1
Evaluating the Trainability of Stroke Patients in
Comparison with Healthy Elderly in a Cognitive
Flexibility Training
Iuno Groot
University of Amsterdam May 2015
Student number: 6221831/10011463
Specialization: Research Master - Clinical Neuropsychology Supervisor: Jaap Murre, Renate van de Ven
2 Index
Abstract P. 3
Evaluating the Trainability of Stroke Patients in Comparison with Healthy Elderly in a CognitiveFlexibility Training
- Introduction P. 4
- Trainability of stroke patients P. 6
-Hypothesis P. 8 Methods -Sample characteristics P. 9 -Materials P. 10 -Procedure P. 13 -Statistics P. 14 Results P. 15 Discussion P. 21 List of References P. 26
3 Abstract
In this study we test the effectivity of a new online cognitive flexibility training for stroke patients and healthy elderly. First we tested whether this cognitive training would improve general cognitive functioning in stroke patients. Next we compared the trainability of 31 stroke patients with 35 healthy elderly. We also tested the hypothesis whether training is more effective when the training is started soon after stroke-onset. Results show that stroke patients in the switch training did not improve more than patients who followed the mock-training. However, both groups did improve on general cognitive functioning, showing far transfer effects of non-trained cognitive domains. Furthermore, there was no difference in trainability between stroke patients and healthy elderly. And finally, no relationship was found between time since stroke and effectivity of the training. Implications and advice for further research are discussed.
4 Evaluating the Trainability of Stroke Patients in Comparison with Healthy Elderly in a
Cognitive Flexibility Training
When the blood supply to the brain is interrupted, neurons in need of oxygen and energy cannot function anymore and incur structural damage in minutes. This neurological condition is named stroke. In the Netherlands, approximately 13.000 people per year do not survive a stroke (World Health Organization [WHO], 2003). The people who do survive often end up having cognitive and motoric disabilities. Even after a year and after following rehabilitation programs, almost 60 percent of stroke-victims maintain cognitive impairments (Rasquin et al. 2004). These cognitive impairments have an enormous impact on patient’s ability to perform activities in daily life (Claesson, Lindén, Skoog & Blomstrand, 2005) and can have a negative influence on stroke recovery and quality of life (Nys et al., 2005; del Ser et al., 2005). A serious need for the optimization of rehabilitation techniques, therefore, exists.
Over the years extensive research has been conducted to investigate the many existing interventions. To date there are promising, but inconsistent results (Kwakkel, Kollen & Wagenaar, 1999; Langhorne, Bernhardt & Kwakkel 2011). One promising result comes from Westerberg et al. (2007) who conducted a computerized working memory training for stroke patients. In this study, the nine participants improved significantly on attention and working memory tasks after five weeks of training. Another study using working memory training showed significant improvements for stroke patients on memory tasks after a frequent and intensive memory training in comparison to the control group (Hildebrandt, Bussmann-Mork & Schwendemann, 2006). However, Langhorn et al. (2011) conclude in their review that the effects of most cognitive training programs are not consistent enough and are sometimes only effective for small groups of patients. This is partly due to the high diversity in stroke
patients. Because strokes can occur anywhere in the brain, stroke-patients all have their own cognitive profile of impairments. Often these include deficits in attention, language, executive
5
functioning, memory or a combination thereof (Nys et al. 2005). The urge to find an effective training for all stroke patients with their variety of cognitive impairments is high.
Recently several studies have highlighted the importance of using switch-paradigms in cognitive training in healthy elderly (Buitenweg, Murre & Ridderinkhof, 2012; Karbach & Kray, 2009). In a typical switch-paradigm, participants have to switch their attention between different tasks or subtasks. Switching relies on a constantly active working memory, where the different rules or properties are held active (Buchler, Hoyer & Cerella, 2008). Our
working memory seems to be the key to success in cognitive training according to the review of Klingberg (2010). This review underlines the high plasticity and trainability of our working memory, contrary to the old view (Miller, 1956) that described the working memory as
having a static load that could only contain seven items.
Furthermore Klingberg (2010) describes promising results that training our working memory through switch paradigms can have on other cognitive domains. This is promising because one of the most important features of an effective training is that there is not only improvement of the trained function but that there is also transfer to other cognitive functions (i.e. ‘far transfer’). Karbach and Kray (2009) studied whether switching training could
improve other cognitive domains rather than exclusively switching-ability. Results showed that healthy elderly improved after the training not only on switching itself, but also on spatial and verbal working memory, fluid intelligence, and they showed less interference effects on the Stroop task. While switch-training is not yet tested in stroke patients, it seems especially promising for this group because they experience widely varying cognitive impairments. The first aim for this study will therefore be to test the effectiveness of a cognitive training including a switch-element for stroke patients.
6 Trainability of Stroke Patients
Most of the previous studies in cognitive training are based on healthy elderly. Due to the lack of research in the stroke population, it is not clear yet if these promising results would be comparable for stroke patients. It would be of great value for rehabilitation programs to understand the patterns of similarity between stroke recovery and brain training of healthy elderly to enhance effectivity (Cramer & Chopp, 2000). To our knowledge, there is no study yet that compares the (re)learning behavior of stroke patients with healthy elderly.
