Manual Therapy 18 (2013) 458e467

Contents lists available at ScienceDirect

Manual Therapy
journal homepage: www.elsevier.com/math

Review article

Attentional focus of feedback and instructions in the treatment

of musculoskeletal dysfunction: A systematic reviewq
Catrina Sturmberg a, Jodie Marquez a, b, Nicola Heneghan c, Suzanne Snodgrass a, b,
Paulette van Vliet a, b, *
a
Discipline of Physiotherapy, School of Health Sciences, University of Newcastle, Australia
b
Centre for Translational Neuroscience and Mental Health, Hunter Medical Research Institute, Newcastle, Australia
c
School of Health and Population Sciences, University of Birmingham, United Kingdom

a r t i c l e i n f o a b s t r a c t

Article history: Background: An external focus of attention (EFA) during the learning of a motor task improves perfor-
Received 7 January 2013 mance and retention in healthy individuals. People with musculoskeletal dysfunction also learn motor
Received in revised form tasks and could potentially benefit from adopting an EFA during practice.
24 June 2013
Objectives: To determine whether instructions and feedback provided to individuals with musculo-
Accepted 11 July 2013
skeletal dysfunction is more effective in improving function and decreasing pain when inducing an
external rather than an internal focus of attention (IFA).
Design: Systematic review
Search methods: MEDLINE, Embase, CINAHL, AMED, the Cochrane Library and five additional databases
were searched.
Selection criteria: Randomised, quasi-randomised and non-randomised controlled trials, cross over trials
and observational studies involving participants with any form of musculoskeletal dysfunction,
comparing IFA or EFA with a different attentional focus (AF), control, placebo or no focus condition.
Data collection and analysis: Two review authors independently screened titles, abstracts and full texts,
then extracted data and appraised the quality of trials using the GRADE system of rating methodological
quality.
Results: Seven studies were included with a total of 202 participants. Two studies compared an IFA with
an EFA, two compared IFA with biofeedback with a different focus condition, and three compared IFA
with biofeedback with a no focus condition. Statistically significant improvements in motor performance
directly attributable to the focus of attention were only found in the EFA groups. There were no sig-
nificant improvements in function or pain.
Conclusion: There is insufficient evidence to draw conclusions regarding the effects of attentional focus
of instructions and feedback on outcomes in musculoskeletal dysfunction.
Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction dysfunction is a significant drain on the health and social care

Musculoskeletal dysfunction is a major cause of morbidity and
disability worldwide (Mock and Cherian, 2008). As well as the
personal cost to individuals and their families, musculoskeletal
q Institution to which the work should be attributed: Discipline of Physiotherapy,
School of Health Sciences, and Hunter Medical Research Institute, University of
Newcastle, University Drive, Callaghan, NSW 2308, Australia.
* Corresponding author. School of Health Sciences, University of Newcastle,
University Drive, Callaghan, NSW 2308, Australia. Tel.: þ61 2 4921 7833; Fax: þ61 2
4921 7053.
E-mail address: paulette.vanvliet@newcastle.edu.au (P. van Vliet).
system (Handoll et al., 2007) and the leading cost driver in workers
compensation (Cusimano-Reaston and Carney, 2011). Healthcare
costs attributable to musculoskeletal spinal pain alone are esti-
mated at $86 billion annually and predicted to rise (Martin et al.,
2008). The treatment of musculoskeletal dysfunction may be
improved through effective rehabilitation. Indeed, early rehabili-
tation of musculoskeletal dysfunction has demonstrated large
financial benefits, with one onsite industrial physical therapy pro-
gram reporting a benefit-to-cost ratio of greater than nine to one
(Hochanandel and Conrad, 1993).
Rehabilitation in musculoskeletal dysfunction frequently in-
volves relearning motor skills with the use of instructions and

