, has recently led to
implementation of eye tracking within low vision centres, an initiative between
[27]
the Neuroscience department of Erasmus MC and Royal Dutch Visio.
Recently, an increasing number of studies have investigated the advantages
of eye tracking based methods to test amongst others visual acuity and visual
field defects as part of visual function assessments. Still, a very limited num-
ber of studies are available that combine eye tracking with visual perception
tests. Whereas bottom-up processes are stimulus-driven, top-down processes
[28]
are related to inner motivation and cognitive cues. In Chapter 9, we
presented one of the first studies on Gestalt perception in visually impaired
children. Here, the emphasis was not on processing speed or accuracy, but on
fixation qualities as a measure for (sustained) visual attention. Our study not 10
only suggests that children with visual impairments, independent of the cause,
perform worse on Gestalt perception, but it also shows that the performances
of children with CVI are weaker than that of children with OVI. Children with
CVI perform in particular weak on the animal items. Eye tracking results sug-
gest that differences between patients groups might be explained by different
strategies: children with OVI that perform weak tended to fixate shorter,
whereas children with CVI need in total more time to perceive and respond.
Did children with OVI put in less effort than the children with CVI? The addition
of an additional eye tracker parameter, i.e. pupil diameter, might help to solve
this question, because pupil diameter is positively related to the workload: i.e.
[29]
the pupil diameter is larger in tasks with a higher workload. Studying pupil
diameter might also clarify whether animal items are really more difficult than
object items. And does it predict performances in general and at item level?
206 Chapter 10
We conclude that our study has considerably added to insights into the
value and shortcomings of internationally available diagnostic tests for dorsal
stream function problems in children, leading to specific, underpinned recom-
mendations for selection of tests and further steps in research, as well as
technical improvements of the tests, used in the current study. The eye tracker
paradigms introduced in this thesis even enable quantitative testing of motion
perception problems in very young and intellectually disabled children and
Gestalt perception in children suspected of brain damage. This new approach
will definitely add value to the existing neuropsychological assessments.
We established that dorsal stream problems in children with (suspected)
brain damage are rather isolated than generalised, so multiple aspects should
be part of neuropsychological assessments. And we evaluated crowding as a
screening method for cerebral visual dysfunction. Although a lot of work has
yet to be done before the tests are completely fit for use in clinical practice, we
could formulate first recommendations for improvement of both the referral
process towards specialised low vision services and within these services, the
selection of clients for neuropsychological assessment.
References
3. Kooiker, M., Visual information processing in children. An eye movement approach
2018, Erasmus MC: Rotterdam, The Netherlands.
4. Jans, S., et al., Landelijke Eerstelijns Samenwerkings Afspraak Onderzoek van de
pasgeborene (National primary care collaboration agreement Assessment of the
neonate). Huisarts en Wetenschap, 2008. 51(13): p. S15-S19.
5. Boere-Boonekamp, M., et al., Landelijke Eerstelijns Samenwerkings Afspraak
Visuele stoornissen bij kinderen en jongeren (National primary care collaboration
agreement Visual impairments in children and adolescents). Huisarts en Weten-
schap, 2010. 53(7): p. S11-14.
6. NVK, Evidence based richtlijn voor de initiële etiologische diagnostiek bij kinderen
met een globale ontwikkelingsachterstand /mentale retardatie (Evidence based
guideline for the initial etiological diagnostics in children with a global develop-
mental delay/mental retardation). 2005, Nederlandse Vereniging voor Kinderge-
neeskunde (Dutch Association of Pediatrics).
7. Coenen-van Vroonhoven, E.J.C., et al., Opsporing visuele stoornissen 0-19 jaar.
Eerste herziening (Identification of visual impairments. First revision), in JGZ-
richtlijn. 2010, RIVM: Centrum Jeugdgezondheid.
8. NHG, Standaard Visusklachten (Standard Visual Problems). 2015, Nederlands
Huisartsen Genootschap (Dutch College of General Practitioners).
9. van Rens, G., H. Vreeken, and R. van Nispen, Richtlijn Visusstoornissen, revalidatie
en verwijzing (Ophthalmological guideline Visual impairments, Rehabilitation and
Referral). 2011, Nederlands Oogheelkundig Gezelschap (Dutch Ophthalmological
Society).
10. Borstlap, R., et al., Een update van de multidisciplinaire richtlijn voor de medische 10
begeleiding van kinderen met Downsyndroom (An update of the multidisciplinary
guideline for medical guidance of children with Down syndrome). 2011, Neder-
landse Vereniging voor Kindergeneeskunde (Dutch Association of Pediatrics).
12. Oogheelkundige richtlijn. Verwijzing van slechtzienden en blinden. 2004, Neder-
lands Oogheelkundig Gezelschap.
13. van der Zee, Y.J., P. Stiers, and H.M. Evenhuis, Should we add visual acuity ratios
to referral criteria for potential cerebral visual impairment? J Optom, 2017. 10(2):
p. 95-103.
14. Huurneman, B., et al., A systematic review on ‘Foveal Crowding’ in visually im-
paired children and perceptual learning as a method to reduce Crowding. BMC
Ophthalmol, 2012. 12: p. 27.
15. Huurneman, B., et al., Crowding in central vision in normally sighted and visually
impaired [corrected] children aged 4 to 8 years: the influence of age and test
design. Strabismus, 2012. 20(2): p. 55-62.
208 Chapter 10
16. Levi, D.M., S. Song, and D.G. Pelli, Amblyopic reading is crowded. J Vis, 2007.
7(2): p. 21 1-17.
17. van der Zee, Y., et al., Clinical assessment of visual motion perception in children
with brain damage and the risk of overdiagnosis. Submitted.
18. Boot, F.H., et al., Factors related to impaired visual orienting behavior in children
with intellectual disabilities. Res Dev Disabil, 2012. 33(5): p. 1670-6.
19. Boot, F.H., et al., Quantification of visual orienting responses to coherent form and
motion in typically developing children aged 0-12 years. Invest Ophthalmol Vis
Sci, 2012. 53(6): p. 2708-14.
20. Boot, F.H., et al., Delayed visual orienting responses in children with develop-
mental and/or intellectual disabilities. Journal of Intellectual Disability Research,
2013. 57(12): p. 1093-1103.
21. Kooiker, M.J., et al., Quantification of visual function assessment using remote eye
tracking in children: validity and applicability. Acta Ophthalmol, 2016. 94(6): p.
599-608.
22. Boot, F.H., et al., Delayed visual orienting responses in children with developmen-
tal and/or intellectual disabilities. J Intellect Disabil Res, 2012.
23. Pel, J.J., J.C. Manders, and J. van der Steen, Assessment of visual orienting be-
haviour in young children using remote eye tracking: methodology and reliability.
J Neurosci Methods, 2010. 189(2): p. 252-6.
24. Kooiker, M.J., J. van der Steen, and J.J. Pel, Development of salience-driven and
visually-guided eye movement responses. J Vis, 2016. 16(5): p. 18.
25. Pel, J.J.M., et al., Remote eye tracking assesses age dependence processing of
coherent motion in typically-developing children. Journal of Medical Engineering
& Technology, 2013. 37(2): p. 109-115.
26. Kooiker, M., et al., Combining Visual Sensory Functions and Visuospatial Orienting
Functions in Children with Visual Pathology: A Longitudinal Study. Accepted for
publication in Brain & Development, 2018.
27. Kooiker, M.J., et al., A Method to Quantify Visual Information Processing in Chil-
dren Using Eye Tracking. J Vis Exp, 2016(113).
tention in the brain. Nat Rev Neurosci, 2002. 3(3): p. 201-15.
29. Hosseini, S.M.H., et al., Neural, physiological, and behavioral correlates of visuo-
motor cognitive load. Sci Rep, 2017. 7(1): p. 8866.
Chapter
Chapter 10:
10: Appendix
Appendix 1
1
Chapter 10: Appendix 1 211
Appendix 1. Care supply chain for children with visual impairments according to existing Dutch
guidelines for different professional disciplines
Pregnancy and birth
Complications?
No Yes
[1, 2]
General Practitioner or Midwife Pediatrician, neonatologist or child
Health screening, including inspection of the eyes neurologist
[3]
Referral to ophthalmologist within a week if: Developmental delay and brain damage:
o Suspicion of congenital ocular abnormalities, As part of diagnostic procedure: referral
specifically cataract and glaucoma for orthoptic/ophthalmologic assessment
Consultation or referral if: including strabismus, slitlamp biomicroscopy,
o Conjunctivitis ophthalmoscopy
[4]
Down Syndrome:
Referral for active orthoptic/ophthalmologic
screening at different ages
[6] [4]
Youth healthcare physician/ General Practitioner Down Syndrome Team
[5]
nurse Assessment of distance visual Ophthalmologist or orthoptist in
Preschool and school age: acuity and refractive errors if the team?
Scheduled screening of health parents or patient complain about
and development, including vision
inspection of the eyes and Referral to ophthalmologist if:
screening of distance visual o Visual acuity ≤0.2
acuity o Wearing a correction
- spherical >± 7 D or
cylindrical >± 3 D
Referral to ophthalmologist in o Monocular distance visual
case of: acuity: ≤ 6 years >0.2 and
- Fixation/following problem <1.0 6-20 years >0.2 and
- Abnormal ocular structures <1.0 and myopia and/or
- Monocular/binocular distance astigmatism or other ocular
visual acuity < 5/10 abnormalities or unrecognised
- Suspicion of cerebral visual amblyopia
impairment (risk groups)
Specialist assessment
[7]
Optometrist, Optician Ophthalmologist and orthoptist Ophthalmologist and orthoptist in
Checks of ocular condition and Detection or exclusion of ocular case of Down Syndrome:
[4]
10A1
refractive errors, prescription of abnormalities Active screening every 2 years,
glasses Referral to low vision service if specifically including cataract,
there is a request for help strabismus, amblyopia, refraction,
and one of the following criteria: nystagmus, visual acuity,
o Visual field defect: accomodation, keratoconus
- Concentrical <30°
- Hemianopsia
o Distance visual acuity:
- <0.3
- 0.3-0.5 + need special
support
o Near visual acuity:
- < 0.25
- Insufficient reading with
addition of +4
o Severe light sensitivity
Special services for visually
impaired people or low vision
services
Diagnostics, treatment and
habilitation
1. Boere-Boonekamp, M., et al., Landelijke Eerstelijns Samenwerkings Afspraak