Learning or training of (cognitive) functions appeals to a variety of brain processes. These processes are based on neural plasticity, which refers to the brains ability to create new neural pathways and alter the structure of existing ones as a result of new experiences (Kleim & Jones, 2008). In other words, neural plasticity is the way that the brain encodes new learned behaviors. Cognitive training with stroke patients applies to the same neural processes as learning in a healthy brain does, however the post-stroke brain is changed in the way it
responds to learning (Kleim & Jones, 2008). (Re)Learning in the post-stroke brain is therefore more complex and somewhat contradicting.
On the one hand, the post-stroke brain has to deal with a significant loss of neurons and neuronal connections (Hasbani, Underhill, de Erausquin & Goldberg, 2000). Without these neurons and their connections it is not surprising that the (re)learn ability is dramatically altered (Kleim & Jones, 2008). The ‘learning organ’ is damaged and therefore stroke patients can have a hard time to learn or train cognitive functions (Ponds, van Heugten, Fasotti, & Wekking, 2010). Even though stroke patients have a higher need for cognitive training than healthy elderly, because they experience more cognitive dysfunction (Rasquin et al. 2004), based on the above we might assume that a cognitive training would be less effective for stroke patients than for healthy elderly.
7
On the other hand, the damaged brain seems to have an extraordinary ability to use and expand the role of neural plasticity in comparison to a healthy brain. Through neural plasticity, the post-stroke brain is able to fix a part of the damaged area and regain lost function. This self-sufficient process called ‘spontaneous recovery’ works through roughly two processes. First, there exists a remarkable increase of connectivity existing in the central nervous system. And second, through this boost of connectivity, there is formation of new functional and structural neuronal circuits through the remapping of related regions (Murphy & Corbett, 2009).
From recent animal studies it becomes clear that the post-stroke brain may have an advantage in plasticity over healthy brains through induced growth-associated genes (Carmichael, Archibeque, Luke Nolan, Momiy & Li, 2005). These growth genes are important in axonal sprouting in peripheral nervous system regeneration and cortical development. Cramer & Chopp (2000) argue that the pattern of these growth genes is comparable with the plasticity of the brain earlier in our development. Only this time the period of high plasticity is of short duration; most research indicates till three months after stroke onset (Kleim & Jones, 2008; Murphy & Corbett, 2009). This heightened plasticity can make it easier for the brain to (re)learn certain capacities (Kleim & Jones, 2008).
There is indeed evidence from behavioral studies that this limited period seems to be the most effective period for (re)training lost functions. Horn et al. (2005) found that three
different therapies (psychical and cognitive) where far more effective if the training was given closer to stroke onset. Another study also found a relationship between the time of admission in a rehabilitation center and recovery (Salter et al. 2006).
Even though sooner is probably better, Desmond et al. (2002) found that among patients with cognitive impairment 35% showed improved on a battery of neuropsychological tests until at least a year post-stroke. A study evaluating 154 stroke patients also reported that most
8
improvement in cognitive functions were seen from onset until three months post-stroke, but 24 % kept improving in intelligence and memory at least till one year after stroke onset (Kotila et al., 1984). Based on these results the trainability of stroke patients might even be better than the trainability in healthy elderly if the training is given within a limited time post-stroke.
Hypotheses
Preceding research revealed the urge for two important issues to be addressed: first, the fact that most stroke-victims keep having cognitive complaints after rehabilitation programs and as a result: the subsequent need to develop a training that will improve the lasting cognitive impairments that stroke patients possess. A newly developed computer-based switch-training seems promising. This study will examine the effect of switch-training on cognitive
functioning in stroke patients. Based on previous research we expect that switch-training will improve performances of stroke patients in attention, executive functioning, memory, and (fluid) intelligence.
Second, there is many promising literature on the cognitive training of healthy elderly. However, it is not known whether stroke patients are comparable in their trainability. It would be of great value for rehabilitation programs to understand the patterns of similarity between stroke recovery and normal development. We will therefore test and compare the trainability in stroke patients with healthy elderly. Based on previous literature we expect that in general the cognitive training will be more effective for healthy elderly than for stroke patients due to the loss of necessary neuronal connections after stroke. Because stroke patients’ brains have possible more growth promoting factors in the earlier post-stroke phase; we expect that stroke patients who start the training within three months post-stroke will be more trainable than stroke patients who start the training in a later post-stroke phase. Summarized, we expect to
9
see more progress in general cognitive functioning after training for healthy elderly than for stroke patients, unless the patients started the training within three months post-stroke. We also expect to find a negative relationship between improvement in general cognitive functioning after the training in comparison to baseline and time since stroke onset.
Methods
Sample characteristics
For this study 31 stroke patients and 35 healthy elderly were recruited. Inclusion criteria for both stroke patients and healthy elderly consisted of: basic computer skills and a computer available at home for training. Exclusion criteria for both groups were: decreased cognitive functioning (TICS-score lower than 27) or severe hearing or vision problems.
