A longitudinal study of computerized cognitive training in stroke patients - effects on cognitive function and white matter.

Background Computerized cognitive training is suggested to enhance attention and working memory functioning following stroke, but effects on brain and behavior are not sufficiently studied and longitudinal studies assessing brain and behavior relationships are scarce. Objective The study objectives were to investigate relations between neuropsychological performance post-stroke and white matter microstructure measures derived from diffusion tensor imaging (DTI), including changes after 6 weeks of working memory training. Methods In this experimental training study, 26 stroke patients underwent DTI and neuropsychological tests at 3 time points - before and after a passive phase of 6 weeks, and again after 6 weeks of working memory training (Cogmed QM). Fractional anisotropy (FA) was extracted from stroke-free brain areas to assess the white matter microstructure. Twenty-two participants completed the majority of training (≥18/25 sessions) and were entered into longitudinal analyses. Results Significant correlations between FA and baseline cognitive functions were observed (r = 0.58, p = 0.004), however, no evidence was found of generally improved cognitive functions following training or of changes in white matter microstructure. Conclusions While white matter microstructure related to baseline cognitive function in stroke patients, the study revealed no effect on cognitive functions or microstructural changes in white matter in relation to computerized working memory training.