1356-689X/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.math.2013.07.002
 C. Sturmberg et al. / Manual Therapy 18 (2013) 458e467
feedback. Feedback, by increasing an individual’s knowledge and
understanding of how they are performing a skill or of their success
in achieving a desired result, helps a person learn motor skills more
quickly and easily, motivating a learner to continue to strive to-
wards achieving a skill goal (Magill, 2007). A large body of research
in healthy people has investigated the effects of an individual’s
focus of attention during their movement performance (Wulf et al.,
2001; Wulf and Prinz, 2001). Instructions that direct an individual’s
attention to the effects of their movements on the environment
(e.g. the tool they are using) are said to induce an external focus of
attention (EFA). Alternatively, when an individual is focussing on
movement of the body or body parts, they have an internal focus of
attention (IFA).
A convincing body of evidence in healthy subjects shows an EFA
during the learning of a motor task improves performance and
retention of that motor task (for a review, see Wulf et al., 2001b). To
explain this phenomenon, McNevin et al. (2003, p.22) propose a
‘constrained action hypothesis.’ They argue that consciously trying
to control one’s movements constrains the motor system by
interfering with automatic motor control processes that would
“normally” regulate the movement. Supporting this notion, an EFA
during a balance task has been found to result in better task per-
formance, increased frequency of responding to permutations and
decreased attention demands compared to an internal focus (Wulf
et al., 2001). This increased efficiency is indicative of more auto-
matic motor control processing (Wulf et al., 2001). Moreover, the
further the attentional focus (AF) is shifted from the body, the
better the individual’s response frequency and overall performance
(Wulf et al., 2001; McNevin et al., 2003). It is thought that focussing
on a movement effect near to the body invites more conscious
intervention in the movement process than when the focus is
further away and hence easier to distinguish from the body
(McNevin et al., 2003). However, the movement effect must be near
enough to the body to be easily related to the movement performed
(Wulf et al., 2001b). Thus the proposed optimal focus is directed to
an effect that is as remote as possible from the body, whilst being as
closely related as possible to the action that caused it (Wulf et al.,
2001b).
Though an EFA is reported as more effective for skill mastery in
asymptomatic persons (Wulf et al., 2001b) and in patients with
stroke (van Vliet and Wulf, 2006), it is not known if this also occurs
in patients with musculoskeletal dysfunction, nor whether this
improved learning effect results in improved patient outcomes.
This review aims to examine and synthesise the available evidence
regarding the impact of AF during motor learning among people
with musculoskeletal dysfunction. The objective of this systematic
review was to determine whether instructions and feedback
inducing an external focus of attention are more effective than
instructions and feedback inducing an internal focus of attention, in
improving function and decreasing pain in patients with muscu-
loskeletal dysfunction.
2. Method
The method followed a review protocol established before the
review was conducted that included details of the research ques-
tion, eligibility criteria, search methods, selection of studies,
grading of methodological quality, data extraction and data
synthesis.
2.1. Eligibility criteria
2.1.1. Study design
This review was inclusive with regard to study design due to the
small number of published studies. Randomised, quasi-randomised
459
and non-randomised controlled trials, cross over trials, observa-
tional studies and case-control studies were included. Studies had
to be in English, and were included with or without blinding of
participants, treating therapist(s) and assessor(s), and with any
length of follow-up.
2.1.2. Comparisons
For inclusion in the review, studies had to report an intervention
or experimental manipulation with instructions or feedback
inducing an IFA or an EFA, compared with: 1) an opposing attention
condition; 2) a control condition; 3) a placebo condition; or 4) a no
feedback intervention. Studies were only included if the exercise or
activity performed by participants was the same in both compari-
son groups. For example, if the AF intervention was part of an
additional exercise or activity practiced only by the experimental
group, this would be excluded. Therefore, studies with additional or
different exercises or activities in one group compared to the other
were excluded.
2.1.3. Participants
We included trials of human participants of any age with
musculoskeletal dysfunction in any body part, with any diagnosis,
e.g. sprains, strains, dislocation, or chronic pain conditions.
2.1.4. Interventions
The included studies were required to contain at least one
intervention inducing an IFA or an EFA. An EFA was defined as the
participant focussing their attention on the effects of their body’s
actions on the environment. Conversely, an IFA was defined as the
participant focussing their attention on their own body and it ac-
tions. Studies of any duration, frequency or intensity of program
were included.
2.1.5. Outcome measures
Function and pain were the primary outcomes. Examples of
measures of function include walking distance, ability to perform
overhead activities, return to sport, and return to work. Examples of
pain measures include visual analogue scale (VAS) for pain, and the
McGill pain questionnaire. Secondary outcomes were examined
when reported. Examples of secondary measures include gait pa-
rameters, balance, co-ordination, proprioception, muscle weakness,
muscle length, and range of motion, incidence of dysfunction reoc-
currence, prevalence and nature of adverse events and other
complications.
2.2. Search methods for identification of papers
Electronic searches of the following databases were undertaken:
MEDLINE (1946eSeptember 2012); Embase Classic þ Embase
(1947e2012, September 4); CINAHL Plus Full Text (1982eSeptember
2012); AMED (1985eSeptember 2012); Cochrane Library; Web of
Science (1956eSeptember 2012); Scopus (1966eSeptember 2012);
OTseeker (http://www.otseeker.com/) (September 2012); Physio-
therapy Evidence database (PEDro, http://www.pedro.fhs.usyd.edu.
au/index.html) (September 2012); REHABDATA (1956e2012)
(http://www.naric.com/research/rehab/default.cfm) (September
2012). Search strategies were developed in MEDLINE using a com-
bination of controlled vocabulary (MeSH) and free text terms
(Table 1). This was modified to suit other databases. Grey literature
outside the above databases was not searched.
2.3. Selection of studies
Two review authors (CS and PvV) read the titles of the iden-
tified references and eliminated any obviously irrelevant studies.
460 C. Sturmberg et al. / Manual Therapy 18 (2013) 458e467
Table 1
Medline search strategy.
1. exp musculoskeletal system/
2. abdomen/or groin/or inguinal canal/or exp back/or exp extremities/or exp
head/or exp neck/or exp pelvis/or exp perineum/or thorax/or thoracic wall/
3. (head or fac* or temperomandibular joint* or neck or shoulder or arm or
elbow or forearm or wrist or hand or finger* or thumb* or vertebral column or
cervical vertebra* or thoracic vertebra* or lumbar vertebra* or sacral
vertebra* or back or chest or abdomen or pelvis or perine* or hip or thigh or
knee or leg or ankle or foot or toe*).mp. [mp ¼ protocol supplementary
concept, rare disease supplementary concept, title, original title, abstract,
name of substance word, subject heading word, unique identifier]
4. 1 or 2 or 3
5. exp musculoskeletal system/in [Injuries]
6. exp arm injuries/or exp athletic injuries/or exp back injuries/or exp pelvic
floor/in or exp burns/or exp dislocations/or exp fractures, bone/or exp
fractures, cartilage/or exp hand injuries/or exp hip injuries/or exp leg
injuries/or exp neck injuries/or exp soft tissue injuries/or exp “sprains and
strains”/or exp tendon injuries/or exp pain or exp Neck Pain/or exp Facial
Pain/or exp Back Pain/or exp Patellofemoral Pain Syndrome/or exp Pain
Perception/or exp Pelvic Girdle Pain/or exp Pain Management/or exp
Myofascial Pain Syndromes/or exp Shoulder Pain/or exp Low Back Pain/or
exp Nociceptive Pain/or exp Pain, Referred/or exp Pain, Intractable/or exp
Pain Clinics/or exp Chronic Pain/or exp Pelvic Pain/or exp Musculoskeletal
Pain/or exp Complex Regional Pain Syndromes/
7. (ligament sprain* or ligament strain* or muscle strain* or muscle sprain* or
cartilage tear* or meniscal tear* or muscle tear* or ligament tear* or tendon
tear* or subluxation or dislocation or fractur* or stress fractur* or overuse
injur* or overuse syndrome* or repetitive strain injur* or rsi or muscle
imbalance or musculoskeletal dysfunction or muscular dysfunction or
skeletal dysfunction or fascial dysfunction or connective tissue dysfunction or
movement dysfunction or pain or chronic pain or complex regional pain
syndrome or reflex sympathetic dystrophy or causalgia or sudeck’s atrophy
or headache or facial pain or neck pain or back pain).mp. [mp ¼ protocol
supplementary concept, rare disease supplementary concept, title, original
title, abstract, name of substance word, subject heading word, unique
identifier]
8. 5 or 6 or 7
9. exp Motor Skills/or exp Motor Activity/or exp Learning/or movement/or
locomotion/or exp Psychomotor Performance/or exp intention/
10. exp Rehabilitation/or exp “Physical Therapy (Specialty)”/or exp Physical
Therapy Modalities/or Kinesiology, Applied/or exp “Physical Education and
Training”/or exp Exercise Therapy/or exp “Activities of Daily Living”/or exp
Occupational Therapy/or exp “Task Performance and Analysis”/or exp
Athletic Performance/
11. (motor learning or rehabilitation or physiotherapy or physical therapy or
manual therapy or exercise therapy or motor skills or material based
occupation or material-based occupation or strength* or endurance training
or endurance exercise or muscle retraining).mp. [mp ¼ protocol
supplementary concept, rare disease supplementary concept, title, original
title, abstract, name of substance word, subject heading word, unique
identifier]
12. 9 or 10 or 11
13. exp Feedback/or exp Cognition/or exp “Task Performance and Analysis”/or
exp Biofeedback, Psychology/or exp Attention/or exp Reaction Time/
14. (feedback or feedforward or augmented feedback or biofeedback or
instruction* or direction* or cue*).mp. [mp ¼ title, abstract, original title,
name of substance word, subject heading word, protocol supplementary
concept, rare disease supplementary concept, unique identifier]
15. 13 or 14
16. (attentional focus or attentional foci or internal focus or internal-focus or
intrinsic focus or intrinsic-focus or external focus or external-focus or
extrinsic focus or extrinsic-focus or extrinsic feedback or attention or Singer
five-step approach or Singer five step approach or attention conditions or
action effect hypothesis or action-concept model or effect anticipation or
theory of event coding or attention to movements or attention to effects of
movements or movement effects or occupationally embedded exercise).mp.
[mp ¼ protocol supplementary concept, rare disease supplementary concept,
title, original title, abstract, name of substance word, subject heading word,
unique identifier]
17. 4 and 8 and 12 and 15 and 16
18. limit 17 to (English language and humans)
Studies denoted irrelevant by both authors were excluded. Ab-
stracts for the remaining studies were then obtained. Using the
inclusion criteria listed hierarchically (study design, English lan-
guage, comparison, participants, interventions, outcomes), two
review authors (CS and PvV) independently classified studies as
‘possibly relevant’ or ‘definitely irrelevant’, according to the first
inclusion criteria a study failed to meet. If both review authors
identified a study as ‘definitely irrelevant’ it was excluded at this
stage. Following this process the full text of those studies cat-
egorised as ‘possibly relevant’ were retrieved and independently
reviewed by two review authors (CS & PvV) who classified them as
‘include’, ‘exclude’ or ‘unsure’. Trials classified as ‘exclude’ by both
review authors were excluded. At any stage where disagreement
occurred between review authors, or a decision could not be
made, consensus was reached through discussion, including a
third review author (JM) as necessary. Reference lists of all
included articles were hand searched and retrieved titles under-
went the same process described above.
2.4. Documentation of methodological quality
Two review authors (CS and PvV) independently assessed the
methodological quality of included studies using the Grading of
Recommendations Assessment, Development and Evaluation
(GRADE) system (Guyatt et al., 2008) (Table 3). Any disagreements
concerning quality assessment between the two review authors
were resolved through discussion, including a third review author
(JM) as necessary. Level of agreement between the two review
authors was assessed using the Cohen’s Kappa statistic.
2.5. Data extraction and synthesis
Two review authors (CS and PvV) independently extracted data
from the studies. Details of the included studies were recorded
including sample size, patient demographics (e.g. musculoskeletal
dysfunction, inclusion and exclusion criteria of participants)
(Table 2) and treatment interventions, outcome measures and re-
ported effect (Table 4). Interventions and AF instructions/feedback
were described in detail. Means, ranges and standard deviations of
outcome measures were reported where available.
Comparative statistical analyses were to be performed if suffi-
cient number of studies within each comparison type category
studied similar populations and employed the same measures.
Studies to be included in meta-analysis were to be analysed using
the Cochrane Collaboration’s Review Manager software, RevMan 5,
and following the Cochrane Handbook for Systematic Reviews of
Interventions (Higgins and Green, 2011). For RCTs, for each com-
parison, the study results for measures of function, pain, secondary
outcomes, dysfunction reoccurrence and adverse events were to be
used if documented. If possible, we would use intention-to treat
analyses. We intended to analyse dichotomous data using the odds
ratio and 95% confidence interval employing a fixed-effect model
with exploration of sources of heterogeneity. We intended to
analyse continuous outcomes as the standardised mean difference
and 95% confidence intervals. If statistical pooling of results was not
appropriate the findings were to be summarised in narrative form.
3. Results
3.1. Flow of studies through the review
Searches of the electronic bibliographic databases identified 2676
records (317 from Medline, 757 from Embase Classic þ Embase, 614
from CINAHL, 16 from AMED, 270 from The Cochrane Library (9 from
CENTRAL, 4 from DARE, 251 from Trials, 6 from Methods Studies,
0 from Technology Assessment, Economic Evaluations and Cochrane
Groups), 53 from Web of Science, 126 from Scopus, 0 from OTseeker,
523 from PEDro, and 0 from REHABDATA). Duplicates were removed
Table 2
Patient demographics.

