Professional Documents
Culture Documents
Accelerated Long-Term Forgetting After TIA or Minor
Accelerated Long-Term Forgetting After TIA or Minor
ScienceDirect
Journal homepage: www.elsevier.com/locate/cortex
Sofie Geurts a,*, Sieberen P. van der Werf b,c,d, Vincent I.H. Kwa e and
Roy P.C. Kessels f,g
a
Canisius Wilhelmina Hospital, Department of Medical Psychology, Nijmegen, The Netherlands
b
University of Amsterdam, Department of Psychology, Brain and Cognition, Amsterdam, The Netherlands
c
OLVG, Department of Psychiatry and Medical Psychology, Amsterdam, The Netherlands
d
Rijnstate Hospital, Department of Medical Psychology, Arnhem, The Netherlands
e
OLVG, Department of Neurology, Amsterdam, The Netherlands
f
Radboud University, Donders Institute for Brain, Cognition and Behaviour, Nijmegen, The Netherlands
g
Radboud University Medical Center, Department of Medical Psychology, Nijmegen, The Netherlands
Article history: Cognitive changes after minor stroke or TIA have been reported, with studies describing a
Received 30 March 2017 ‘vascular’ cognitive profile with spared episodic memory. Still, many patients also report
Reviewed 28 September 2017 memory complaints. Studies using long-term forgetting paradigms have detected memory
Revised 4 December 2017 impairment after prolonged intervals in contrast to standard delayed testing in other pa-
Accepted 3 April 2018 tient groups. This study examined whether accelerated long-term forgetting (ALF) is pre-
Published online xxx sent in patients with minor stroke or TIA by comparing one-week delayed recall and
recognition with the performance of a healthy control group. Results revealed that the
Keywords: patients' performance after one week was worse than the controls, in the absence of an
Long-term memory impairment after a short delay. Patients did, however, not report more memory worries
Episodic memory than controls. Possibly, reduced effort, attention or mnemonic strategies may contribute to
Stroke subtle consolidation problems, which go undetected in daily functioning.
Transient ischemic attack © 2018 Elsevier Ltd. All rights reserved.
Vascular cognitive impairment
* Corresponding author. Canisius Wilhelmina Hospital, Department of Medical Psychology, Weg door Jonkerbos 100, 6500 GS, Nijmegen,
The Netherlands.
E-mail address: s.geurts@cwz.nl (S. Geurts).
https://doi.org/10.1016/j.cortex.2018.04.002
0010-9452/© 2018 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Geurts, S., et al., Accelerated long-term forgetting after TIA or minor stroke: A more sensitive measure
for detecting subtle memory dysfunction? Cortex (2018), https://doi.org/10.1016/j.cortex.2018.04.002
2 c o r t e x x x x ( 2 0 1 8 ) 1 e7
Please cite this article in press as: Geurts, S., et al., Accelerated long-term forgetting after TIA or minor stroke: A more sensitive measure
for detecting subtle memory dysfunction? Cortex (2018), https://doi.org/10.1016/j.cortex.2018.04.002
c o r t e x x x x ( 2 0 1 8 ) 1 e7 3
Table 1 e Demographic characteristics for all participants, as well as depressive symptoms, anxiety level, stroke details and
cerebrovascular risk factors for the patients.
Note:
a
Mode; education level was classified using a 7-point scale based on the Dutch educational system, with ‘1’ reflecting less 6 years of primary
school and ‘7’ university degree. HADS ¼ Hospital Anxiety and Depression Scale: 0e7 (normal), 8e11 (mild), 11e14 (moderate) and 14e21
(severe); note that HADS scores were missing for two patients.
b
mRS was not available for 5 patients.
c
For the minor stroke patients only.
research ethical committee of the faculty of social sciences of controls underwent memory testing with the RAVLT and
Radboud University in Nijmegen. completed the Multifactorial Memory Questionnaire (MMQ). It
was agreed upon that they would be contacted by telephone to
2.2. Materials and procedure provide feedback about their baseline results. After one week
the same testing procedure was followed as described for the
The Dutch version of the RAVLT was used to assess verbal patients. At baseline neither the patients nor the controls
learning, recall and recognition performances. This version of were informed about the one-week follow-up assessment of
the RAVLT consists of a list of fifteen non-associated words the word list.