Stroke patients were recruited in three rehabilitation centers in the Netherlands; Reade, Heliomare and De Trappenberg. Stroke patients were selected if they had experienced a stroke within the past five years. All participants should have received cognitive rehabilitation but still maintain cognitive complaints. Additional exclusion criteria for stroke patients were: severe aphasia or neglect or sufferings from a severe psychiatric of physical diseases like epilepsy, psychoses, substance abuse or severe depression.
For the healthy elderly group, 35 participants were recruited with an age between 60 and 85 years. This group was recruited through advertisements in local newspapers. Additional exclusion criteria for elderly were: sufferings from a neurodegenerative disease like Parkinson or Alzheimer dementia or experience of a psychiatric disease like severe depression or anxiety disorder.
Personal data of the participants was anonymized. After completing the study, participants received a lifelong access to the brain training games and travel allowances for visiting the
10
University of Amsterdam (UVA). This study was approved by the medical ethical commission (METC) of the VU medical center.
Materials
Telephone Interview Cognitive Status (TICS; Kempen, Meier, Brouwens, van Deursen, & Verhey, 2007)
The TICS is a telephonic cognitive screening, consisting of 25 questions. The total score range is from 0 (severe cognitive constraints) till 41 (no cognitive constraints). The
correlation for test-retest reliability is high, r = .93 and .92. The Pearson correlations between the TICS and the cognitive screening measure MMSE is high; r = .77 (Kempen, Meier, Brouwens, van Deursen & Verhey, 2007).
Assessment of Neuropsychological Functioning
A subset of five neuropsychological tests was taken together to measure general cognitive functioning. These tests were selected to all represent a different cognitive function: memory, working memory, attention, logical reasoning and executive functioning (word fluency). A composite score (as suggested by Rasmussen et al. 2001) was calculated based on the following five neuropsychological tests to reduce the number of dependent variables. We have looked exploratively at the different tests separately.
Rey’s auditory verbal learning test (RAVLT; Sean & Deelman, 1986)
The RAVLT is an often used verbal memory test where participants have to remember a list of 15 words. A total score of correct remembered words after five trials is conducted and used for analyses. The test-retest reliability for the different items are high, between r = .60
11
and r = .70 (Geffen, Butterworth & Geffen, 1994). RAVLT is proved to predict memory impairment with a hit-rate of 75.3 % (p < .001) (Ryan & Geisser, 1986).
Letter-Number Sequencing (WAIS-IV)
This subtest of the WAIS-IV measures working memory capacity. In this task participants have to repeat and order sequences of numbers and letters. A total score of correct sequences will be conducted and used for the analyses. The test-retest reliability for the subtests of the WAIS-IV range from r = .70 till r = .90 (Wechsler, 2008). The WAIS correlated highly with the Stanford-Binet IQ test with r = .88 (Wechsler, 2008).
Trail Making Test (D-Kefs TMT; Delis, Kaplan & Kramer, 2007)
The TMT measures attention and cognitive flexibility. In this test participants have to draw a line between dots. The dots have numbers or letters and participants have to draw a line as fast as possible while alternating between numbers and letters. For the analyses we used a scaled time-score corrected for motion speed (version B corrected for A). The test-retest reliability of the TMT is normal to high, r = .56 and r = .77 (Bouma, Mulder, Lindeboom & Schmand, 2012). Scores on the TMT correlated with different executive functioning tests in a range of r =.14 to .73 (Sanchez-Cubillo et al., 2009).
Shipley (Shipley institute of Living Scale)
The Shipley test is a measure for logical reasoning. In this study only the verbal reasoning section was used, where participants have to identify logical sequences and complete these sequences. The Shipley consists of 20 items. Participants got one point per correct answer, resulting in a minimum score of 0 and a maximum score of 20. For the analyses the total score of correct answers was used. The vocabulary scale has a high reliability estimate ranging from
12
r = .85 to r = .92 (Kaya & Delen, 2009). The test is correlated to different measures of
intelligence, like the verbal GAMA IQ score with a correlation of r = .29 (Matthews, Lassiter & Habedank, 2001).
Category Fluency (D-Kefs TMT; Delis, Kaplan & Kramer, 2007)
A variant of the verbal fluency task is used to measure executive functioning including the ability to switch. Participants have to name items from two categories alternately. The total score of items that are named is used for further analyses. The test-retest reliability of the D-Kefs fluency task is high, with a Pearson correlation between r = .70 and r = .88 (Spreen & Straus, 1998, as cited in Wecker, Kramer, Hallam & Delis, 2004). There is a moderate correlation between the switching condition from de D-Kefs verbal fluency and another switch-paradigm, the TMT, r = .38 (Pearson correlation; Wecker et al. 2004).
Training
Participants trained for twelve weeks, five days a week for 30 minutes a day, online on their computer at home.
Switch Training. The switch-training consisted of nine different tasks in three cognitive
domains: attention, reasoning, and working memory. To provide the switch element, participants switched between the tasks every three minutes, resulting in ten different tasks per day. Task difficulty was adjusted to the participant’s level. In total, there were 20 levels with different degrees of difficulty. Participants were provided with feedback on their performance after each completed trainings sessions.