Study Sample size Age in yearse (Mean) Female: Type of dysfunction Duration of dysfunction Inclusion criteria Exclusion criteria
Expa:Compb [Range] {SD} Male (Mean) {SDa}

Laufer et al., 2007 20:20 [19e33] 4:36 Lateral ankle sprain, Not reported First or recurrent lateral ankle sprain Evidence of concomitant additional injury,
grade 1 or 2 at grade 1 or 2, sustained less than history of prior surgical procedures in
4 months earlier, able to bear full lower extremities, neurologic or
weight on injured leg vestibular disorder, previous training
on stabilometer
Rotem-Lehrer and IFAc group: 20 (20.9) 0:36 Lateral ankle In weeks: First or recurrent lateral ankle sprain at Evidence of concomitant additional
Laufer, 2007 EFAd 16 [19e33] sprain, grade 1 or 2 IFA group: (4.3) {3.1} grade 1 or 2 in accordance to the injury, previous ankle surgery or
EFA group: (4.1) {2.4} classifications described by Crichton other pathological conditions or
et al. (1995), sustained less than surgical procedures in either lower
4 months previously, able to bear extremity, neurological, vestibular
full weight on injured leg with only or other balance disorder, previous
mild discomfort training on stabilometer
Budzynski et al., 1973 A: 6 (36) 16:2 Frequent tension In years: Frequent tension headaches, average 22 question telephone questionnaire
B: 6 [22e44] headaches group A (9.6) intensity 0.3 or above (moderate) to screen out applicants who appeared
C: 6 group B (6.8) during the two week baseline period to have other than muscle contraction
group C (6.7) headaches; thorough medical and