Visuele stoornissen bij kinderen en jongeren (National primary care collaboration
agreement Visual impairments in children and adolescents). Huisarts en Weten-
schap, 2010. 53(7): p. S11-14.
2. Jans, S., et al., Landelijke Eerstelijns Samenwerkings Afspraak Onderzoek van de
pasgeborene (National primary care collaboration agreement Assessment of the
neonate). Huisarts en Wetenschap, 2008. 51(13): p. S15-S19.
3. Evidence based richtlijn voor de initiële etiologische diagnostiek bij kinderen
met een globale ontwikkelingsachterstand /mentale retardatie (Evidence based
guideline for the initial etiological diagnostics in children with a global develop-
mental delay/mental retardation). 2005, Nederlandse Vereniging voor Kinderge-
neeskunde (Dutch Association of Pediatrics).
4. Een update van de multidisciplinaire richtlijn voor de medische begeleiding van
kinderen met Downsyndroom (An update of the multidisciplinary guideline for
medical guidance of children with Down syndrome) B.e.a. (Eds), Editor. 2011,
Nederlandse Vereniging voor Kindergeneeskunde (Dutch Association of Pediat-
rics).
5. Basispakket Jeugdgezondheidszorg 0-19 jaar (Basic package youth healthcare
0-19 years). 2003, Ministerie van Volksgezondheid, Welzijn en Sport.
6. Standaard Refractiewijkingen (Standard refractive errors). 2001, Nederlands
Huisartsen Genootschap (Dutch College of General Practitioners).
7. Richtlijn Visusstoornissen, revalidatie en verwijzing (Ophthalmological guideline
Visual impairments, Rehabilitation and Referral), ed. V.H.V.N.R. Van Rens GHMB.
2011: Nederlands Oogheelkundig Gezelschap (Dutch Ophthalmological Society).
Summary
Summary
Summary 215
This study was designed to increase the knowledge on neuropsychological di-

agnosis of cerebral visual dysfunction (‘cerebral visual impairment’) in children
with brain damage or an increased risk of brain damage, as well as to evaluate
available diagnostic methods. We focused on motion perception and other
dorsal stream functions and on the age group 4 to 7 years.
In Chapter 1, a general introduction to the theme is given, with information

on current barriers to early diagnosis and intervention by specialised low vi-
sion centres in the Netherlands. Groups considered at risk are children with
cerebral palsy and/or intellectual disability, and prematurely born children.
Factors hampering epidemiological research as a basis for evidence-based
diagnostic procedures and habilitation are elucidated, the lack of easily ap-
plicable and valid quantitative diagnostic tests being one of them. We decided
to specifically address cerebral functions associated with the dorsal stream
of the visual system: object recognition under suboptimal conditions, motion
perception, visuomotor abilities and visual attention.
In Chapter 2, we focus on early recognition of children at risk for cerebral

visual problems from the perspective of different levels in the diagnostic
chain: youth healthcare, ophthalmology, and expert low vision centre. We
assessed abnormal crowding (difficulties handling complex visual information)
in 60 typically developing school children, 21 typically developing children with
ocular abnormalities, and 26 children with indications of brain damage, using
Teller Acuity Cards-ll and Cambridge Crowding Cards. The aim was to evaluate
whether the crowding ratio and the ratio between grating acuity and optotype
acuity differentiate better than visual acuity alone between children without
neurological and ocular abnormalities, children with ocular disorders only, and
children with (potential) brain damage.
Sum
For youth healthcare, subnormal crowded acuity had the best sensitivity
(76%) and specificity (70%), and for ophthalmology and low vision centres,
the crowding ratio, with a sensitivity of 67% and a specificity of 79 and 86%,
respectively.
Proportions of weak performers on cerebral visual function tests did not
significantly differ between groups with and without abnormal crowding.
We conclude that crowding and perception problems seem to be indepen-
dent but co-occurring impairments, implying that crowding cannot be applied
directly to screen for cerebral visual dysfunction. Youth health professionals
best continue screening with crowded acuity, whereas ophthalmologists and
low vision experts best add the crowding ratio to their routine diagnostics,
216 Summary
to distinguish children at risk for visual dysfunction in the context of brain

damage from children with ocular pathology.
Chapter 3 is a short introduction into the concept of motion perception,

including current insight into its representation in the brain and different
aspects of motion perception: global or coherent motion perception, percep-
tion of complex motions, perception of biological motion, and perception of
motion-defined forms. Quantitative methods to measure motion perception
are usually based on moving dots: the random dot kinematogram (RDK) is
used for perception of global or coherent motion, motion direction and motion-
defined form. For biological motion, point-light figures are mostly applied.
The advantage of such tasks above observational tasks is, that they provide
a threshold value. A variety of computerised tasks have been developed by
investigators around the world. However, at this stage, most have only been
applied in research settings.
Chapter 4 is a systematic review of the scientific literature on computerised

motion perception tasks, with the aim to judge their current applicability in
clinical practice for children. We focussed on confounding risks, task aspects,
thresholds used, sample limits or cut-off values, and performance of risk
groups compared to typically developing children. Twenty-six articles were
included, all cross-sectional, of which 22 focussed on a single motion percep-
tion aspect. Global motion has been studied most intensively, whereas smaller
numbers of studies focussed on motion-defined form and biological motion.
The confounding risk was low in 6, moderate in 10 and high in 10 of these
studies. In eight out of 16 studies with a low or moderate confounding risk,
some kind of sample limit or cut-off value was used. There is a wide variety in
task characteristics, and confidence intervals for normal limits are wide as a
result of small age-groups. Combining outcomes of studies using similar tasks,
we observed developmental trends for global motion and motion-defined form.
Eight studies addressed performance of patient groups (mean group size
21, range 8-93) compared to typically developing children. Taking confidence
intervals for normal limits into account, there is some evidence for an in-
creased risk of abnormal performance in prematurely born children, children
with autism and children with amblyopia.
We conclude that so far, computerised motion perception tasks can be used
in clinical practice as observational instruments with quantifiable outcomes,
but because of unreliable normal limits, weak performance cannot be identi-
fied reliably.
Summary 217
Chapter 5 addresses the relation between chronological age versus develop-

mental age and the outcomes of four motion perception tasks. We administered
a global motion task consisting of a random dot kinematogram (RDK) with a
variable proportion of dots coherently oscillating in a horizontal direction, a
motion-defined form task consisting of objects hidden in an RDK, a biological
motion task consisting of a human dot figure walking and a motion speed task
consisting of two identical cars filled with dots moving with different speed. We
investigated these tasks in 49 children with indications of brain damage and
60 out of 119 controls. A second aim was to construct and evaluate normal
limits for motion perception, to be used in clinical practice.
In children with brain damage, all aspects of motion perception were related
to developmental age only. In controls, no significant difference in motion
perception performance was found applying developmental or chronological
age. We conclude that in children with brain damage, developmental age
should be used as input for the norm tables, to avoid overestimation of motion
perception problems.
Data of all 119 controls were used to construct preliminary normal limits, as
well as developmental trends, based on percentile scores of different chrono-
logical age groups. A bottom effect for the biological motion task was found.
Therefore, clinical evaluation on this task is not useful. Suggestions for careful
clinical application of the outcomes of the other motion perception tasks are
given.
Chapter 6 is the report of an experimental study into motion perception of