Mock Training. In the active control training, participants trained on four different tasks
that are not likely to improve cognitive functioning. There was no switch element included; participants trained ten minutes per task, resulting in three tasks per day. Participants trained
13
on a predetermined level for the whole week. In total there were nine different levels that varied in difficulty. Participants in the mock-training group were provided with feedback after each training session, as in the switch-training group.
Procedure
A controlled intervention study with an active control group was conducted. This study was part of the TAPASS (Training project Amsterdam Seniors and Stroke) study, only part of the collected data is used here.
All participants for this study were selected by an expert at the different rehabilitation centers. When registered with the project, all patients were screened following the telephonic interview cognitive status (TICS). For the remaining in- and exclusion criteria, participants filled in a questionnaire about their medical status. If participants met the criteria, they were randomized over the intervention- (switch-training) and the active control-condition (mock-training). Both groups followed a twelve week training program. For five days a week, 30 minutes per day participants trained online on their computer at home on different brain tasks. Participants were monitored daily on their training schedule and called weekly to discuss progress and to give them feedback on their performances.
Within this study there were two measurements were participants visited the UVA: a baseline measure, before the training began, and a post-treatment measure, when the training was finished. Here the patients received a neuropsychological assessment, including the RAVLT, letter and number sequencing, TMT, category fluency, and Shipley test. The duration of the baseline measurement was four hours, including a detailed instruction of the training. Duration of the post-treatment measurement was approximately three hours. All researchers who were present at the neuropsychological assessment were unaware of the training condition of participants.
14
Figure 1; research design
Statistics
All analyses were conducted using SPSS.22. A Chi-square test was used to test whether men and women were equally divided over the two conditions. A Shapiro-Wilk test was conducted to test if the data was normally distributed. Outliers were detected using a boxplot of the data.
From the neuropsychological assessment, we created a composit score for general
cognitive functioning. Raw scores on the RAVLT, letter-number sequencing, TMT, category fluency and the Shipley were translated to z-scores. Mean scores and standard deviations of the baseline measure were taken and also applied for the post-treatment measure. The z-scores from the five different neuropsychological tests were added to form one score of general cognitive functioning.
For the first hypothesis we have examined whether the switch- training improves cognitive functioning for stroke patients. Differences in z-scores between the switch- and mock-training on the neuropsychological assessment between baseline and post-treatment were analysed with a repeated measure ANOVA. We expected an interaction effect between group and time; scores on the neuropsychological assessment will be higher post-treatment in comparison to baseline, but only for the switch-training group. For the active control group we expected no significant changes in scores on the assessment between baseline and post-treatment.
TICS Screening Measurement 1: Neuropsychological Assessment 1. Switch - Training --- 2. Mock- Training Measurement 2: Neuropsychological Assessment
15
In an explorative analysis, we examined if the training had more effect on a specific test from the neuropsychological assessment. This will be conducted by examining possible improvement of the scores on the different test independently through a repeated measure MANOVA.
The second hypothesis tests whether there is a difference in improvement between stroke patients and healthy elderly on general cognitive functioning post-treatment in comparison with baseline. Differences in scores between the group of healthy elderly and stroke patients on the neuropsychological assessment between baseline and post-treatment were analysed with a repeated measure ANOVA. We expected an interaction effect between group and time; scores on the neuropsychological assessment will be higher post-treatment in comparison to baseline, but mainly for the healthy elderly group.
With the third hypothesis we tested whether there is a relationship between time since stroke and training effectivity. We used a correlation analyses to examine whether there is a (negative) relation between the time since stroke and the difference score from baseline to post-treatment of neuropsychological functioning. Because the assumptions for the Pearson correlation were not met, we used the non-parametric equivalent the Spearman correlation.
Results
In total 31 stroke patients completed the training. Three of them had missing data on at least one of the test in the neuropsychological assessment, resulting in 28 participants, of which 16 men and 12 women, with an average age of M = 60.25 (SD = 1.53). Men and women were equally divided over the two conditions, χ(1) = 2.054, p = .152.
In the healthy elderly group 35 participants completed the training, one of them had
missing data on the on at least one of the test in the neuropsychological assessment. Resulting in a group of 34 participants, with respectively 10 men and 24 women, with an average age of
16
M = 66.15 (SD = .744). Men and women were equally divided over the two condition; χ(1) = .008, p = .928. Demographic and clinical characteristics of the participants per condition are shown in table 1.
Table 1; Means and Standard Deviations (SD) of Demographic and Clinical Characteristics
of the Participants; Age (in years), Gender, Educational Level and Time since Stroke (in days).
Switch Training Mock Training Total
Stroke Patients
Age 58.92 (10.02) 61.25 (6.5) 60.25 (8.13)
Gender (% Female) 58.33 % 68.75 % 64.28%
Educational Level a 5.83 (0.39) 5.88 (0.88) 5.86 (0.75) Time since Stroke 1299.42 (352.23) 749.19 (416.12) 985 (472.90)
Healthy Elderly
Age 66.2 (4.78) 66.07 (3.79) 66.15 (4.35)
Gender (% Female) 30 % 71.43 % 47.05%
Educational Level a 5.65 (1.3) 5.79 (0.89) 5.71 (1.14)
aEducation is coded comparable to Verhage (1964); ranging from 1 (elementary education not completed) to 7
(university degree).