that is presented orally over five trials. The total learning score The MMQ was used to assess subjective memory func-
is the sum of these five trials, with a possible score ranging tioning. In this study, we used two of its three subscales, that
from 0 to 75. The standard testing procedure involves a 30-min is, MMQ Contentment, reflecting the degree of concern and
delayed-recall test (score range: 0e15), followed by a delayed contentment regarding memory functioning (18 items, score
recognition test. In the recognition test, the 15 target words range 0e72) and MMQ Ability, assessing daily forgetfulness (20
were presented among 15 distracters (scoring range: 0e30, i.e., items, score range 0e80). The MMQ Contentment and Ability
hits þ correct rejections). subscales are positively defined, with lower scores indicating
In addition to the standard protocol, we added a long-term less memory contentment and reflecting more daily forget-
delayed recall and recognition test. In the long-term delayed fulness (Troyer & Rich, 2000; Van der Werf & Vos, 2011).
recognition condition, the original 15 distracters were
replaced with 15 alternative distracters that were chosen from 2.3. Data analysis
a parallel RAVLT version (with matching word frequencies for
the distractor items). After one week, all patients were asked First, we examined whether the groups differed on the RAVLT
to come back to discuss the results; during that visit the standard outcome measures (i.e., total number of words
additional recall and recognition tests were administered. The recalled during the 5 learning trials, the 30-min delayed recall
Please cite this article in press as: Geurts, S., et al., Accelerated long-term forgetting after TIA or minor stroke: A more sensitive measure
for detecting subtle memory dysfunction? Cortex (2018), https://doi.org/10.1016/j.cortex.2018.04.002
4 c o r t e x x x x ( 2 0 1 8 ) 1 e7
performance and the 30-min delayed recognition) using Stu- analysis on the free-recall scores, the sphericity assumption
dent's t-tests. Subsequently, we performed a mixed-design was violated [W ¼ .658, c2 (2) ¼ 23.0, p < .0005], so Greenhouse-
GLM analysis using the recall scores of the RAVLT as within- Geisser corrections were applied. The results showed a sig-
subject factor (three levels: trial 5, 30-min and 1-week recall) nificant main effect of Time [F (1.49, 83.5) ¼ 142.1, p < .0005,
and group (patients vs controls) as between-subject factor. h2p ¼ .717], a Time Group interaction [F (1.49, 83.46) ¼ 5.3,
Next, we performed a mixed-design GLM analysis on the p ¼ .013, h2p ¼ .086], but no main effect of Group [F (1,56) ¼ 1.4,
recognition data (two levels: 30-min and 1-week recognition), p ¼ .291, h2p ¼ .020]. Follow-up t-tests showed that while the
again with group (patient vs controls) as between-subject free recall performance did not differ between the two groups
factor. The assumption of normality of the data was exam- after 30 min (see above, d ¼ .04), the patients showed
ined using the skewness and kurtosis statistics for all vari- a worse free-recall compared to the controls after 1 week
ables and Mauchly's W was computed to examine the [t (56) ¼ 2.318, p < .024; d ¼ .61]. With respect to the recog-
sphericity assumption. nition performance (see Table 2), the GLM analysis showed a
Using the two delayed recall scores (i.e., after 30 min and significant main effect of Time [F (1, 56) ¼ 63.9, p < .0005,
one week) we calculated the percentage of information loss h2p ¼ .533], but neither a main effect of Group [F (1, 56) ¼ 1.0,
(Recall Decline) as follows: [100 (standard 30-min recall p ¼ .334, h2p ¼ .017] nor a Time Group interaction
score e one-week recall score)/standard 30-min recall score]. [F (1, 56) ¼ .3, p ¼ .606, h2p ¼ .005].