46 children with indications of brain damage with the aim to investigate which
aspects of motion perception should be addressed, i.e. whether children with
brain damage have isolated or multiple motion perception weaknesses. We
also evaluated the clinical impact of chronological age and developmental
Sum
age as reference level when interpreting motion perception performances by
applying the above preliminary limits for normal performance (Chapter 5).
We evaluated results of a global motion task, of motion-defined task and a
motion speed task. A significantly increased risk of abnormal performance was
specifically found for global motion and motion-defined form. We primarily
found that specific and not multiple aspects of motion perception were af-
fected and the use of chronological age as reference level increases the risk of
overdiagnosis significantly.
Chapter 7 describes a study into different dorsal stream functions. It was

specifically aimed at the question whether children with indications of brain
218 Summary
damage and problems with object recognition under suboptimal conditions,

have a general or a specific dorsal stream dysfunction. Object recognition in
sub-optimal conditions was assessed with five computerised subtasks of the
L94. Motion perception was assessed with computerised motion perception
tasks for global motion, motion-defined form, biological motion, and motion
speed (see Chapter 5). Visuoconstructive skills were assessed with the Beery
VMI and the subtest mosaics of the SON-R, and visual attention with a com-
puterised visual search task. Missing normal limits were based on outcomes in
60 typically developing children aged 4-7 years. Forty-eight children at risk for
visual perception problems were divided in a group with normal and a group
th
with abnormal object recognition (score below 5 percentile). A general dorsal
stream dysfunction was considered present if a majority of participants with
th
object recognition problems would perform weak (score below 10 percentile)
on at least two additional dorsal stream functions.
Compared to participants with normal object recognition, participants with
object recognition problems performed significantly more often weak on mo-
tion perception and visual attention, but not on visuoconstructive skills. A
minority had a general dorsal stream dysfunction. We conclude that in children
with object recognition problems, dorsal stream dysfunctions seem to be
rather specific than general. Therefore, multiple aspects and functions should
be assessed in neuropsychological assessment of children at risk.
The used visual search task was first developed in 2002 by master thesis
students at Leuven Catholic University. In the Appendix, its procedure is de-
scribed, as well as our construction of normal limits in a group of 60 typically
developing school children, aged 4-7 years, divided in three age groups.
Chapter 8 is a description of a paradigm using a different, innovative diag-

nostic approach. Its development was initiated by a collaboration with the
Neuroscience department, Erasmus MC. The rationale for this collaboration
was the urgent need of objective quantitative motion perception tests ap-
plicable to children aged younger than 4 years and children as well as adults
with severe intellectual disabilities. Based on remotely tracking of reflexive
eye movements towards specific visual stimuli, shown on a computer screen,
objective outcomes can be obtained in a very short time.
The aim of the current study was, to quantify processing of different types
of coherent motion in terms of ocular motor fixation times in a group of 188
typically developing children (age 0-13 years), divided into two age groups (0-
3+ and 4-11+ years). Motion coherence was applied in three different types
of random dot kinematograms (RDK): vertical (RDK1) and diagonal (RDK2)
Summary 219
motion, and expansion (RDK3). Orienting eye movements were quantified

measuring reaction time to fixation (RTF).
RTF was significantly prolonged in the young group compared to the older
group, whereas in the older group, RTF was significantly affected by the type
of RDK shown.
The results suggest that based on ocular motor responses, differences in
processing of types of coherent motion can be revealed. This result has en-
couraged us to further develop and evaluate different types of visual stimuli.
Chapter 9 is again a description of an eye tracker study. This time, the focus
was not on assessment of visually-guided responses, but on assessment of
different fixation strategies during a Gestalt perception task. Such task as-
sesses the ability to visualize a complete whole when presented incomplete
information or a partial picture, a so-called Gestalt item. To our knowledge,
this is one of the first studies on Gestalt perception in visually impaired chil-
dren with a special focus on (sustained) visual attention. The database of a
longitudinal eye tracking study consisted of 72 patients/clients of Royal Dutch
Visio who underwent the Gestalt closure test. Digitised Gestalt items were
shown one-by-one on a monitor with an integrated eye tracker to test the
performance and to quantify the number of fixations and the mean and total
fixation duration. The children were divided in three different groups: ocular
visual impairments (OVI; n = 38), ocular and cerebral visual impairment
(OCVI; n = 14) or cerebral visual impairments (CVI; n = 20). Children with
visual impairments performed more often weak than age-matched American
controls. Differences remained significant even after controlling for differences
in cognitive level. Children with brain damage performed significant worse on
the animate items than the group without brain damage. The data collected
on orienting attention revealed that this performance in the CVI children could
Sum
not be attributed to a lack of effort.
Chapter 10 is a general discussion of all findings and their implications for

clinical practice and research. Most important outcomes are, that in children
with (suspected) brain damage:
1. the crowding ratio can be used as a detection instrument for possible brain
dysfunction in the population that is referred for ophthalmological or spe-
cialised low vision assessment, but is insufficiently effective for screening
in the unselected children’s population;
2. results of motion perception tasks should be evaluated against develop-
mental age and not chronological age;
220 Summary
3. problems are specifically found for global motion and motion-defined form
perception;
4. motion speed problems are scarce, but motion speed discrimination is a
complex task and cognitive limitation hampers understanding of the test;
5. biological motion tests are too difficult for this developmental age group
and/or its’ validity is questionable;
6. other dorsal stream problems (object recognition, visual attention, visuo-
contructive functioning) are more often isolated than simultaneous pres-
ent;
7. object recognition problems are associated with weak motion perception
and weak visual attention, but not with weak visuoconstructive function-
ing;
8. the performances on Gestalt perception are weak compared to the children
without brain damage. Specifically recognising the animate items seems
impaired.
Main recommendations for improvement of clinical practice are:

1. Addition of crowding assessment to routine orthoptic diagnostics of children
in (paediatric) ophthalmology clinics;
2. Addition of crowding assessment to routine orthoptic intake assessments in
specialist low vision services; and
3. addition of ‘confirmed brain damage’ or ‘explicit risk of brain damage’ to
the current Dutch eligibility requirements for specialised diagnostic assess-
ment in low vision services.
These measures will improve detection and expert diagnosis of children with
cerebral visual dysfunction.
Most of the quantitative neuropsychological tests used appear applicable and

valid. The eye tracker study presented in Chapter 8 was part of the evaluation
and validation studies that were conducted by the Neuroscience department
of Erasmus MC and Royal Dutch Visio. This has recently led to implementation
of eye tracking within low vision centres. In Chapter 9, the joining of forces
between the departments of Intellectual Disability Medicine and Neuroscience,
Erasmus MC, took a new turn in developing eye tracking paradigms with a
special focus on testing neuropsychological aspects.
Before these new paradigms can be added to routine diagnostic assessment
in low vision services, normal limits have to be evaluated in larger samples, so
this is our most relevant recommendation for collaborative (international) re-
search. To date, together with the Vestibular and Ocular Motor Research Group
Summary 221
of the dept. Neuroscience, Erasmus MC, the dept. of Paediatrics, Subdivision

of Neonatology and Pediatric Intensive Care, Erasmus MC – Sophia Children’s
Hospital and Royal Dutch Visio new future paradigms and applications are
designed, evaluated and validated.
Sum
Samenvatting
Samenvatting
Samenvatting 225
Dit proefschrift behandelt de neuropsychologische diagnostiek van cerebrale

visuele functiestoornissen (‘cerebrale slechtziendheid’) bij kinderen met her-
senschade of een verhoogd risico daarop. Het doel was om te onderzoeken in
hoeverre beschikbare diagnostische tests toepasbaar zijn in de praktijk van
de visuele revalidatie- en adviescentra. Ik heb me hierbij in de eerste plaats
gericht op kinderen in de leeftijdsgroep van 4 tot 7 jaar en op functies van de
zogenaamde ‘dorsale stroom’ van het visuele systeem.
Hoofdstuk 1 is een algemene inleiding op het thema: cerebrale visuele stoor-

