Computerized working memory training after stroke – A pilot study

Institution: Karolinska Institute

Title: Computerized working memory training after stroke – A pilot study

Researcher(s): H. Westerberg; H. Jacobaeus; T. Hirvikoski; P. Clevberger; M-L. Östensson; A Bartfai; T. Klingberg.

Program: Cogmed RM

Published: Brain Injury, June 2007

Funding: The Swedish Stroke Foundation supported this study. Training program, investigator training, and technical support provided by Cogmed. No funding provided by Cogmed.

Overview
Deficits in working memory (WM) and attention are frequent outcomes after stroke and the degree of WM impairment following stroke predicts recovery and the probability of returning to work. For these reasons, training WM in stroke victims could play an important role in maximizing their recovery.

This study was a pilot investigation testing whether Cogmed Working Memory Training (CWMT) could enhance WM functioning in adult stroke victims. Participants were eighteen 34-65 year olds who had suffered strokes one to three years earlier. All were former patients on a stroke rehabilitation unit at a hospital in Stockholm, Sweden. They were randomly assigned to receive five weeks of Cogmed Working Memory Training or to a no treatment control condition.

CWMT included eight different computer exercises designed to train visuo-spatial WM – remembering the position of objects that were presented briefly on the screen – as well as verbal tasks – remembering sequences of letters, sounds, and digits. As an example of a visuo-spatial task, lights would appear at different locations on the screen and participants had to remember the order and click the locations in the correct sequence. On a verbal working memory task, participants heard a series of digits and had to click the numbers in the reverse order in which they were presented.

Training was conducted five days a week over five weeks with patients completed the training at home on their computer. Daily training results were transferred via the internet so that progress could be monitored remotely. Weekly phone calls made by a Cogmed coach were used to provide feedback and encouragement,and to instruct participants on how to maximize their training.

Each training session provided exposure to 90 WM tasks and required about 40 minutes to complete. Throughout training, the difficulty level of the WM tasks was adjusted to match participants’ ability by modifying the number of elements to be recalled on a trial-by-trial basis. For example, if the location of four lights was correctly recalled, the next trial would require participants to recall the location of five lights. If a trial was missed, the next trial was made easier by dropping the number to three. Through this method of ‘adaptive’ training, participants were challenged to work at their highest level with the goal of enhancing WM capacity over time.

Outcome Measures
Training impact was measured by administering a series of eight neuropsychological tests before and after training. These tests measured a variety of cognitive functions including visuo-spatial and verbal WM, inhibitory control, and problem solving. Importantly, the tests of WM were different from those used during the training itself. Thus, improvement from pre- to post-testing on these tests would indicate that training benefits generalized to non-trained measures of WM.

Participants also completed the Cognitive Functioning Questionnaire (CFQ) before and after training. This is a 25-item measure to assess cognitive functioning in daily life, e.g., how frequently one experiences attention lapses and memory problems. Positive change on the CFQ would suggest that CWMT resulted in improved cognitive functioning that individuals noticed in their daily life and represents an important addition to using neuropsych assessments alone.

Results
To measure training effectiveness, researchers compared the amount of change made by training and control participants. This enabled them to learn whether CWMT yielded gains above and beyond those attributable to practice effects on the neuropsych measures or simply to the passage of time.

Significant training effects were found on four of the eight neuropsych measures; all occurred on tests that assessed WM or attention. The size of these training effects were in a range that would be considered moderate to large by conventional standards, i.e., effect sizes ranged from .61 to 1.58. Training-related gains on tests of inhibitory control or simple short term memory tasks that did not explicitly tax WM were not evident.

Participants’ self report of daily cognitive functioning also showed significant benefits of CWMT. Problems reported by the training group declined significantly over time (reduced problems were reported by 8 of the 9 participants) while problems reported by control participants increased slightly. Items on which training participants reported the greatest improvements were those that dealt most directly with issues related to attention.

Summary and Implications
Results from this small pilot investigation suggest that intensive working memory training can help stroke victims improve WM and attentional functioning in ways that individuals notice in their daily lives. This is an important demonstration and suggests that working memory training may prove to be an important tool in rehabilitation from stroke.

This study has several limitations that the researchers are clear in acknowledging. First, the sample was small and replicating these findings with a larger and more diverse sample of stroke victims would be an important next step. Second, the control group was a passive one, i.e., they did not receive any form of intervention to which CWMT was compared. As a result, one cannot rule out that engaging in any type of intensive computer activity for five weeks – as opposed to their being specific benefits of CWMT, would have resulted in similar gains.

Brain Injury, 2007; 21(1): 21–29

Link to abstract