C. Sturmberg et al. / Manual Therapy 18 (2013) 458e467

psychiatric examination to rule
out the possibility of neurological
and other organic disorders and
to confirm the diagnosis of tension
headache; evidence of severe
psychological problems as detected
by Minnesota Multiphasic Personality
Inventory were eliminated from
main study
Stenn et al., 1979 6:5 (23) 9:2 Myofascial pain Not reported Temperomandibular joint pain for Radiologic evidence of bony joint
[16e34] dysfunction syndrome at least 1 year; complete failure of damage; evidence of physical trauma
conservative treatment (physiotherapy, to head or neck in subjects history
occlusal splint, ultrasound, or tranquilizers)
to provide any lasting relief of symptoms ;
age between 16 and 55 years;
Yip and Ng, 2006 13:13 (32.5) 16:10 Patellofemoral pain Not reported Insidious onset of patellofemoral pain Degenerative changes on radiography,
[22e55] for >6 months; positive apprehensive chondral damage, meniscal lesion,
{8.8} test; knee pain with at least 2 of: ascending ligamentous instability, previous
stairs, descending stairs, squatting, kneeling, knee surgery or traumatic injury,
prolonged sitting, hopping, jumping signs of acute inflammation
Thiengwittaporn 30:30 Control 54:4 Knee osteoarthritis Not reported Diagnosis of knee osteoarthritis; over Experiencing knee operation; history
et al., 2009 (60.7) 50 years old; able to walk without of neuromuscular disorder, flexion
{8.260} a supporting device; have previously contracture of >10 degrees; unable
Intervention been instructed in the quadriceps exercise to talk or communicate
(61.9) however are unable to perform it correctly
{8.717} or have never been instructed in the
quadriceps exercise
Alexander and 6:5 (57.2) 5:6 Above knee or below Not reported Patients with amputation referred for None reported
Goodrich, 1978 knee amputation, for prosthetics training during a
reasons of diabetes or six-month period
trauma
a
Experimental group.
b
Comparison group.
c
Internal focus of attention.
d
External focus of attention.
e
Where means, SDs or range are missing, these were not reported; SD e standard deviation.

461
462 C. Sturmberg et al. / Manual Therapy 18 (2013) 458e467
Table 3
Quality of evidence summary using GRADE system.
Study High Moderate
Laufer et al., 2007 U
Rotem-Lehrer and Laufer, 2007 U ance by stabilising the platform”; (Laufer et al., 2007, p 106)).
Budzynski et al., 1973
Stenn et al., 1979
Yip and Ng, 2006
Thiengwittaporn et al., 2009
Alexander and Goodrich, 1978
by the first author. Fig. 1 shows the study selection process and
reasons for abstract exclusion.
In all, agreement was made to include 7 relevant studies
(Budzynski et al., 1973; Alexander and Goodrich, 1978; Stenn et al.,
1979; Yip and Ng, 2006; Laufer et al., 2007; Rotem-Lehrer et al.,
2007; Thiengwittaporn et al., 2009). Excluded studies and reason
for exclusion are shown in Appendix A.
Three studies were excluded as the AF intervention was part of
an additional exercise or activity practiced only by the experi-
mental group (Asfour et al., 1990; Dursun et al., 2001; Magnusson
et al., 2008). For example, Magnusson et al. (2008) compared a
standard back rehabilitation program with the standard program
plus a separate exercise using an EFA where a triaxial goniometer
was strapped to participants’ backs and linked to a computer screen
displaying an icon representing the position of the participants’
trunk. The EFA group had significantly greater improvements in
VAS pain scores, short form-36 physical functioning, roles limita-
tions and bodily pain scores and ROM, which were maintained at 6
month follow-up.
3.2. Study characteristics
3.2.1. Design
Six of the included studies were randomized controlled trials
(Budzynski et al., 1973; Stenn et al., 1979; Yip et al., 2006; Laufer
et al., 2007; Rotem-Lehrer et al., 2007; Thiengwittaporn et al.,
2009), and one was a controlled cohort study (Alexander et al.,
1978). Sample sizes ranged from 11 to 60. Overall the mean number
of participants was 28.86 (SD 17.83). Table 2 provides an overview
of study sample sizes.
3.2.2. Comparisons
All studies compared the effects of two types of interventions, at
least one of which contained instructions or feedback inducing an
IFA or an EFA. Budzynski et al. (1973) also compared a third, no
intervention group. This group was excluded from our compari-
sons, as this comparison did not meet the inclusion criteria.
3.2.3. Participants
The participant demographics of the total 202 participants in
the seven included studies are summarised in Table 2. Studies
involved participants with a variety of types of musculoskeletal
dysfunction, with two studies focused on lateral ankle sprain
(Laufer et al., 2007; Rotem-Lehrer et al., 2007), one on patellofe-
moral pain (Yip et al., 2006), one on myofascial pain dysfunction
syndrome in the temporomandibular (TMJ) region (Stenn et al.,
1979), one on frequent tension headache (Budzynski et al., 1973),
one on osteoarthritis (OA) in the knee(s) and one investigated on
participants with an amputated lower limb (Alexander et al., 1978).
3.2.4. Interventions
3.2.4.1. Internal focus versus external focus. Two of the included
studies (Laufer et al., 2007; Rotem-Lehrer et al., 2007) reported on
single-leg stance training on an unstable surface in participants
with lateral ankle sprain. The authors compared postural stability
performance following the provision of instructions inducing an
Low IFA (“keep your balance by stabilising your body”; (Laufer et al.,
2007, p 106)) with instructions inducing an EFA (“keep your bal-
U
U 3.2.4.2. Internal focus with biofeedback versus other focus condition.
U
In two other studies (Budzynski et al., 1973; Stenn et al., 1979), an
U
U IFA with biofeedback was compared to another AF during relaxa-
tion training. In both studies, both the experimental and control
groups were to focus their attention on relaxation. The experi-
mental groups received auditory biofeedback related to the EMG
activity of the masseter (Stenn et al., 1979) and frontalis (Budzynski
et al., 1973) muscles respectively, thus inducing an IFA in the
participant. The comparison groups received no biofeedback (Stenn
et al., 1979) or were told to focus on pseudofeedback consisting of
meaningless clicks, to assist them in keeping out intruding
thoughts (Budzynski et al., 1973).
3.2.4.3. IFA with biofeedback versus no focus condition. Two further
studies (Yip et al., 2006; Thiengwittaporn et al., 2009) examined
the effectiveness of home exercise programs alone (no AF)
compared with the same home exercise program undertaken with
an IFA. Thiengwittaporn et al. (2009) constructed a quadriceps
education device (QED) to prompt patients through visual and
auditory feedback to complete seated knee extensions correctly.
Yip and Ng (2006) provided real-time normalised visual feedback
via home EMG units to the experimental group, with instructions to
selectively increase activity in vastus medialis oblique (VMO) and
decrease activity in vastus lateralis (VL) during home exercise. The
comparison groups in both these studies received no feedback. The
remaining study included in the review compared customary
prosthetic training with and without video feedback (Alexander
et al., 1978). Both groups were filmed once weekly, from front, side
and back, whilst walking 100 feet. The IFA group watched this
footage of themselves immediately after completing the walk and
were encouraged to comment on their gait pattern and how to
improve it. They then repeated the walking task and viewed the
footage again. The comparison group did not have access to the
footage until after the final assessment.
3.2.5. Outcome measures
The primary outcomes of interest were function and pain.
Approximately half the studies included some measure of function
and/or pain as follows: presence/absence of limitation and/or di-
rection of mandible on opening (Stenn et al., 1979); perceived pain
severity using the Patellofemoral Pain Syndrome Severity Scale (Yip
and Ng 2006); subjective pain ratings on a 10-point scale, (Stenn
et al., 1979); average headache scores on a 5-point scale
(Budzynski et al., 1973).
Included studies also reported the measurement of the
following secondary outcomes: gait parameters via the Amputee
Gait Rating Scale (Alexander and Goodrich, 1978); postural stability
(Laufer et al., 2007; Rotem-Lehrer and Laufer, 2007); isokinetic
strength of knee extension (Yip and Ng, 2006); muscle tension in
frontalis and masseter muscles respectively (Budzynski et al., 1973;
Stenn et al. 1979); patellar alignment (Yip and Ng 2006); person-
ality changes on the Minnesota Multiphasic Personality Inventory
(Budzynski et al., 1973) and ability to correctly perform seated knee
extension exercises (Thiengwittaporn et al., 2009). None of the
studies reported any adverse events.
3.2.6. Methodological quality
The methodological assessment is reported in Table 3 and the
percentage agreement between reviewers on items on the GRADE
Table 4
Results.