nissen. Vooral kinderen met cerebrale motorische stoornissen, verstandelijk
gehandicapte kinderen en veel te vroeggeboren kinderen hebben een ver-
hoogd risico op cerebrale visuele functiestoornissen. Dit hoofdstuk beschrijft
de huidige barrières voor een tijdige diagnose en behandeling door de gespe-
cialiseerde visuele revalidatie- en adviescentra. Ook wordt uitgelegd waarom
grootschalig wetenschappelijk onderzoek naar het voorkomen van cerebrale
visuele functiestoornissen tot nu toe nog niet haalbaar is, terwijl dat juist
belangrijk is om tot een effectieve aanpak van de diagnostiek en revalidatie
te komen. Eén van de redenen hiervoor is het ontbreken van gemakkelijk
toepasbare en betrouwbare kwantitatieve tests voor het meten van de ver-
werking van visuele informatie.
De laatste drie decennia is door onderzoek aangetoond dat een groot deel
van de hersenen betrokken is bij de verwerking van wat je ziet. Wetenschap-
pers hebben modellen ontwikkeld om de samenhang tussen hersengebieden
te verklaren. Een belangrijk model binnen de neuropsychologie is de karakte-
risering van twee informatiestromen in de hersenen: een ventrale stroom en
een dorsale stroom. De hoofdfunctie van de ventrale stroom is het waarnemen
en de herkenning van wat je ziet, terwijl de dorsale stroom gespecialiseerd
is in het omzetten van visuele informatie in acties. Functies van de dorsale
SNL
stroom zijn bijvoorbeeld het waarnemen van beweging, het vermogen om
motorisch te reageren op visuele prikkels, en het efficiënt selecteren van
relevante visuele informatie door middel van de visuele aandacht. Hoewel de
herkenning van voorwerpen een typische functie van de ventrale stroom is,
blijkt de dorsale stroom van belang te zijn bij de herkenning van voorwerpen
onder lastige omstandigheden, bijvoorbeeld als ze vanuit een ongebruikelijk
perspectief worden weergegeven. In deze studie zoals beschreven in dit
proefschrift richtten we ons op de functies van het dorsale visuele systeem:
objectherkenning onder minder gunstige omstandigheden, bewegingspercep-
tie, visuomotorische vaardigheden en visuele aandacht.
226 Samenvatting
Hoofdstuk 2 gaat over de vroege herkenning van kinderen die een verhoogd
risico hebben op cerebrale visuele problemen. De verschillende stappen in
de diagnostische keten, namelijk de jeugdgezondheidszorg, de oogheelkunde
en de visuele revalidatiecentra, stonden hierbij voor ons centraal. Het doel
van deze studie was te onderzoeken of in de uitkomsten van verschillende
gezichtsscherptetests (detectie vs. herkenning) en het meten van ‘crowding’
een extra bijdrage leveren bij de opsporing en doorverwijzing van kinderen
met een verhoogd risico op cerebrale visuele stoornissen. Crowding is het
verschijnsel dat iemand hinder heeft van omringende visuele informatie bij
het vaststellen van wat zij of hij ziet, ook bij een normale gezichtsscherpte.
Kinderen met verhoogde crowding kunnen bijvoorbeeld problemen hebben
bij het spelen (een vloer vol speelgoed), bij het eten (een volle gedekte tafel)
bij het winkelen (volle schappen in een supermarkt of een onbekende drukke
straat) of bij het lezen van lange woorden en teksten. Crowding kan worden
onderzocht door de prestaties op verschillende gezichtsscherptetests met
elkaar te vergelijken: testen met losse letters, waarbij de letter niet omringd
wordt met visuele informatie, en testen met letters omringd door andere let-
ters.
Uit ons onderzoek bleek dat in de jeugdgezondheidszorg de toevoeging van
diagnostiek gericht op crowding bij de screening op visuele problemen niet
wezenlijk bijdraagt aan de opsporing van kinderen met een verhoogd risico op
cerebrale visuele problemen. Daarentegen draagt deze ‘crowding diagnostiek’
in de oogheelkundige praktijk en het visueel revalidatiecentrum wel bij om on-
derscheid te maken tussen kinderen met alleen een oogheelkundig probleem
en kinderen met een verhoogd risico op een cerebraal visueel probleem.
Wij vonden echter ook dat bij het neuropsychologische onderzoek kinderen
met verhoogde crowding even vaak cerebrale visuele problemen hadden als
kinderen zonder verhoogde crowding. Onze conclusie is dan ook, dat crowding
en cerebrale visuele problemen wel vaak samengaan, maar dat de afgenomen
testen verschillende functies en vaardigheden meten. Verhoogde crowding kan
dan ook niet gebruikt worden als directe screeningsmethode voor cerebrale
visuele problemen, alleen voor screening op een verhoogd risico: op basis van
een verlaagde gezichtsscherpte zou slechts 29% van de onderzochte kinderen
met een verhoogd risico op een cerebraal visueel probleem worden doorver-
wezen, terwijl er bij 67% van deze groep sprake is van verhoogde crowding.
Verhoogde crowding is daarom mogelijk een waardevol aanvullend verwijscri-
terium. Wij adviseren oogartsen en visuele revalidatie- en adviescentra dan
ook om ‘crowding diagnostiek’ toe te voegen aan de standaard diagnostiek.
Samenvatting 227
In Hoofdstuk 3 geef ik achtergrondinformatie over het waarnemen van bewe-

ging oftewel bewegingsperceptie. Allereerst beschrijf ik wat er op dit moment
Samenvatting | 209
bekend is over de betrokkenheid van de verschillende hersengebieden als
beweging wordt waargenomen. Vervolgens ga ik in op de verschillende aspec-
de betrokkenheid van de verschillende hersengebieden als beweging wordt
ten van bewegingsperceptie en de bewegingsperceptietaken die beschreven
waargenomen. Vervolgens ga ik in op de verschillende aspecten van bewegingsperceptie
worden in de wetenschappelijke literatuur. De huidige taken richten zich op
en de bewegingsperceptietaken die beschreven worden in de wetenschappelijke
globale bewegingswaarneming, waarneming van de richting van beweging,
literatuur. De huidige taken richten zich op globale bewegingswaarneming, waarneming
herkennen van vormen alleen door beweging, en waarneming van biologische
van de richting van beweging, herkennen van vormen alleen door beweging, en
beweging, zoals de bewegingen van mens of dier. Om bewegingspatronen te
waarneming van biologische beweging, zoals de bewegingen van mens of dier. Om
creëren, wordt gebruik gemaakt van een centraal principe, het zogenaamde
bewegingspatronen te creëren, wordt gebruik gemaakt van een centraal principe, het
‘random dot kinematogram’ (RDK). Dit is meestal een zwart vlak met bewe-
zogenaamde ‘random dot kinematogram’ (RDK). Dit is meestal een zwart vlak met
gende witte stippen dat wordt aangeboden op een computerscherm. Een deel
bewegende witte stippen dat wordt aangeboden op een computerscherm. Een deel van
van de stippen beweegt in dezelfde richting, terwijl de overige stippen in alle
de stippen beweegt in dezelfde richting, terwijl de overige stippen in alle richtingen door
richtingen door elkaar heen bewegen. Als er maar genoeg stippen in dezelfde
elkaar heen bewegen. Als er maar genoeg stippen in dezelfde richting bewegen, en je
richting bewegen, en je hersenen zijn is staat dit patroon te verwerken, dan
hersenen zijn is staat dit patroon te verwerken, dan zie je dit verschil in beweging. Op
zie je dit verschil in beweging. Op deze manier kun je niet alleen de globale
deze manier kun je niet alleen de globale richting van de beweging bepalen (Figuur 1A),
richting van de beweging bepalen (Figuur 1A), maar ook vormen herkennen
maar ook vormen herkennen (Figuur 1C). Bij onderzoek naar biologische beweging, zoals
(Figuur 1C). Bij onderzoek naar biologische beweging, zoals het lopen van
het lopen van een mens, zijn de witte stippen juist de gewrichten of andere kenmerkende
een mens, zijn de witte stippen juist de gewrichten of andere kenmerkende
elementen van het lichaam (Figuur 1B). Uit het bewegingspatroon van de stippen is dan
elementen van het lichaam (Figuur 1B). Uit het bewegingspatroon van de stip-
bijvoorbeeld te achterhalen of de mens naar rechts loopt of rent.
pen is dan bijvoorbeeld te achterhalen of de mens naar rechts loopt of rent.
SNL
Figuur 1. Een
Een voorbeeld
voorbeeld van
vantaken
takenvanvandrie
driesoorten
soortenbewegingswaarneming.
bewegingswaarneming. InIn
dede praktijk
praktijk is de
is de
achtergrond meestal
achtergrond meestal zwart
zwart en
en zijn
zijn de
de stippen
stippen meestal
meestal wit.
wit.
A. Globale
A. Globale beweging:
beweging: eeneen RDK
RDK waarin
waarin20%20%vanvandedestippen
stippentegelijk
tegelijkomhoog
omhoogbeweegt
beweegt(de
(destippen
stippen
met een
met een pijltje)
pijltje) en
en de
de rest
rest alle
alle kanten op. B.
kanten op. B. Biologische
Biologische beweging.
beweging. EenEen mensenfiguur
mensenfiguur met
met lichtjes
lichtjes
op de gewrichten, die naar rechts loopt. C. Vorm die bestaat uit beweging: een RDK waarin stippels
op de gewrichten, die naar rechts loopt. C. Vorm die bestaat uit beweging: een RDK waarin stippels
in de vorm van een vierkant omhoog bewegen en de rest omlaag. De grens van het vierkant is in
in de vorm van een vierkant omhoog bewegen en de rest omlaag. De grens van het vierkant is in
het echt niet te zien.
het echt niet te zien.
Met behulp van deze taken kan in getallen uitgedrukt worden hoe goed iemand beweging
Met
kan behulp
zien; van deze
wij noemen taken kan intaken.
dat ‘kwantitatieve’ getallen uitgedrukt
In voorbeeld A is worden hoe goed
het percentage ie-
stippen
mand beweging
dat iemand kan
minimaal zien;
nodig wijom
heeft noemen dat ‘kwantitatieve’
de samenhangende bewegingtaken.
te zien,In voorbeeld
zijn
‘drempel’. In voorbeeld C wordt de drempel meestal bepaald door het minimale
percentage stippen in het vierkant dat nodig is om het vierkant te onderscheiden. Het
228 Samenvatting
A is het percentage stippen dat iemand minimaal nodig heeft om de samen-

hangende beweging te zien, zijn ‘drempel’. In voorbeeld C wordt de drempel
meestal bepaald door het minimale percentage stippen in het vierkant dat
nodig is om het vierkant te onderscheiden. Het kunnen kwantificeren is een
groot voordeel boven de nu gebruikte werkwijze, waarin iemands reacties op
beweging door middel van observatie wordt vastgesteld.
Hoofdstuk 4 beschrijft het resultaat van de bestudering van 26 internati-