Within the group of stroke patients, two outliers were found (outside two standard deviations from the mean). In the group for healthy elderly one outlier was detected. All analyses are conducted with and without the outliers. None of the results differed between the group with and without the outliers, so the data with outliers is presented.
17 Hypothesis 1
To test the first hypothesis, whether stroke patients profit from the switch-training in comparison to the mock-training, we created a composit score for general cognitive functioning by adding Z-scores of the five neuropsychological tests. Mean and standard deviations for this measure of cognitive functioning for the two conditions of stroke patients are shown in Table 2.
Table 2. Mean scores, Standard Deviation (SD) and Number of Participants (N) of Cognitive
Functioning for Stroke Patients in the Switch Training Group and the Mock Training Group at Baseline and Post-Treatment.
Baseline Post-Treatment Switch-training (N=12) -.359 (.582) .497 (.572)
Mock-training (N=16) -.012 (.811) .553 (.742)
A repeated measure ANOVA is conducted to test whether the mean scores of cognitive functioning differ post-treatment in comparison to baseline for the switch- and mock-training. A Shapiro-Wilk test showed that both training groups were distributed normally (p > .05). However, the assumption of sphericity was violated, we have therefore taken the values of the Greenhouse-Geisser test.
Against expectation; there was no significant interaction effect between time and training; cognitive functioning scores were not higher post-treatment in comparison to baseline for participants of the switch-training in comparison to participants of the mock-training; F(1,26) = .199, p = .659. However, we can observe a main effect for time. We can see that both groups improved post-treatment in comparison to baseline, F(1,26) = 4.71, p = .039. The
18 Figure 2; Means and standard errors of general cognitive functioning for the switch- and
mock-training on baseline and post-treatment.
Exploratively we tested if the training had more effect on a specific cognitive domain for stroke patients. Mean and standard deviation for RAVLT, letter-number sequencing, TMT, category fluency and the Shipley for stroke patients in the two training groups are shown in table 3. -0,6 -0,4 -0,2 0 0,2 0,4 0,6 0,8 1 Measurement Ge n e ral C o gn itiv e Fu n ction in g switch-training mock-training
19
Table 3. Means and Standard Deviation for stroke patients on the RAVLT, Letter-Number
Sequencing, TMT, Category Fluency and Shipley in the Switch- and Mock-training condition at Baseline and Post-Treatment.
A repeated measure MANOVA was conducted to test whether mean scores on the RAVLT, NDS, TMT, Category Fluency, and the Shipley differed post-treatment in
comparison to baseline for the switch- and mock-training. The assumption of sphericity was violated so we have taken the values of the Greenhouse-Geisser test. No significant
interaction effect between time and training was found; mean scores on the individual tests were not higher post-treatment compared to baseline for the switch- or mock-training. There was one main-effect found, for the letter-number sequencing test. Here there was a significant improvement over time; both groups improved on letter and number sequencing
post-treatment in comparison to baseline F(1,6.159) = 10.620, p < .05.
Hypothesis 2
In the second hypothesis we tested whether the training was more effective for stroke patients than for healthy elderly. Because there was no difference between the switch- and mock-training group for the stroke patients (See the first hypothesis), we merged these two groups to one group. The group of healthy elderly also did not differ between the two trainings;
Test Switch-Training (n= 12) Mock-training (n = 16)
Baseline Post-Treatment Baseline Post-Treatment
RAVLT -.053 (.955) .001 (.999) .156 ( 1.074) .017 (.683)
letter-number sequencing .061 (1.159) .524 (1.389) -.001 (1.07) .0453 (1.158)
TMT -.161 (1.07) -0.148 (1.411) -.421 (.961) .203 (.997)
Category Fluency -.049 (.815) -.085 (.929) .229 (1.158) .024 (1.004)
20 F(1,32) = .253 , p = .619. We therefore also merged these groups into one. Mean and standard
deviations for this measure of cognitive functioning for the merged groups of stroke patients and healthy elderly are shown in Table 4.
Table 4. Mean scores, Standard Deviation (SD) and Number of Participants (N) on Cognitive
Functioning for stroke patients and healthy elderly at Baseline and Post-Treatment.
Baseline Post-Treatment
Mean SD Mean SD
Stroke Patients (N=28 ) -.161 .519 .529 .482
Healthy Elderly (N=34) -.106 .522 .688 .599
Next a repeated measure analysis was conducted. The assumption of sphericity had been violated; we have therefore taken the values of the Greenhouse-Geisser test. There was no significant interaction effect between time and group; cognitive functioning scores did not improve more for the stroke patients in comparison to the healthy elderly; F(1,60) = . 041, p = .840. However, there was a significant main effect for time; both the stroke patients and healthy elderly have significantly improved post-treatment in comparison with baseline;
F(1,60) = 8.319 , p < .01.
Note that baseline levels of general cognitive functioning do not differ between stroke patients and healthy elderly; an independent t-test shows no difference t (60) = 0.73, p = .942.