Study Comparison interventions Long term Relevant outcome measures Reported effect Authors conclusion
follow up

Laufer et al., 2007 Single-leg stance on injured ankle on Biodex None Measures of postural stability: Overall No significant differences Superior improvement
Stability System (BSS) at stability levels stability Index (OSI) e the variance between groups in EFA group during
6 and 4, ten 20-s trials with 30 s rest of foot platform displacement in Result at Stability level 6: acquisition phase at
between each trial, performed on three degrees in all directions; Significant improvement level 6. Improvements
consecutive days. Anterior/Posterior Stability between pre-training and maintained in the
Instructions before each training session Index (APSI); Medial/Lateral Stability post-training OSI and APSI retention phase by
to EFA group “keep your balance by Index (MLSI) at stability levels 6 and 4. scores only in EFA group both groups, at
stabilising the platform”; to IFA group Assessed pre- and posttraining and (mean change scores: OSI levels 4 and 6.
“keep your balance by stabilising your body” 48 h after the last session (retention test) 2.40 versus-1.33; APSI
3.22 versus 0.49);
Result at Stability level 4:
Significant improvement
between pre-training and
post-training OSI and APSI
scores in both groups
(OSI: IFA pre-test: 9.02,

C. Sturmberg et al. / Manual Therapy 18 (2013) 458e467

post-test 8; EFA pre-test
10.31, post-test 7.42; APSI:
IFA pre-test: 6.24, post-test
4.75; EFA pre-test 7.16,
post-test 3.67)
Rotem-Lehrer Single-leg stance on injured ankle on None Measures of postural stability: Overall No significant difference Use of EFA in training
and Laufer, 2007 Biodex Stability System (BSS) at stability stability Index e the variance of foot between groups of postural control
levels 6 and 4, ten 20-s trials with platform displacement in degrees No significant difference on moderately
30 s rest between each trial, performed in all directions; Anterior/Posterior between post-training unstable surfaces
on three consecutive days. Stability Index; Medial/Lateral Stability and retention in improves postural
Instructions before each training Index at stability level 2. Assessed either group control on an
session to EFA group “keep your before the initial training session and Significant improvement untrained, less
balance by stabilising the platform”; 48 hours following the last session between pre-training and stable surface.
to IFA group “keep your balance by post-training scores only in
stabilising your body” EFA group (mean change
scores OPI 5.45, APSI 4.14,
MLSI 3.36)
Budzynski Groups A and B: 16 relaxation sessions of 3 months (all patients Frontalis electromyographic (EMG) Significant differences between Training in relaxation
2et al., 1973 30 min each in laboratory, ideally twice in groups A and B) levels, measured during two no-feedback groups favouring Group A: of the forehead
weekly, for total of 16 sessions over 8 weeks. 18 months (4 of sessions pre-training and during all  decline in frontalis EMG muscles with EMG
Group A (experimental group): auditory 6 patients in Group A) training sessions; average headache scores, activity during last two weeks feedback is useful
feedback proportional to EMG activity from taken from patients charting of headache of treatment and at 3 month in alleviating
frontalis muscle. Instructed to decrease click intensity recorded every waking hour follow up (Group A mean tension headache
rate to lower muscle tension on scale of 1 to 5; Minnesota Multiphasic 3.92 versus 8.43 mV in Group B)
Group B (comparison group): noncontingent Personality Inventory completed before Mean difference ¼ 4.51 mV
“pseudofeedback” consisting of a recording and after training period  decline in average headache scores
of the biofeedback generated by Group A.  reduction in hysteria profile
Instructed to keep attention focused on
varying rate of clicks to help keep out
intruding thoughts
Groups A and B: relaxation practice outside
of laboratory for two 15e20 min periods
every day. No specific relaxation instructions
given except to relax in the same way they
had in the lab (without any instruments)
Group C: no training
Stenn et al., 1979 All subjects: provided in vivo relaxation 3 months after EMG measures of masseter muscle Significant differences The addition of
instructions (Jacobson, 1964) in seated position end of treatment tension; subjective pain ratings favouring experimental biofeedback to a

463
(continued on next page)
Table 4 (continued )