onaal gepubliceerde wetenschappelijke artikelen over computertaken voor
bewegingswaarneming. Het doel van deze systematische review was het
beoordelen welke bewegingsperceptietaken al geschikt zijn voor gebruik in de
klinische praktijk van de visuele revalidatie- en adviescentra. De studies die
wij vonden beschreven allerlei verschillende taken ontworpen om hetzelfde
aspect van bewegingsperceptie te meten. Verder bleken er vooral taken te
zijn ontwikkeld voor het meten van globale beweging, en minder voor het
onderscheiden van vormen uit beweging en biologische beweging. Slechts van
een paar studies konden de data worden gecombineerd, omdat daarin (bijna)
dezelfde taken waren gebruikt. Dit resulteerde in de bevinding dat jonge
kinderen zonder een visueel probleem slechtere scores halen dan oudere
kinderen, dat jongere kinderen minder goed zijn in bewegingsperceptie. De
drempels voor bewegingsperceptie worden dus beter met de leeftijd. Hiermee
moet rekening gehouden worden bij het opstellen van normaalwaarden voor
kinderen van verschillende leeftijden. Helaas was het aantal deelnemers in de
verschillende studies klein, zeker per leeftijdsgroep, waardoor er vooralsnog
nog geen betrouwbare normaalwaarden zijn voor gebruik ervan in de klinische
praktijk.
In acht studies werden kinderen uit patiëntengroepen vergeleken met
kinderen zonder problemen. We vonden aanwijzingen dat kinderen met een
lui oog, kinderen met autisme en kinderen die te vroeg geboren zijn een
verhoogd risico hebben op problemen in de bewegingsperceptie.
Onze conclusie was, dat computertaken voor bewegingsperceptie in de
klinische praktijk best al gebruikt kunnen worden om te kijken of een kind
duidelijk slechter presteert dan andere kinderen. Het vaststellen van een
perceptieprobleem op basis van normaalwaarden is op dit moment nog niet
mogelijk.
Hoofdstuk 5 beschrijft de studie naar bewegingsperceptie bij 49 kinderen

met (mogelijke) hersenschade en 119 schoolkinderen zonder medische
problemen. Kinderen met hersenschade hebben vaak een achterstand in
Samenvatting 229
hun ontwikkeling. Daarom werd eerst bestudeerd of voor het beoordelen

van bewegingsperceptie uitgegaan moet worden van de gewone leeftijd of
de ontwikkelingsleeftijd. Naast globale bewegingswaarneming, herkennen
van vormen door beweging, waarnemen van de snelheid van beweging, en
waarneming van biologische beweging werd daarom de performale ontwikke-
lingsleeftijd (niveau van handelend vermogen) onderzocht bij de kinderen met
(mogelijk) hersenschade en 60 van de 119 schoolkinderen zonder problemen.
Het bleek dat we voor kinderen met (mogelijke) hersenschade de ontwik-
kelingsleeftijd moeten hanteren om overschatting van de problemen in de be-
wegingsperceptie te voorkomen. Bij kinderen zonder problemen bleek dat het
niet uitmaakt welke leeftijd je hanteert. Op basis van de uitkomsten van alle
119 schoolkinderen zonder problemen heb ik voorlopige normgegevens voor
verschillende leeftijdsgroepen en ontwikkelingstrends bepaald. De biologische
bewegingstaak bleek te moeilijk voor alle leeftijdscategorieën, waardoor deze
vooralsnog nog niet goed bruikbaar is voor de klinische praktijk. Resultaten
van de overige bewegingsperceptietaken kunnen door toepassing van deze
nieuwe normwaarden wel met enige voorzichtigheid gebruikt worden in de
klinische praktijk.
Hoofdstuk 6 rapporteert over de prestaties van 46 kinderen met (mogelijke)

hersenschade op 3 bewegingsperceptietaken: globale beweging, herkennen
van vormen door beweging en snelheid van beweging. De scores van deze
46 kinderen werden vergeleken met de voorlopige normgegevens zoals we
die hadden gevonden in de studie beschreven in Hoofdstuk 5. Hiermee wilden
wij bestuderen of bepaalde problemen in de bewegingsperceptie speciaal
voorkomen bij kinderen met hersenschade, of dat er een enkel aspect of
meerdere aspecten zijn aangedaan bij hetzelfde kind. Er werd bij de kinderen
met (mogelijke) hersenschade een verhoogd risico gevonden voor problemen
SNL
bij globale beweging en de herkenning van vormen door beweging. Daarom
moet in de klinische praktijk in elk geval problemen met globale beweging
of herkenning van vormen door beweging onderzocht worden. De meeste
kinderen hadden maar één probleem en er waren maar een paar kinderen die
beide problemen hadden.
In dit onderzoek hebben wij verder onderzocht wat het effect is van het ge-
bruik van de normtabel passend bij de chronologische leeftijd i.p.v. ons advies
om de tabel passend bij de ontwikkelingsleeftijd te gebruiken. Hieruit bleek
dat er inderdaad een risico is op het overschatten van bewegingsperceptiepro-
blemen: het aantal kinderen met bewegingsperceptieproblemen nam toe van
230 Samenvatting
19/46 naar 24/46, ook het aantal kinderen met zwakke prestaties op 2 taken
nam significant toe en er was zelfs 1 kind dat zwak presteerde op alle 3 taken. 
Hoofdstuk 7 beschrijft een studie naar verschillende functies van de dorsale

visuele stroom bij achtenveertig kinderen met een verhoogd risico op cerebrale
visuele problemen. Daarmee wilden we de vraag beantwoorden of kinderen
met problemen met het herkennen van voorwerpen, bijv. bij een afwijkende
kijkhoek, ook andere problemen hebben die toegeschreven kunnen worden
aan de dorsale stroom. Oftewel, of het om een op zichzelf staand probleem
gaat of dat het een breder probleem betreft. Dit is van belang voor het bepalen
welke taken er moeten worden afgenomen in het diagnostiektraject.
Naast herkenning van voorwerpen onder suboptimale omstandigheden (sub-
taken van de L94), werden ook bewegingsperceptie, het namaken van visueel
ruimtelijke constructies (tekeningen en blokpatronen), en visuele aandacht
(een zoektaak op de computer) onderzocht. De kinderen werden verdeeld
in twee groepen: groep 1 met kinderen die geen problemen hadden met de
herkenning van voorwerpen en groep 2 met kinderen die daarmee wel pro-
blemen hadden. Voor de visuoconstructieve taken werd gebruik gemaakt van
bestaande Nederlandse normen, voor bewegingsperceptie en de visuele zoek-
taak werd gebruik gemaakt van normaalwaarden gebaseerd op de prestaties
van schoolkinderen zonder medische problemen (voor bewegingsperceptie zie
Hoofdstuk 5 en 6, voor de zoektaak zie Appendix bij dit hoofdstuk).
De kinderen die problemen hadden met de herkenning van voorwerpen,
presteerden significant slechter op taken gericht op bewegingsperceptie en op
de taak voor visuele aandacht (zoektaak) dan de kinderen die daarmee geen
problemen hadden. Slechts een kleine minderheid van groep 1 had meerdere
problemen tegelijk, een breed dorsale stroom probleem. Onze conclusie is dan
ook, dat dorsale stroomproblemen bij kinderen met (mogelijke) hersenschade
meestal op zichzelf staan. Dat betekent dat bij kinderen in risicogroepen
steeds meerdere functies moeten worden onderzocht, om te zien wat het
specifieke probleem bij dit kind is.
In de Appendix van dit hoofdstuk wordt informatie gegeven over de ge-
bruikte visuele zoektaak, ontwikkeling van aan de Katholieke Universiteit van
Leuven. Ook voor deze taak hebben wij nu normaalwaarden ontwikkeld op
grond van gegevens van 60 schoolkinderen zonder medische problemen.
Hoofdstuk 8 beschrijft de resultaten van een studie waarbij de oogbewe-

gingen worden vastgelegd met een eye tracker tijdens de afname van bewe-
gingsperceptietaken. Het doel was om een methode te ontwikkelen voor het
Samenvatting 231
meten van visuele functies bij personen die niet kunnen meewerken met huidig
neuropsychologisch onderzoek. Het gaat hierbij bijvoorbeeld over kinderen
jonger dan 4 jaar of kinderen (en volwassenen) met verstandelijke beperkin-
gen. Bij deze methode maakten we gebruik van het feit, dat als de hersenen
een visuele stimulus goed kunnen verwerken, de ogen in een reflex naar deze
stimulus gaan kijken. We boden een drietal bewegingsperceptiestimuli op
basis van Random Dot Kinematograms (RDK’s) aan op een beeldscherm met
een ingebouwde camera voor het meten van oogbewegingen (de zogenaamde
eye tracker).
Deze methodiek hebben wij eerst op haalbaarheid en betrouwbaarheid ge-
test bij 188 kinderen (0-13 jaar) zonder visuele of neurologische problemen.
Wij toonden op het scherm drie verschillende RDK’s zien: één met verticale
beweging (RDK1), één met diagonale beweging (RDK2), en één met beweging
vanuit het centrum naar buiten (RDK3). De maat die met de software werd
gemeten, was de reactietijd van het kind: hoe lang deed het erover om zijn
ogen te fixeren op de goede plek.
De reactietijd was bij jonge kinderen (0-3 jaar) duidelijk langer dan bij
oudere kinderen (4-11+ jaar). Voor de oudere kinderen maakte ook het type
RDK uit voor de reactietijd.
Onze conclusie van deze eerste verkennende studie is dat het mogelijk is om
op basis van oogbewegingen uitspraken te doen over verschillende aspecten
van globale bewegingsperceptie. Dit is een bemoedigend resultaat voor de
verdere ontwikkeling van deze methode.
Ook in Hoofdstuk 9 wordt een studie naar oogbewegingen gepresenteerd. Nu