The same formula was used with the delayed recognition For the decline scores (see Table 2), a significant group
scores for the percentages of information loss on the recog- difference was found on Recall Decline [t (56) ¼ 3.286, p < .002]
nition condition (Recognition Decline). The Recall and Recog- but not on Recognition Decline [t (56) ¼ .614, p ¼ .542], indi-
nition Decline scores as well as the MMQ scores were cating that the percentage of information loss after one week
compared across the groups using Student's t-tests. Effect was significantly higher for the patients than the controls on
sizes were computed (h2p for the GLM factors and Cohen's d for the recall measures but not on the recognition scores.
t-tests). Pearson correlation coefficients (r) were computed for With respect to the MMQ (see Table 2), three patients did
the patients and controls separately to examine the associa- not fully complete the questionnaire. MMQ analyses were
tions between the Recall and Recognition Decline measures, performed with the remaining 27 patients and all controls. No
the MMQ subscales, and the time post-event. significant differences were found between patients and
controls on MMQ Ability [t (53) ¼ .974, p ¼ .34] or MMQ
Contentment [t (56) ¼ 1.221, p ¼ .23]. A significant positive
3. Results correlation was found in the patient group between the MMQ
Ability score and Recognition Decline (r ¼ .39, p ¼ .04). Higher
Skewness and kurtosis statistics showed all variables were subjective memory ability scores were associated with more
normally distributed. The two groups did not differ signifi- long-term recognition decline. A significant negative correla-
cantly with respect to the total score on the five learning trails tion was found in the control group between the MMQ
[Mcontrol ¼ 47.9, SD ¼ 7.4; Mpatients ¼ 47.2, SD ¼ 9.5; t (56) ¼ .31, Contentment score and Recall Decline (r ¼ .39, p ¼ .04). In
p ¼ .76], the 30-min delayed recall score [t (56) ¼ .165, p ¼ .87, this group higher memory contentment scores were associ-
see Fig. 1] and the 30-min delayed recognition score ated with less decline on long-term recall. Time after stroke or
[t (56) ¼ .953, p ¼ .35, see Table 2] of the RAVLT. TIA was not significantly correlated with Recall Decline
Fig. 1 shows the RAVLT free recall scores immediately after (r ¼ .317) or Recognition Decline (r ¼ .247).
trial 5, the 30-min delay and after 1-week. For the GLM
4. Discussion
Please cite this article in press as: Geurts, S., et al., Accelerated long-term forgetting after TIA or minor stroke: A more sensitive measure
for detecting subtle memory dysfunction? Cortex (2018), https://doi.org/10.1016/j.cortex.2018.04.002
c o r t e x x x x ( 2 0 1 8 ) 1 e7 5
Table 2 e Results for RAVLT delayed recognition and long-term decline as well as MMQ scores.
RAVLT ¼ Rey Auditory Verbal Learning Test (Dutch version); MMQ ¼ Multifactorial Memory Questionnaire; *p < .05.
distinct processes, that is, familiarity (a sense of having seen There are some limitations that should be addressed. First,
an item before) and recollection (remembering that a partic- we analyzed the minor stroke and TIA patients together, as the
ular item have been previously seen, as well as the context in sample size was too small to perform subgroup analyses.
which it appeared) (Merkowa, Burke, & Kahanac, 2015). These Future studies should include a larger group so that TIA and
processes combined may also have facilitated the recognition minor stroke patients can be analyzed separately, to investigate
performance after 1 week, resulting in a performance similar to what extent each diagnosis contributes to long-term forget-
to controls in the patients. Alternatively, this discrepancy may ting. Also, we only included patients below 75 years of age, as a
also be the result of a retrieval deficits due to subtle executive result of which the ALF cannot be generalized to minor stroke at
impairments (Unsworth, 2015; Davidson, Troyer, & older age. Future studies should also investigate the clinical
Moscovitch, 2005). significance of ALF in minor stroke and TIA, for example using a
Furthermore, we investigated the association between longitudinal approach, to identify those who are at risk for
subjective memory complaints and the actual memory per- further cognitive decline. Furthermore it would be interesting to
formance. The results demonstrated comparable levels of relate ALF in patients with TIA or minor stroke to neuroimaging
memory contentment and complaints of forgetfulness in the findings, for instance diffusion-weighted imaging (DWI), as
patient group compared to the controls. In addition, the as- DWI abnormalities have been found in minor stroke and TIA
sociations between subjective memory functioning and the patients (Brazzelli et al., 2014; Kate et al., 2015; Schulz, Briley,
memory performances were in general low. However, the two Meagher, Molyneux, & Rothwell, 2003; Van Rooij et al., 2017).