Hypotheses 3
Within the third hypothesis we tested whether there is a relationship between the effectivity of the training and time since stroke. The average time since stroke was approximately M = 985 days (SD = 472.90). A correlation analysis was conducted between the difference score from
21
baseline to post-treatment of neuropsychological functioning, and the time since stroke. Because the variables did not have a linear relationship, we have chosen a non-parametric equivalent to Pearson’s correlation, the Spearman test. Against expectation, there was no (negative) relationship between time since stroke and effectivity of the training; r = .112, p = .593. The relationship between the deviation score (post-treatment minus baseline) for general cognitive functioning and time since stroke is shown in figure 3.
Figure 3; Scatterplot showing the relationship between the deviation score (post-treatment
minus baseline) of general cognitive functioning and time since stroke.
Discussion
In this study a newly developed cognitive flexibility training was tested for stroke patients and healthy elderly. The first aim of this study was to determine the effectivity of the switch-training for stroke patients in comparison to a mock-switch-training. Against expectation, there were no higher scores on the neuropsychological assessment for stroke patients who followed the
-4 -3 -2 -1 0 1 2 3 4 5 0 500 1000 1500 2000 2500 D e vi ation S co re
22
switch-training in comparison to stroke patients who followed the mock-training. Both groups improved on general cognitive functioning after training.
Next we studied whether there was a difference in trainability between healthy elderly and stroke patients. Against expectation, there was no difference in improvement for the healthy elderly in comparison to the stroke patients. Finally, we tested whether there was a negative relationship between the time that had passed since stroke and improvement in general cognitive functioning after the training. Again, against expectation, we did not find a (negative) relationship between time since stroke and effectivity of the training.
First of all, although we did not find a difference of improvement in general cognitive functioning between the switch- and mock-training, we did find that both groups improved after training. As argued in the introduction, one of the most important features of an effective training is that there is far transfer to other cognitive domains. It is a very promising result that the general cognitive functioning improved after following the training, because it shows improvement on other cognitive domains than the trained functions itself.
The absence of a difference in improvement between the switch- and mock-training can possible be explained by the fact that we had a small group of participants and therefore a lack of power. In the results we can observe a trend where participants in the switch-training improved more on general cognitive training than participants following the mock-training. Possible this trend would be clearer if there were more participants. For further research it would be of great value to test the training on a bigger population of stroke patients.
Another possible explanation can be that de mock-training did not function as a reliable control condition. Patients who followed the mock-training also trained everyday on different ‘games’, but without the switch element. Originally it was thought that these brain-games would not improve cognitive functioning. However, it is possible that they did. The results showed that both groups did improve on general cognitive functioning after following
23
the switch- and mock-training. Future research should focus on adding a waiting-list condition. In that way it can become clear whether the mock-training was too good for a control-training or that neither trainings really worked.
It is also possible that the switch element in the switch-training was not as effective as we expected. This may be explained by the fact that this study did not manipulate the switching effect in the same way as previous studies have done. As mentioned in the introduction, the effectivity of switching may rely on a constantly active working memory (Buchler, Hoyer & Cerella, 2008), which would improve working memory capacity and would transfer to other cognitive domains according to Klingberg (2010). In the study by Karbach and Kray (2009) this is exactly what they did: participants had to switch constantly (within tens of seconds) between two tasks. However, in our study, participants had to switch every three minutes between different brain-games. It is arguable that this switch-manipulation was not challenging enough to improve working memory.
Another important aspect to take into consideration for future research is to specify the training more per patient. As mentioned in the introduction, the cognitive complaints of stroke patients are very diverse. One patient might have more memory problems, where another patient has a deficit in attention. It is maybe too simplistic to think that a single training can improve all those different cognitive impairments by also improving cognitive flexibility. Previous research has shown promising results where switch-training did not only improve the ability to switch but also improved other cognitive functions (Karbach & Kray, 2009). However, specialized trainings also show promising results for stroke patients with attention deficits (Lincoln, Majid, & Weyman, 2000) and memory dysfunction (Kaschel, Sala,
Cantagallo, Fahlböck, Laaksonen, & Kazen (2002). Therefor a combination of switch-training specialized in the cognitive domain where patients have the most disturbances can be a
24
memory games, while constantly switching between these games. Further research should study this interesting possibility.
The second hypothesis about the difference in trainability between stroke patients and healthy elderly was also not confirmed. Against expectation, no difference was found in the way stroke patients and healthy elderly are trainable. This hypothesis was based on the fact that stroke patients can experience trouble in (re)learning because of the significant loss in neurons. However, both groups improved on general cognitive functioning after cognitive training, indicating that stroke patients are in general just as trainable in their cognitive functions as healthy elderly are. This can have positive implications in that previous research on cognitive training for healthy elderly is comparable for stroke patients.