richtten we ons niet op automatische reflexmatige oogbewegingen die door
een stimulus worden uitgelokt, maar op kijkstrategieën die gehanteerd worden
bij het aanbieden van onvolledig weergegeven plaatjes (Gestalts genoemd).
SNL
Bij een Gestalt-perceptie taak wordt onderzocht of het kind in staat is de
onvolledige figuur mentaal aan te vullen en te herkennen. Deze kijkstrate-
gieën kunnen mogelijk iets zeggen over gerichte visuele aandacht. Voor zover
wij weten is dit een van de eerste studies naar Getalt-perceptie bij kinderen
met een visuele beperking. Bij deze studie hebben we gebruik gemaakt van
gegevens van 72 kinderen die bij Koninklijke Visio bekend zijn en die deel
hebben genomen aan een longitudinale studie en waarbij o.a. de digitale ver-
sie van de Gestalt Closure test is afgenomen. De kinderen kregen een reeks
Gestalt-items aangeboden. De oogbewegingen werden d.m.v. de eye tracker
geregistreerd en de gegeven antwoorden werden genoteerd. Met behulp van
de gegevens van de oogbewegingen werden het aantal fixaties van de ogen op
232 Samenvatting
het plaatje en de gemiddelde en totale fixatieduur van de ogen op het plaatje

bepaald. De deelnemers werden verdeeld in drie groepen, afhankelijk van de
oorzaak van de visuele beperking: kinderen met een oogheelkundig probleem
(OVI; n = 38), kinderen met een oogheelkundig probleem en hersenschade
(OCVI; n = 14) of kinderen met alleen hersenschade (CVI; n = 20). Kinderen
met een visuele beperking presteerden vaker zwak op de Gestalt Closure test
in vergelijking met de bestaande Amerikaanse normgegevens. Dit betekent
dat ze plaatjes vaker niet herkenden. Ook als er gecontroleerd werd voor het
geschatte ontwikkelingsniveau presteerde deze groep vaker zwak. Opvallend
was dat de kinderen met hersenschade met name de plaatjes van dieren min-
der vaak herkenden dan de kinderen zonder hersenschade. Uit het kijkgedrag
bleek dat dit resultaat niet verklaard kan worden door een gebrek aan inzet,
omdat ze juist langer en vaker op deze plaatjes fixeerden dan de anderen.
Hoofdstuk 10 is een algemene beschouwing van de uitkomsten gepresen-

teerd in dit proefschrift en wat de betekenis ervan is voor de klinische praktijk
en voor het doen van verder wetenschappelijk onderzoek. De belangrijkste
uitkomsten voor kinderen met (mogelijke) hersenschade zijn dat:
1. de crowding ratio gebruikt kan worden om kinderen met een verhoogd
risico op cerebrale visuele problemen op te sporen als ze de oogarts bezoe-
ken of verwezen zijn naar visuele revalidatie- en adviescentra. Onderzoek
naar crowding is niet effectief in de algemene populatie die het consultatie-
bureau bezoekt;
2. bij de beoordeling van de prestaties op bewegingsperceptietaken rekening
gehouden zou moeten worden met de performale ontwikkelingsleeftijd en
niet (alleen) gekeken zou moeten worden naar de prestaties in vergelijking
met de gewone leeftijd;
3. deze kinderen een verhoogd risico hebben op specifieke bewegingspercep-
tieproblemen, namelijk bij globale beweging en het herkennen van vormen
door beweging;
4. problemen met waarnemen van de snelheid van beweging zelden voor-
komen, maar het waarnemen van snelheidsverschillen is een complexe
taak en het begrip van de taak lijkt beperkt te worden door cognitieve
beperkingen;
5. de onderzochte biologische bewegingstaak met een lopend persoon te
moeilijk is voor kinderen met een ontwikkelingsleeftijd van 4-7 jaar en/of
dat de validiteit van deze taak te betwijfelen valt;
Samenvatting 233
6. andere dorsale visuele stroom problemen (herkennen van voorwerpen on-

der suboptimale omstandigheden, visuele aandacht en visuoconstructieve
vaardigheden) vaker geïsoleerd voorkomen dan tegelijk;
7. problemen met het herkennen van voorwerpen geassocieerd zijn met
problemen in de bewegingsperceptie en visuele aandacht, maar niet met
problemen in visuoconstructieve vaardigheden;
8. ze vaak zwak presteren op de Gestalt-perceptietaak in vergelijking met
kinderen zonder hersenschade. Met name het herkennen van onvolledige
dierenplaatjes lijkt beperkt.
Onze voornaamste aanbevelingen voor verbetering van de klinische praktijk

zijn de volgende:
1. Diagnostiek van crowding toevoegen aan de routine orthoptische diagnos-
tiek bij kinderen in de oogheelkundige praktijk;
2. Diagnostiek van crowding toevoegen aan het routine orthoptisch intake
onderzoek van de gespecialiseerde visuele revalidatie- en adviescentra, en
3. Toevoegen van ‘bevestigde hersenschade’ of ‘expliciet risico op hersen-
schade’ aan de huidige Nederlandse indicatiecriteria voor gespecialiseerd
diagnostisch onderzoek door visuele revalidatie- en adviescentra.
Deze maatregelen zullen de opsporing en diagnostiek van kinderen met
cerebrale visuele problemen verbeteren.
Wij hebben aangetoond dat de meeste van de door ons gebruikte kwantita-
tieve neuropsychologische tests bij kinderen toepasbaar en valide zijn. De
eye tracking studie, zoals beschreven in Hoofdstuk 8, was onderdeel van een
evaluatie- en validatiestudie die werd uitgevoerd door de afdeling Neurowe-
tenschappen van het Erasmus MC en Koninklijke Visio. Dit heeft recent geleid
tot de implementatie van het gebruik van de eye tracker bij onderzoeken in
SNL
de klinische praktijk bij Koninklijke Visio. De samenwerkende onderzoekers
van de afdelingen Geneeskunde voor Verstandelijk Gehandicapten en Neuro-
wetenschappen van het Erasmus MC sloegen een nieuwe weg in door ook een
andere neuropsychologische tests, namelijk een Gestalt-perceptie taak, zoals
beschreven in Hoofdstuk 9, te onderzoeken met de eye tracker.
Maar voordat deze taken kunnen worden toegevoegd aan het routine
diagnostisch onderzoek in visuele revalidatie- en adviescentra, moeten de
bijbehorende normaalwaarden beter worden onderzocht in grote groepen
kinderen. Dat is dus onze belangrijkste aanbeveling voor verder onderzoek.
Dit zal samenwerking van onderzoeksgroepen vergen, eventueel internatio-
naal. Op dit moment werken professionals van de Vestibulaire en Oculomotore
234 Samenvatting
Onderzoeksgroep, afdeling Neurowetenschappen, Erasmus MC, de afdeling

Kindergeneeskunde, subafdeling Neonatologie en Kinder Intensive Care,
Erasmus MC-Sophia, en Koninklijke Visio al samen en ontwikkelen, evalueren
en valideren nieuwe onderzoeksmethoden en -toepassingen op het gebied van
de visuele ontwikkeling bij kinderen, o.a. bij kinderen met een verhoogd risico
op cerebrale visuele functiestoornissen.
Dankwoord
Dankwoord
222 | Dankwoord
Dank jullie wel!