significant correlations are interesting. That is, patients who In sum, this study is the first to investigate long-term
reported less memory complaints demonstrated more forget- forgetting after cerebrovascular disease, showing that accel-
ting over time on the recognition task. In contrast, healthy erated forgetting may be a more sensitive measure in detecting
participants who were more content with their memory per- memory difficulties after minor stroke and TIA than traditional
formance showed less recall decline over time. This suggests memory assessment. Based on our results, the delay interval in
that healthy controls may be better at rating their own memory testing in clinical practice (typically between 15 and
memory functioning in comparison to the patients. Only a few 30 min) may be too short. This is also in agreement with
previous studies have addressed subjective cognitive func- neurobiological findings on episodic memory formation. That
tioning in minor stroke and TIA. Two studies reported the is, there is evidence that consolidation takes place over days
presence of memory complaints after minor stroke or TIA and weeks rather than minutes, possibly mediated by sleep
(Dassonville et al., 2016; Fens et al., 2013), while one study (Chambers, 2017; Dudai, Karni, & Born, 2015). For instance, ALF
reported that TIA patients did not experience cognitive com- for context location was found in thalamic stroke patients after
plaints (Van Rooij et al., 2014). In (major) stroke, subjective 24 h, despite normal context memory performances after one
cognitive functioning has been studied more extensively, but hour (Tu, Miller, Piguet, & Hornberger, 2014).
conflicting results on the association between subjective Future neuroimaging studies combined with more exten-
memory complaints and performances on memory tests have sive cognitive testing (also including measures of executive
been found (e.g., Van Rijsbergen, Mark, de Kort, & Sitskoorn, function and mental effort) may provide further insights into
2014). A lack of insight or emotional distress are possible the underlying mechanisms of ALF in TIA and minor stroke
causes. An alternative explanation for this discrepancy is that patients.
the subtle memory decline after one week may not have
resulted in noticeable daily problems in our sample, especially
given the relatively short time since the vascular event for Acknowledgements
most of the patients. Both assumptions could contribute to the
lack of concordance in this study and even explain why pa- The authors would like to thank the BAC team consisting of
tients who reported less complaints showed more long-term I. Bekker, J. Kramer, and B. Notz for their assistance in the data
forgetting on recognition. collection.
Please cite this article in press as: Geurts, S., et al., Accelerated long-term forgetting after TIA or minor stroke: A more sensitive measure
for detecting subtle memory dysfunction? Cortex (2018), https://doi.org/10.1016/j.cortex.2018.04.002
6 c o r t e x x x x ( 2 0 1 8 ) 1 e7
Please cite this article in press as: Geurts, S., et al., Accelerated long-term forgetting after TIA or minor stroke: A more sensitive measure
for detecting subtle memory dysfunction? Cortex (2018), https://doi.org/10.1016/j.cortex.2018.04.002
c o r t e x x x x ( 2 0 1 8 ) 1 e7 7
patients referred for neuropsychological assessment. Frontiers Zamboni, G., Griffanti, L., Jenkinson, M., Mazzucco, S., Li, L.,
in Psychology, 7, 605. https://doi.org/10.3389/fpsyg.2016.00605. Küker, W., et al., Oxford Vascular Study. (2017). White matter
Van der Werf, S. P., & Vos, S. H. (2011). Memory worries and self- imaging correlates of early cognitive impairment detected by
reported daily forgetfulness: A psychometric evaluation of the the Montreal Cognitive Assessment after transient ischemic
Dutch translation of the multifactorial memory questionnaire. attack and minor stroke. Stroke, 48, 1539e1547. https://doi.org/
Clinical Neuropsychologist, 25, 244e268. https://doi.org/10.1080/ 10.1161/STROKEAHA.116.016044.
13854046.2010.543290.
Please cite this article in press as: Geurts, S., et al., Accelerated long-term forgetting after TIA or minor stroke: A more sensitive measure
for detecting subtle memory dysfunction? Cortex (2018), https://doi.org/10.1016/j.cortex.2018.04.002