464
Study Comparison interventions Long term Relevant outcome measures Reported effect Authors conclusion
follow up

for 30 min session, EMG electrodes attached period (9 of on 10-point scale, patient recorded group: relaxation and
over masseter area 11 patients) 5 times a day; symptoms and signs  decline in perceived pain cognitive behaviour
Experimental group: auditory biofeedback of myofascial pain dysfunction (mean 0.817, SDa 0.818 versus modification
throughout the relaxation period with tone syndrome (MPDS) i.e. presence/absence mean 2.55, SD 1.74); program increases
varying with muscle tension. Control: of limitation and/or direction of Mean difference ¼ 1.733 the effectiveness
no biofeedback mandible on opening, abnormal joint  decline in signs and of the program in
All subjects: following EMG, 30 min session with sounds, tenderness in muscles of symptoms of MPDS (mean 2.33, decreasing pain
psychologist for cognitive behaviour modification mastication or condylar leads, assessed SD 1.51 versus mean 5.0, SD 3.39); and signs and
(total 7 sessions over 8 weeks) by blinded physician Mean difference ¼ 2.67 symptoms of MPDS
No difference between groups
in muscle tension
Yip and Ng, 2006 Comparison group: home exercise protocol: 15 min, None Perceived pain severity using No difference between The addition of
daily. 5 min of flexibility exercises; 3 sets of Patellofemoral Pain groups in all outcome EMG biofeedback
10 repetitions of quadriceps strengthening Syndrome Severity measures to exercise
exercises with emphasis on vastus medialis Scale; patellar alignment program on VMO
obliquus (VMO) recruitment; balance and proprioception using McConnel test of activation had
training, plyometric and agility training gliding, tilting and rotation; no measurable