However, we should be cautious with this conclusion. In our hypothesis we assumed that stroke patients would have a lower baseline score of general cognitive functioning through their brain damage than healthy elderly. As argued in the introduction, this was the reason why we expected that in general, healthy elderly would be more trainable than stroke patients. However, results showed that baseline scores of general cognitive functioning did not differ between stroke patients and healthy elderly. It is possible that our exclusion criteria for stroke patients were too strict, resulting in a well-functioning group of stroke patients. To answer the initial research question whether stroke patients are different in their trainability than healthy elderly, a more diverse group of stroke patient with more objective cognitive dysfunctions is needed.
Finally, we did not find a relationship between time since stroke and trainability. This hypothesis was based on the idea that there is a heightened plasticity of the brain after stroke. This possible learning advantage would be the strongest till three months post-stroke.
Preceding research indicated however, that this high plasticity might continue until a year post-stroke for about 24% of the patients (Kotila et al., 1984). In this study there were no
25
patients who experienced a stroke in the past three months. This can be an explanation why we did not obtain a relation between effectivity of the training and time since stroke. Possibly, our group of stroke patients was not diverse enough in time since stroke to show such
differences in trainability. For future research it would be recommended to recruit more patients and start the training closer to stroke onset.
This study tested a new cognitive training for the population of stroke patients and compared their trainability with healthy elderly. The new designed switch-training was not more effective than an active control condition. The new developed switch-training can and has to be improved in multiple features to become a more effective cognitive training for stroke patients. Next to that, a lot of questions are still to be solved around the trainability of stroke patients and the role that heightened brain-plasticity can play in rehabilitation
programs. Summarizing, this study addressed the high urge to find an effective cognitive training for stroke patients. Furthermore, it has given some clear direction for future research to expand the field of cognitive rehabilitation and hopefully develop a more effective training within the next years.
26 References
Buchler, N. G., Hoyer, W. J., & Cerella, J. (2008). Rules and more rules: The effects of multiple tasks, extensive training, and aging on task-switching performance. Memory &
Cognition, 36(4), 735-748.
Buitenweg, J. I. V., Murre, J. M. J., & Ridderinkhof, K. R. (2012). Brain training in progress: A review of trainability in healthy seniors. Frontiers in Human Neuroscience, 6, 1-11.
Bouma, A., Mulder, J., Lindeboom, J., & Schmand, B. (red) (2012). Handboek
neuropsychologische diagnostiek. Amsterdam: Pearson.
Carmichael, S. T., Archibeque, I., Luke, L., Nolan, T., Momiy, J., & Li, S. (2005). Growth-associated gene expression after stroke: Evidence for a growth-promoting region in peri-infarct cortex. Experimental Neurology, 193(2), 291-311.
Carpenter, P. A., Just, M. A., & Shell, P. (1990). What one intelligence test measures: A theoretical account of the processing in the Raven Progressive Matrices
Test. Psychological Review, 97(3), 404-431.
Claesson, L., Lindén, T., Skoog, I., & Blomstrand, C. (2005). Cognitive impairment after stroke: Impact on activities of daily living and costs of care for elderly people.
Cerebrovascular Diseases, 19(2), 102-109.
Cramer, S. C., & Chopp, M. (2000). Recovery recapitulates ontogeny. Trends in
Neurosciences, 23(6), 265-271.
Desmond, D. W., Moroney, J. T., Sano, M., & Stern, Y. (2002). Incidence of dementia after ischemic stroke results of a longitudinal study. Stroke, 33(9), 2254-2262.
27
Geffen, G. M., Butterworth, P., & Geffen, L. B. (1994). Test-retest reliability of a new form of the auditory verbal learning test (AVLT). Archives of Clinical Neuropsychology, 9(4), 303-316.
Hasbani, M., Underhill, S. M., de Erausquin, G., & Goldberg, M.P. (2000). Synapse loss and regeneration: A mechanism for functional decline and recovery after cerebral ischemia?
The Neuroscientist; 6(2), 110–119.
Hedden, T., & Gabrieli, J. D. (2004). Insights into the ageing mind: A view from cognitive neuroscience. Nature Reviews Neuroscience, 5(2), 87-96.
Hildebrandt, H., Bussmann-Mork, B., & Schwendemann, G. (2006). Group therapy for memory impaired patients: A partial remediation is possible. Journal of Neurology, 253, 512-519.
Horn, S. D., DeJong, G., Smout, R. J., Gassaway, J., James, R., & Conroy, B. (2005). Stroke rehabilitation patients, practice, and outcomes: Is earlier and more aggressive therapy better? Archives of Physical Medicine and Rehabilitation, 86(12), 101-114.
Karbach, J., & Kray, J. (2009). How useful is executive control training? Age differences in near and far transfer of task‐switching training. Developmental Science, 12(6), 978-990. Kaschel, R., Sala, S. D., Cantagallo, A., Fahlböck, A., Laaksonen, R., & Kazen, M. (2002).
Imagery mnemonics for the rehabilitation of memory: A randomised group controlled trial.
Neuropsychological Rehabilitation, 12(2), 127-153.
Kaya, F., & Delen, E. (2009). Test review: Shipley-2 manual. Journal of Psychoeducational
28
Kempen, G. I. J. M., Meier, A. J. L., Bouwens, S. F. M., van Deursen, J., & Verhey, F. R. J. (2007). Telefonisch Interview Cognitieve Status (TICS): Psychometrische
aspecten. Tijdschrift voor Gerontologie en Geriatrie, 38(1), 34-40.