Dankwoord 237
Het is zover! Ik heb toestemming voor het drukken van dit proefschrift, dus
het einde van het promotietraject is in zicht. “Eindelijk!”. Ik zal vast niet de
enige zijn die dit denkt. Het is een lang traject geweest met zo z’n ups en z’n
downs en velen van jullie hebben op een of andere manier een deel van deze
reis met mij meegemaakt.
Allereerst dank aan alle deelnemende kinderen. Dank jullie, dat jullie mee
wilden doen! Meedoen was voor de een gemakkelijker dan voor de ander:
soms ging alles goed, soms ontdekte je dat je iets niet zo goed kon. De een
kon er dan nog om lachen, de ander werd er verdrietig of boos van. Toch bleef
iedereen meedoen. Ik heb bewondering voor jullie doorzettingsvermogen.
Dank dat ik van jullie mocht leren en dat ik dat weer mag delen met anderen.
Natuurlijk ook dank aan de deelnemende ouders en medewerkers van de
deelnemende scholen: Eduard van Beinum, Openbare basisschool Charlois, De
Bergse Zonnebloem, Visio-school. Dank ook aan de organisaties die bij mijn
onderzoek betrokken waren: Rijndam Revalidatiecentrum, Koninklijke Visio en
Erasmus MC. Zonder jullie had ik dit onderzoek en het proefschrift niet van de
grond gekregen en had ik veel kennis en ervaring niet op kunnen doen.
Ook ben ik veel dank verschuldigd aan Prof.dr. Heleen Evenhuis. Heleen,
zonder jouw initiatief, coaching, geduld, en eindeloze vertrouwen in mijn mo-
gelijkheden had ik deze wetenschappelijke en persoonlijke reis niet kunnen
maken en afronden. Jij trok de conclusie dat ook onderzoek naar Cerebrale
Visuele Informatieverwerkingsproblemen (CVI; Cerebral Visual Impairment)
van belang is. Door de samenwerking met Annemieke Blokker, destijds regio-
directeur van Koninklijke Visio, verbond je de wetenschap, vanuit de leerstoel
Geneeskunde voor Verstandelijk Gehandicapten bij het Erasmus MC, met de
praktijk van Koninklijke Visio en kreeg ik de kans die ik zocht: werken in het
veld van de neuropsychologie en het kunnen combineren van wetenschap
en praktijk. Deze combinatie is niet altijd gemakkelijk gebleken, maar was
D
zeker interessant en leerzaam. Dank voor het feit dat je bleef zoeken naar
alternatieve mogelijkheden, je concrete feedback en de tijd, die je ook na je
pensioen voor me vrij hebt gemaakt.
De inhoud van mijn onderzoek heb ik grotendeels te danken aan de voortrek-
kers in het internationale en nationale veld van de visuele neuropsychologie
en neurologie. In het bijzonder dank aan Dr. Peter Stiers. Dank je, Peter, dat je
me de eerste jaren van het traject hebt willen begeleiden. Ik heb fijne herin-
neringen aan onze inhoudelijke discussies en mijn bezoeken aan Leuven en
Maastricht. Dank, dat je jouw verzamelde data en het testmateriaal dat door
jou en jouw studenten ontwikkeld is, beschikbaar hebt gesteld voor nader
onderzoek. Ook wil ik o.a. de managers, collega-gedragswetenschappers,
238 Dankwoord
oogartsen en orthoptisten binnen Visio en Bartiméus bedanken. Dank dat jullie

met mij van gedachten wilden wisselen over welke hindernissen er waren/zijn
binnen het veld van visuele revalidatie. Dit gaf richting aan mijn onderzoek. In
het bijzonder nog dank aan Dr. Paul Looijestijn. Dank je, Paul, voor je bijdrage
aan de begeleidingsgroep, het delen van je ervaring en kennis.
Gedurende vrijwel het hele traject waren er contacten met de afdeling Neu-
rowetenschappen van het Erasmus MC, in het bijzonder met Prof.dr. Hans van
der Steen, Dr.ir. Johan Pel en Dr. Marlou Kooiker. Hans, dank voor je aanhou-
dende belangstelling voor het verloop van mijn promotietraject. Johan, dank
je, dat je me bleef prikkelen en betrekken bij wetenschappelijke projecten en
aanverwante zaken, waardoor het ook leuk bleef om aan mijn eigen artikelen
te werken. Niet in de laatste plaats bedankt voor het feit dat je in de laatste
periode van mijn promotietraject het co-promotorschap op je hebt willen ne-
men. Met jouw positieve instelling en steun, gaf je me opnieuw perspectief op
een positieve afronding van dit traject en waren de laatste loodjes draagbaar
en interessant. Milou, dank voor de inspirerende discussies, de gezamenlijke
zoektochten naar houdbare interpretaties van de data en je feedback.
Het doen van wetenschappelijk onderzoek, vraagt ook regelmatig aanpas-
singen van de dagelijkse praktijk. Meerdere (oud-)collega’s zijn tussen hun
gewone werkzaamheden door met mij op stap geweest om de basale visuele
functies van de deelnemende kinderen te onderzoeken en vast te leggen.
Naast het uitvoeren van de gebruikelijke testmethoden, waren ze bereid
om dingen net even anders te doen in het belang van het onderzoek. Ook
binnen de regionale centra van Koninklijke Visio in Goes, Rotterdam, Den
Haag/Leiden hebben orthoptisten hun werkwijze aangepast door ook crowding
(visuele ruis) te onderzoeken. Amanda, Ingrid, Julia, Lyanne, Mariëtte, Marlien
en Yvonne, dank voor jullie flexibiliteit, de fijne samenwerking en bijdrage aan
het onderzoek. Verder wil ik Suzanne en Apiradee (Kwang) bedanken voor hun
bijdrage aan de dataverzameling, Gerda en Corrie voor hun ondersteuning bij
de dataverwerking.
Een goede voorbereiding is het halve werk. Het promotietraject startte in
2005, de reis ernaartoe al veel eerder. Wat mij betreft is het begonnen met
mensen die mij inspireerden en nieuwsgierig maakten, zoals vele docenten
op mijn weg hebben gedaan. Dank jullie allen voor het leggen van het funda-
ment. Dank je Már, dat je me in de zomer van 2000 mee nam in de wereld van
de klinische psychologie, het evidence-based werken en de congressen. Takk,
Már! Thank you, for your hospitality, the introduction to clinical psychology,
evidence-based working, congresses and for showing me Iceland! Bij Kind
& Adolescent, Gent 2003, werd de kloof tussen wetenschap en praktijk voor
Dankwoord 239
mij voor het eerst gedicht en werd het zaadje geplant dat wetenschappelijk
onderzoek ook wel iets voor mij zou kunnen zijn. Dank jullie voor de gezellige
en leerzame periode!
Werken en leren is het leukst in een omgeving met een warme belangstel-
lende sfeer. Wat heb ik toch veel fijne mensen om me heen! Beste (oud-)
collega’s van de Leerstoel Geneeskunde voor Verstandelijk Gehandicapten,
Huisartsgeneeskunde en van Koninklijke Visio en meiden van de GZ-oplei-
dingsgroep GZ17-E, dank voor jullie steun, lieve en motiverende woorden en
de benodigde afleiding als ik even geen zin had om met mijn promotietraject
bezig te zijn.
Lieve heit en mem, lieve familie en vrienden. Ik weet het, ik ben een eeu-
wige student. Ik zit vaak met mijn neus in de boeken, artikelen of werk aan
opdrachten en artikelen. Tijd en energie zijn net als geld, je kunt ook tijd en
energie maar 1x besteden, helaas. Dit had regelmatig tot gevolg dat ik jullie
minder vaak zag en sprak, dan dat ik zou willen. Wat heb ik genoten van het
samen zijn, de gesprekken, de lekkere kopjes thee, het samen lekker eten,
puzzelen, naar de film of theater gaan en uitwaaien. Tussendoor waren er de
gezellige telefoontjes, lieve kaartjes en appjes. Dank jullie wel hiervoor! Ho-
pelijk komt er met het afronden van dit traject een zee van tijd vrij, waardoor
we weer vaker samen kunnen genieten.
Lieve Bionka en Sjoukje, dank dat jullie mijn paranimfen willen zijn, me bij
willen staan bij het verdedigen van mijn proefschrift. Dank voor jullie hulp bij
de voorbereidingen, het meelezen en de gezelligheid.
Last but not least, lieve Roelof, vrijheid en onszelf verder ontwikkelen staan
hoog in ons vaandel. Dank voor de ruimte die je me hiervoor geeft. En wat
is het fijn om dit niet alleen te doen. Ik waardeer dat je mij achterna bent
gekomen en gezellig bij mij in Rotterdam bent komen wonen. Dank voor de
lieve zorg als ik weer eens geen puf meer had, de gezellige uitjes, de fijne
D
wandelingen en fietstochten in binnen- en buitenland. Ik hoop dat we nog lang
samen mogen genieten!
About
About the
the author
author
About the author 243
th
Ymie van der Zee was born on December 26 , 1978 in Oosterwolde (FRL),
the Netherlands. She finished her pre-university education at the CSG Liudger
in Drachten in 1998. That same year, she started as a student Psychology
at the Rijksuniversiteit Groningen. In 2000 she visited Iceland to combine a
vacation with a short introduction in clinical psychology in a hospital in Reyk-
javik (Sjúkrahús Reykjavíkur). After writing a master thesis titled “Parkinson’s
Disease: questionnaire-based research on sensation seeking combined with
experimental research on stimulus novelty effects in EEG-recordings”, she
started a six months traineeship at ‘Kind & Adolescent’, a university psycho-
logical centre in Ghent, Belgium. In 2003 she finished her study in psychology
with two majors, Clinical Psychology and Methodology, and two minors Neuro-/
Biopsychology and Developmental psychology. During the summer breaks of
2002 and 2003, and in 2004 till February 2005 she worked as professional
caregiver in a residential care centre for people with an intellectual disability
and additional problems like Borderline personality disorder and Gilles de la
Tourette (Talant, Beetsterzwaag, The Netherlands).
In February 2005 she started her PhD track on Cerebral Visual Impairment
(CVI), which started out as a collaborative effort of the Department of Intel-
lectual Disability Medicine of the Erasmus MC, Rotterdam, The Netherlands,
Royal Dutch Visio, Dutch Centre of Excellence for visually impaired and blind
people, The Netherlands and the Laboratory of Neuropsychology of the K.U.
Leuven, Belgium (later Department of Neuropsychology & Psychopharmacol-
ogy, University Maastricht, Maastricht, The Netherlands). The last years the
collaboration with Vestibular and Ocular Motor Research Group, Department of
Neuroscience, Erasmus MC, Rotterdam, the Netherlands intensified.
Results of these collaborations can be found in this doctor thesis, titled:
“Diagnosing Cerebral Visual Dysfunctions in Children: looking beyond Visual
Acuity and Visual Field”.
cv
Since 2010 she is mainly working as neuropsychologist at Royal Dutch Vi-
sio. She worked in multiple teams and focusses on children with and without
intellectual disabilities or brain damage. In 2017 she started her postmaster
education to obtain the registration as health psychologist (Gezondheidszorg-
psycholoog) to improve her clinical skills and combine research and clinical
practise in the future.
PhD
PhD Portfolio
Portfolio
PhD Portfolio 247
Name PhD student: Y J. van der Zee