C. Sturmberg et al. / Manual Therapy 18 (2013) 458e467

Experimental group: identical home exercise program isokinetic strength of knee effect at 8 weeks
with surface EMG biofeedback providing real-time extension peak torque and
normalised visual biofeedback signals of VMO total work
and vastus lateralis (VL) according to their intensity
of contraction during exercises. Subjects asked to
selectively increase activity of VMO whilst maintaining
relatively stable activity in VL
Thiengwittaporn Comparison group: seated knee extension with None Accuracy of seated knee extension Significant difference Patients who use the
et al., 2009 minimum 6 s hold and 20 s rests, 3 sets of program. After 2 weeks of home between groups in QED can more
10 repetitions, 3 times daily exercise, patients asked to exercise accuracy, accurately perform
Experimental group: seated knee extension as above, demonstrate 10 knee extensions. favouring experimental the knee extension
using Quadriceps Educational Device (QED) which used Considered accurate if patient group (79.3% accurate in exercise than those
lights and auditory feedback, initiated once patients fully extended knee and experimental group; who are instructed
achieved full knee extension, to prompt patients to maintained this position for 6 to 28.6% accurate in by a doctor alone
time exercises correctly and complete 10 repetitions 15 s then returned knee to 90  , comparison group)
for a minimum of 8 repetitions Mean difference ¼ 50.7%
Alexander and Comparison group: customary prosthetic training None A panel of 14 (3 physiatrists and No statistical analyses Video feedback
Goodrich, 1978 program with weekly filming of ambulation over 11 physical therapists) rated the performed, however in improved the
100 feet, not permitted to view footage footage of patients’ initial effort 5 of 6 cases, change scores outcome of training
Experimental group: customary prosthetic training (as soon as able to ambulate outside were larger in the of most amputees
program with weekly filming as above. Patients viewed the parallel bars with their prostheses, experimental group than in this sample
footage immediately after task and were asked to comment usually 1e4 days after artificial limb the comparison group
on their feelings on seeing themselves walk, their gait pattern was delivered) and final effort 4 weeks
and how to improve it. Investigators did not point out errors later, on the Amputee Gait Rating Scale.
or call the patient’s attention to a specific action. Patients Mean rating for initial and final
repeated the task and viewed footage again, performances of each patient was calculated.
without a request for comments The difference between the means
(difference score) was considered the
measure of change
a
Standard deviation.
 C. Sturmberg et al. / Manual Therapy 18 (2013) 458e467
Fig. 1. Flow of information through different phases of the systematic review
system of rating methodological quality was 86% with
Kappa ¼ 0.696 (p ¼ 0.053; CI 0.170e1.0). Generally, studies scored
well with regards to the inclusion of intention to treat analysis,
blinding of outcome assessor, baseline similarity, identical treat-
ment of groups, outcome measurement and appropriate statistical
analysis. However, methods of randomisation and allocation
concealment were poorly described. As studies investigated phys-
ical therapy modalities, participants were generally aware of their
treatment allocation.
3.3. Effects of interventions
Due to the heterogeneity of studies with regard to type of
musculoskeletal dysfunction and outcome measures, it was inap-
propriate to perform meta-analysis. Therefore the results are
summarized in narrative form.
3.3.1. IFA versus EFA
Neither Laufer et al. (2007) nor Rotem-Lehrer and Laufer (2007)
provided any outcomes related to function or pain. Both Laufer et al.
(2007) and Rotem-Lehrer and Laufer (2007) reported statistically
significant improvements in postural stability at stability level 6 (8
being the most stable foot platform setting and 1 being the least
stable) from pre-training to post-training only in groups receiving
EFA. Rotem-Lehrer and Laufer (2007) found this not only at the
trained levels of stability, but also when participants were tested on
a more unstable surface.
3.3.2. IFA with biofeedback versus other focus condition
Budzynski et al. (1973) and Stenn et al. (1979) both found sig-
nificant differences favouring IFA in measures of pain: average
465
headache scores (Budzynski et al., 1973) and perceived pain in the
TMJ region (Stenn et al., 1979). Stenn et al. (1979) also reported a
decline in symptoms and signs of myofascial pain dysfunction
including the presence/absence of limitation and/or direction of
mandible on opening, again favouring IFA, but reported no differ-
ences between groups in masseter muscle tension, as measured via
EMG electrodes placed over the masseter area. Budzynski et al.
(1973) found differences between participant groups favouring
IFA in both frontalis muscle activity and hysteria profile on the
Minnesota Multiphasic Personality Inventory.
3.3.3. IFA with biofeedback versus no focus condition
Yip and Ng (2006) found no differences between groups in the
outcomes measured: pain, patellar alignment or isokinetic knee
extension strength. Alexander and Goodrich (1978) and
Thiengwittaporn et al. (2009) provided no outcomes related to
function or pain. In terms of secondary outcomes, Thiengwittaporn
et al. (2009) found a significant difference between groups in ac-
curacy of the knee extension exercise favouring IFA. Alexander and
Goodrich (1978) performed no statistical analyses, however in 5 of
6 cases, change scores were larger in the experimental group
(receiving feedback inducing an IFA) than the comparison group.
4. Discussion
The results of this review suggest that there is insufficient evi-
dence to determine if an EFA or an IFA is more effective in
improving function and decreasing pain in patients with muscu-
loskeletal dysfunction. Only two of the included studies, both
investigating ankle sprain (Laufer et al., 2007; Rotem-Lehrer and
Laufer, 2007), were designed specifically to examine the effect of
466 C. Sturmberg et al. / Manual Therapy 18 (2013) 458e467
differing attentional foci. The remaining studies’ results are
confounded by the provision of biofeedback to the experimental
group only, making it impossible to differentiate the effect of the
feedback from the effect of the AF (Budzynski et al., 1973; Stenn
et al., 1979; Yip and Ng, 2006; Thiengwittaporn et al., 2009). The
findings from this review suggest there is limited evidence from
two studies suggesting an EFA may be superior to an IFA during
motor learning for postural stability in participants with lateral
ankle sprain. The only outcome measured in these studies was
variance in stabilometer platform displacement in all directions. As
authors (Laufer et al., 2007) themselves note, the amount of
improvement in postural control that significantly influences
function or is clinically significant is yet to be established. These
two studies did not report any patient-related pain or self-reported
function outcomes, so evidence is insufficient to know whether an
EFA is more effective in the treatment of patients with musculo-
skeletal dysfunction.
The finding that an EFA is more effective for improving motor
performance concurs with the findings of other reviews of AF in
healthy (Wulf and Prinz, 2001) individuals, as well as populations
with stroke (van Vliet and Wulf, 2006) and Parkinson’s disease
(Landers et al., 2005; Wulf et al., 2009). This suggests that an EFA
may be more effective for improving motor performance in patients
with musculoskeletal dysfunction, but it is unknown whether this
also improves patient outcomes.
4.1. Strengths and limitations of the current review
The literature search employed a thorough and systematic
strategy during September 2012, using broad, inclusive criteria,
however there are potential sources of bias. Firstly, additional
studies were potentially not identified at the time or may since
have been published. Secondly, the search was limited to studies in
the English language.
Only seven studies were identified for inclusion in this review.
Meta-analysis was not recommended due to the large hetero-
geneity of musculoskeletal conditions and outcome measures in
the studies we retrieved. In order to describe all studies
comparing different attentional foci, studies were included that
did not have a comparison of AF conditions as their stated aim,
but included at least one intervention that clearly induced par-
ticipants to focus their attention internally or externally. While
confounding factors made it impossible to differentiate AF effects
from the effects of feedback versus no feedback in these studies
(Budzynski et al., 1973; Stenn et al., 1979; Yip and Ng, 2006;
Thiengwittaporn et al., 2009), they demonstrate how AF is
inherent to many common clinical scenarios and provide a
template which can be refined by future researchers to enable
the isolation of AF effects.
The duration of symptoms was not reported by the majority of
authors (Alexander et al., 1978; Stenn et al., 1979; Yip et al., 2006;
Laufer et al., 2007; Thiengwittaporn et al., 2009), however, the
stage of dysfunction may have a major bearing on patients’ motor
control. As such, the strategies to improve motor learning may well
vary at different stages of the dysfunction. As duration of symptoms