Kleim, J. A., & Jones, T. A. (2008). Principles of experience-dependent neural plasticity: Implications for rehabilitation after brain damage. Journal of Speech, Language, and
Hearing Research, 51(1), 225-239.
Klingberg, T. (2010). Training and plasticity of working memory. Trends in Cognitive
Sciences, 14(7), 317-324.
Kotila, M., Waltimo, O., Niemi, M. L., Laaksonen, R. I. T. V. A., & Lempinen, M. (1984). The profile of recovery from stroke and factors influencing outcome. Stroke, 15(6), 1039-1044.
Kwakkel, G., Kollen, B. J., & Wagenaar, R. C. (1999). Therapy impact on functional recovery in stroke rehabilitation: A critical review of the literature. Physiotherapy, 85(7), 377-391.
Langhorne, P., Bernhardt, J., & Kwakkel, G. (2011). Stroke rehabilitation. The Lancet,
377(9778), 1693-1702.
Lincoln, N., Majid, M., & Weyman, N. (2000). Cognitive rehabilitation for attention deficits following stroke. The Cochrane Library , 4. Art. No.: CD002842. DOI:
10.1002/14651858.CD002842.
Matthews, T. D., Lassiter, K., & Habedank, H. (2001). Validity of two brief measures: The General Abilities Measure for Adults and the Shipley Institute of Living Scale. Perceptual
29
Miller, G. A. (1956). The magical number seven, plus or minus two: Some limits on our capacity for processing information. Psychological Review, 63, 81-97.
Murphy, T. H., & Corbett, D. (2009). Plasticity during stroke recovery: From synapse to behaviour. Nature Reviews Neuroscience, 10(12), 861-872.
Nys, G. M. S., Van Zandvoort, M. J. E., De Kort, P. L. M., Jansen, B. P. W., Van der Worp, H. B., Kappelle, L. J., & De Haan, E. H. F. (2005). Domain-specific cognitive recovery after first-ever stroke: A follow-up study of 111 cases. Journal of the International
Neuropsychological Society, 11(07), 795-806.
Ponds, R., van Heugten, C., Fasotti, L., & Wekking, E. (2010). Neuropsychologische
Behandeling. Amsterdam: Boom.
Rasmussen, L. S., Larsen, K., Houx, P., Skovgaard, L. T., Hanning, C. D., & Moller, J. T. (2001). The assessment of postoperative cognitive function. Acta Anaesthesiologica
Scandinavica, 45(3), 275-289.
Rasquin, S., Lodder, J., Ponds, R. W. H. M., Winkens, I., Jolles, J., & Verhey, F. R. (2004). Cognitive functioning after stroke: A one-year follow-up study. Dementia and Geriatric
Cognitive Disorders, 18(2), 138-144.
Robertson, I. H., & Murre, J. M. J. (1999). Rehabilitation of brain damage: Brain plasticity and principles of guided recovery. Psychological Bulletin, 125(5), 544-575.
Ryan, J. J., & Geisser, M. E. (1986). Validity and diagnostic accuracy of an alternate form of the Rey Auditory Verbal Learning Test. Archives of Clinical Neuropsychology, 1(3), 209-217.
30
Salter, K. B. A., Hartley, M. B. A., & Foley, N. B. (2006). Impact of early vs delayed admission to rehabilitation on functional outcomes in persons with stroke. Journal of
Rehabilitation Medicine, 38, 113-117.
Sanchez-Cubillo, I., Perianez, J. A., Adrover-Roig, D., Rodriguez-Sanchez, J. M., Rios-Lago, M., Tirapu, J., & Barcelo, F. (2009). Construct validity of the Trail Making Test: role of task-switching, working memory, inhibition/interference control, and visuomotor abilities. Journal of the International Neuropsychological Society, 15, 438-450.
Schmand, B., Groenink, S. C., & Van den Dungen, M. (2008). [Letter fluency: Psychometric properties and Dutch normative data]. Tijdschrift voor Gerontologie en Geriatrie, 39, 64-76.
del Ser, T., Barba, R., Morin, M. M., Domingo, J., Cemillan, C., Pondal, M., & Vivancos, J. (2005). Evolution of cognitive impairment after stroke and risk factors for delayed progression. Stroke, 36(12), 2670-2675.
Wechsler, D. (2008). Wechsler Adult Intelligence Scale–Fourth Edition (WAIS–IV). San Antonio, TX: NCS Pearson.
Wecker, N. S., Kramer, J. H., Hallam, B. J., & Delis, D. C. (2005). Mental flexibility: Age effects on switching. Neuropsychology, 19(3), 345-352.
Westerberg, H., Jacobaeus, H., Hirvikoski, T., Clevberger, P., Östensson, M. L., Bartfai, A., & Klingberg, T. (2007). Computerized working memory training after stroke-A pilot study. Brain Injury, 21(1), 21-29.
World Health Organization. (2003). The atlas of heart disease and stroke. From: http://www.who.int/cardiovascular_diseases/resources/atlas/en/