Erasmus MC Department: Intellectual Disability Medicine/Dpt. Of General Practice
PhD period: 2005-2018
Promotor: H.M. Evenhuis
Co-promotor: J.J. Pel
1. PhD training Year
General courses
Master class ‘Improving your clinical research’ 2006
Methodologie van patiëntgebonden onderzoek en voorbereiding van subsidieaanvragen 2006
Statistiek in de kliniek, basisgebruik SPSS 14 2007
Specific courses
Paediatric Clinical Epidemiology (Erasmus Winter programme) 2006
International Classification of Functioning, Disability and Health (ICF; Koninklijke Visio) 2007
Active Memory; Pathways through the Brain in Memory and Action (NWO Cognition 2006
Summer School, Doorwerth; Poster presentation)
Kinderoogheelkunde (Boerhaave Nascholing, Leiden; Attendance and oral presentation) 2015
Presentations
Fine scale functional organization of the human ventral stream (Helmholtz lecture) 2006
Space and the parietal cortex (Helmholtz lecture) 2006
Functional Magnetic Resonance Imaging of Human Visual Cortex (Helmholtz lecture) 2007
Learning to see late in life (Lecture Erasmus MC) 2014
Video Interaction Guidance – Light version (Koninklijke Visio) 2015
Afscheidscollege Prof. Dr. Heleen Evenhuis. Van ontginnen naar oogsten. 15 jaar 2015
epidemiologisch onderzoek bij mensen met verstandelijke beperkingen (Erasmus MC)
(Inter)national conferences/Seminar
Kennis InZichtelijk: Hoe gebruik ik kennis om mijn werk met mensen met een visuele 2005
beperking te professionaliseren? (ZonMW InZicht; Attendance)
Bridging research, policy and practice (E-IASSID Congress, Maastricht; Oral 2006
presentation).
Brain Days: functional response to cerebral damage & development and adaptation of 2006
structure and function (Attendance)
VOGG (Vereniging voor ouders en verwanten van mensen met een verstandelijke 2007
handicap) Symposium. (Lelystad; Oral prestentation) P
Focus op onderzoek. Wetenschappelijk onderzoek ten behoeve van (de zorg aan) 2007
mensen met een verstandelijke beperking. (Kennisplein Gehandicaptensector en
ZonMw, Amsterdam; 2 oral prestentations)
Cognitieve revalidatie (ZonMw consortium Cognitieve Revalidatie; Attendance) 2007
Technologie en communicatie in het leven van mensen met een visuele beperking 2007
(Attendance)
Voortgang In Zicht (ZonMW InZicht; Attendance) 2007
10 jaar InZicht: een decennium innoveren en implementeren (ZonMW InZicht; 2008
Attendance)
IASSID Congress, Cape Town, South Africa (Oral presentation) 2008
InZicht Oogopslag (ZonMW InZicht; Attendance) 2010
Vroeghulpsymposium. Terugkijken & Vooruitzien (MEE - Integrale Vroeghulp Rotterdam; 2015
Attendance)
248 PhD Portfolio
Middagsymoposium SOLK oorzakelijk behandelen (Logacom BV, Amsterdam) 2016

Other
Member of Klankbordgroep Integrale Vroeghulp Rotterdam 2012-2017
2. Teaching Year
Lecturing
Guest Lecturer at ID physician training 2008
Regional meeting for orthoptists. Rotterdam 2008
Meeting for personnel of a rehabilitation center. Rotterdam 2008
Meeting for physicians of child health care centers. Rotterdam 2008
Meeting for personnel of special schools for blind and visually impaired children 2008 and
later
Supervising Master’s theses
Apiradee Poungjit. Master thesis Clinical Neuropsychology. Relation of Crowding Effect and 2009
Visual Search in Children: Comparison between children with a high risk of cerebral visual
impairment and normal children.
Suzanne Ton. Master thesis Clinical Neuropsychology and Child & Adolescent Psychology. 2010
Cerebral Visual Impairment in Children. A pilot study on motion perception in children with
cerebral damage and children without cerebral damage.
Jeanne van der Wulp. Master thesis Clinical Neuropsychology. Using eye movements to 2017
predict reading skills in children. An eye tracking study while reading aloud a narrative
Marisabel Talamante Ojeda. Master thesis in Neuroscience and Behavioral Biology. 2017
Perception tests in children with visual impairments: an eye movement approach
A digital version of this thesis is available at
https://epubs.ogc.nl/?epub=y.vanderzee

Thesis Diagnosing Cerebral Visual Dysfunctions in Children-Looking Beyond Visual Acuity and Visual Field

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Thesis Diagnosing Cerebral Visual Dysfunctions in Children-Looking Beyond Visual Acuity and Visual Field

Uploaded by

Copyright:

Available Formats

Diagnosing Cerebral

van der Zee

IJmkjen Janneke van der Zee

Cover design: Optima Grafische Communicatie, Rotterdam

Copyright © 2019, IJmkjen Janneke van der Zee

Het vaststellen van cerebrale visuele stoornissen bij kinderen:

ter verkrijging van de graad van doctor aan de

Prof.dr. R.C.M.E. Engels

en volgens besluit van het College voor Promoties.

IJmkjen Janneke van der Zee

Promotor: Prof.dr. H.M. Evenhuis

Overige leden: Prof.dr. N.E. Schalij-Delfos

Copromotor: Dr.ir. J.J.M. Pel

Chapter 1 General introduction 7

Chapter 2 Should we add visual acuity ratios to referral criteria for 19

Chapter 3 Visual motion perception: brain localisation, terms and 37

Chapter 4 Applicability of computerised motion perception tasks 45

Chapter 5 Chronological age versus developmental age in 99

Chapter 6 Clinical assessment of visual motion perception 121

Chapter 7 Object recognition and dorsal stream vulnerabilities in 135

Chapter 8 Remote eye tracking assesses age dependence 157

Chapter 9 Gestalt perception in children with visual impairments: 173

Chapter 10 General discussion 195

Children that need special support because of a visual

a ball or pouring a cup of tea. From a social interaction perspective, it also

Diagnosing cerebral visual dysfunctions in children

A few reasons why epidemiological knowledge on cerebral visual dysfunc-

Aim of this scientific project and contents of this thesis

In Chapter 2, we focus on early recognition by e.g. youth healthcare physi-

visually integrate parts of the representation into a whole. In clinical practise,

Should we add visual acuity ratios

Childhood cerebral visual impairment often goes unrecognised, especially

For this cross-sectional study, a heterogeneic group of 107 children in the

Five children in group 1 had refractive errors and 12 latent strabismus. In

Groups performed significantly different on the visual acuity tests (Kruskal-

Uncrowded acuity/crowded acuity

Only in group 2, a significant relationship between age and the uncrowded/

Grating acuity/uncrowded acuity

Grating acuity/crowding acuity

Referral criteria for different specialists

This study explored an opportunity to improve clinical identification of children

Visual motion perception: brain

Motion perception is the specific mental function of recognizing and interpret-

Hierarchical brain model of motion perception

Different motion perception paradigms

Another, more complex, perceptual function is the ability to recognize forms

Motion perception is the specific mental function of recognizing and interpret-

Table 1. Search strings used in each database and number of hits.

95%-CI of the intended percentage of excluded controls is the true population

Study characteristics and quality assessment

Table 3. Rating of confounding risk

Motion perception tasks and thresholds

Estimated normal limits

Controls Patients Statistical test

without providing evidence for symmetrical or normal distributions of scores.

In Figure 1C, we present a comparison of two studies on motion-defined

Patients’ performance compared to typically developing

task, we performed a post-hoc t-test on the summary data. Children with

We conclude that, considering group mean performances, multiple patient

Ad 3. Another matter that needs consideration concerns factors that may

Table 1. Characteristics of the global motion studies

Table 2. Characteristics of motion-defined form studies

Table 3. Characteristics of the biological motion studies