increases, one would expect motor control to deteriorate. It is
hypothesised, for example, that chronic back pain patients have
impaired motor control (Panjabi, 2006), a hypothesis supported by
findings of decreased proprioception in specific populations of
patients with low back pain (O’Sullivan et al., 2003).
Lastly, level of ability, or stage of learning, significantly impacts
the strategies and conscious attention used to perform and improve
the execution of a task (Magill et al., 2003). Accordingly, the in-
structions and feedback of benefit to novice learners differs from
that of benefit to accomplished learners perfecting a skill (Magill
et al., 2003). It follows that the effect of AF may also change at
different stages of learning. Laufer et al. (2007) and Rotem-Lehrer
and Laufer (2007) specifically included only novice learners, with
no experience on a stabilometer. Thiengwittaporn et al. (2009)
included participants who have never been instructed in the
quadriceps exercise, as well as those who have been previously
instructed however are unable to perform the exercise correctly.
However, information regarding the stage of learning of partici-
pants is missing from the remainder of studies, limiting their
interpretation.
4.2. Implications for research
Overall, the studies identified are insufficient to address the
objective of the review, to determine whether instructions and
feedback inducing an external focus of attention are more effective
than instructions and feedback inducing an internal focus of
attention, in improving function and decreasing pain in patients
with musculoskeletal dysfunction. Only seven studies could be
included, encompassing six different musculoskeletal dysfunc-
tions. In addition, too few studies were designed to assess the
effect of different attentional foci independent of other in-
terventions. It is therefore not possible at present to make firm
recommendations for clinical practice. However, it is encouraging
that the two studies that were designed specifically to examine
the effect of differing attentional foci (Laufer et al., 2007; Rotem-
Lehrer and Laufer, 2007) did show a positive effect for feedback
inducing an external focus of attention. Further research is needed
to determine whether this will be a consistent finding that can be
applied to musculoskeletal practice in the current health and so-
cial care context.
This review provides a synthesis to assist future researchers to
design studies capable of isolating AF effects. To address the limi-
tations detailed above, future research should be designed with AF
as the only variable; confounding factors, such as feedback given to
one group but not the other, be considered and minimised;
methodology, particularly the randomisation process and alloca-
tion concealment, be clearly documented; participant variables,
such as stage of learning and chronicity of dysfunction be investi-
gated in relation to AF; the scope of musculoskeletal dysfunctions
under investigation be widened; and the longevity of effects be
examined via long term follow-up.
5. Conclusions
There are no studies that directly compare an EFA to an IFA and
report its effect on pain or functional outcomes in patients with
musculoskeletal pain. There is limited evidence from two studies
that an EFA may be superior to an IFA for improving motor per-
formance in patients with lateral ankle sprain, but there is insuf-
ficient evidence at this stage to make firm recommendations for
clinical practice.
6. Conflict of interest statement
I certify that no party having a direct interest in the results of the
research supporting this article has or will confer a benefit on me or
on any organization with which I am associated.
Acknowledgements
The authors would like to thank Debbie Booth, Faculty Librarian
at The University of Newcastle who assisted with search strategies
for this review. The authors report no conflict of interest.
 C. Sturmberg et al. / Manual Therapy 18 (2013) 458e467
Appendix A. Excluded studies
First author, year Reason for exclusion
Lehrer, 2009 Not in English
Brewer, 2006 Participants did not have
musculoskeletal dysfunction
Stevens, 1977 Participants did not have
musculoskeletal dysfunction
Butler, 2010 Intervention not aimed
at inducing an AF
Mosely, 2005 Intervention not aimed
at inducing an AF
Asfour, 1990 Not valid comparison
condition e AF intervention
was part of an additional exercise
or activity practiced only by the
experimental group
Dursun, 2001 Not valid comparison
condition e AF intervention was
part of an additional exercise or
activity practiced only by the
experimental group
Magnusson, 2008 Not valid comparison
condition e AF intervention was
part of an additional exercise or
activity practiced only by the
experimental group
Anderson Worth, 2007 Not valid comparison condition
Bajaj, 2010 Not valid comparison condition
Bush, 1985 Not valid comparison condition
Dalen, 1986 Not valid comparison condition
Dohrmann, 1978 Not valid comparison condition
Keefe, 1980 Not valid comparison condition
Sousa, 2009 Not valid comparison condition
Teyhen, 2005 Not valid comparison condition
Liew, 2011 Full text unobtainable
References
Alexander J, Goodrich R. Videotape immediate playback: A tool in rehabilitation of
persons with amputations. Arch Phys Med Rehabil 1978;59(3):141e4.
Asfour S, Khalil T, Waly S, Goldberg M, Rosomoff M, Rosomoff R. Biofeedback in back
muscle strengthening. Spine 1990;15(6):510e3.
Budzynski T, Stoyva J, Adler C, Mullaney D. EMG biofeedback and tension headache:
a controlled outcome study. Psychometric Med 1973;35(6):484e96.
Cusimano-Reaston M, Carney B. Legal changes necessitates proactive management
of musculoskeletal disorders: the role of electrodiagnostic functional assess-
ment soft tissue management program. IEEE Eng Med Biol Soc 2011:7570e3.
Dursun N, Dursun E, Kilic Z. Electromyographic biofeedback-controlled exercise
versus conservative care for patellofemoral pain syndrome. Arch Phys Med
Rehabil 2001;82(12):1692e5.
467
Guyatt GH, Oxman AD, Vist GE, Kunz R, Falck-Yttr Y, Alonso-Coello P, et al. GRADE e
an emerging consensus on rating quality of evidence and strength of recom-
mendations. Br Med J 2008;336:924e6.
Handoll HH, Gillespie WJ, Gillespie LD, Madhok R, Handoll HHG, Al E. Moving to-
wards evidence-based healthcare for musculoskeletal injuries: featuring the
work of the Cochrane Bone, Joint and Muscle Trauma Group. J R Soc Promot
Health 2007;127:168e73.
Higgins JPT, Green S. Cochrane handbook for systematic reviews of interventions.
Version 5.1.0 (Updated March 2011). The Cochrane Collaboration; 2011.
Hochanandel CD, Conrad DE. Evolution of an on-site industrial physical therapy
program. J Occup Med 1993;35:1011e6.
Landers M, Wulf G, Wallmann H, Guadagnoli M. An external focus of attention
attenuates balance impairment in patients with Parkinson’s Disease who have a
fall history. Physiotherapy 2005;91:152e8.
Laufer Y, Rotem-Lehrer N, Ronen Z, Khayutin G, Rozenberg. Effect of attention focus
on acquisition and retention of postural control following ankle sprain. Arch
Phys Med Rehabil 2007;88(1):105e8.
Magill RA. Motor learning and control: concepts and applications. 9th ed. New York:
McGraw Hill; 2007.
Magnusson M, Chow D, Diamandopoulos Z, Pope M. Motor control learning in
chronic low back pain. Spine 2008;33(16):E532e8.
Martin BI, Deyo RA, Mirza SK, Turner JA, Comstock BA, Hollingworth W, et al. Ex-
penditures and health status among adults with back and neck problems. J Am
Med Assoc 2008;299(6):656e64.
McNevin NH, Shea CH, Wulf G. Increasing the distance of an external focus of
attention enhances learning. Psychol Res 2003;67:22e9.
Mock C, Cherian MN. The global burden of musculoskeletal injuries: challenges and
solutions. Clin Orthop 2008;466:2306e16.
O’Sullivan PB, Burnett A, Floyd AN, Gadsdon L, Logiudice J, Miller D, et al. Lumbar
repositioning deficit in a specific low back pain population. Spine 2003;28:
1074e9.
Panjabi MM. A hypothesis of chronic back pain; ligament subfailure injuries lead to
muscle control dysfunction. Eur Spine J 2006;15:668e76.
Rotem-Lehrer N, Laufer Y. Effect of focus of attention on transfer of a postural
control task following an ankle sprain. J Orthop Sports Phys Ther 2007;37(9):
564e9.
Stenn P, Mothersill K, Brooke R. Biofeedback and a cognitive behavioural approach
to treatment of myofascial pain dysfunction syndrome. Behav Ther 1979;10(1):
29e36.
Thiengwittaporn S, Wetpiryakul P, Foosakun Y, Ngamsom T, Vathanavit P,
Pintongtun J. Comparison of the accuracy of quadriceps isometric exercise be-
tween using Quadriceps Education Device (QED) and not using QED for oste-
oarthritic knee patients: a randomized controlled trial. J Med Assoc Thai
2009;92(Suppl. 6):S33e8.
van Vliet PM, Wulf G. Extrinsic feedback for motor learning after stroke: what is the
evidence? Disabil Rehabil 2006;28(13e14):831e40.
Wulf G, Landers M, Lewthwaite R, Tollner T. External focus instructions reduce
postural instability in individuals with Parkinson’s Disease. Phys Ther
2009;89(2):162e8.
Wulf G, McNevin NH, Shea CH. The automaticity of complex motor skill learning as
a function of attentional focus. Q J Exp Psychol 2001;54:1143e54.
Wulf G, Prinz W. Directing attention to movement effects enhances learning: a
review. Psychon Bull Rev 2001;8:648e60.
Yip S, Ng G. Biofeedback supplementation to physiotherapy exercise programme for
rehabilitation of patellofemoral pain syndrome: a randomized controlled pilot
study. Clin Rehabil 2006;20(12):1050e7.