Increased prefrontal and parietal activity after training of working memory

Institution: Karolinska Institute

Title: Increased prefrontal and parietal brain activity after training of working memory

Researcher(s): Pernille J Olesen, Helena Westerberg & Torkel Klingberg

Program: Cogmed RM

Published: Nature Neuroscience, January 2004

Funding: This study was supported by the Swedish Research Foundation (Vetenskapsrådet), the Wallenberg Global Learning Network, and Cogmed Cognitive Medical Systems AB. Training program, investigator training, and technical support provided by Cogmed. No funding provided by Cogmed.

Working memory capacity has traditionally been thought to be constant. Recent studies, however, suggest that working memory can be improved by training. In this study, we have investigated the changes in brain activity that are induced by working memory training. Two experiments were carried out in which healthy, adult human subjects practiced working memory tasks for 5 weeks. Brain activity was measured with functional magnetic resonance imaging (fMRI) before, during and after training. After training, brain activity that was related to working memory increased in the middle frontal gyrus and superior and inferior parietal cortices. The changes in cortical activity could be evidence of training-induced plasticity in the neural systems that underlie working memory.

Does working memory training lead to increased neural activity in brain regions known to be activated when individuals perform tasks that depend on working memory capactiy? Documenting such changes would be an important step in explaining the positive behavioral and cognitive effects of working memory training that have been reported in several studies by pointing to observable changes in brain functioning that result from training. This was the focus of the two-part study described below.

In Experiment One, three healthy adult subjects completed 90 working memory trials per day for between 20 and 30 days; fMRI scans were completed before and after training. Compared to eleven comparison subjects who were not trained, trained subjects made significant improvements in several non-trained tasks, i.e., the Span Board Task and Ravens Advanced Progressive Matrices. Thus, positive effects of training on several tests of cognitive functioning were demonstrated. Especially noteworthy was that fMRI scans completed before and after training indicated increased activity in prefrontal and parietal cortical areas when subjects performed working memory dependent tasks. This represented initial indication that training produced increased activation of brain areas that are activated during working memory tasks.

Experiment Two was an extended replication of this study conducted with eight healthy adults. These adults completed five weeks of working memory training using three different visuospatial tasks, i.e., recalling the position of items presented briefly on the screen. Training occurred five days per week and participants completed 90 working memory trials during each session. Subjects were scanned weekly during the training while performing both a working memory task and a control task. As in Experiment One, post-training assessments indicated significant improvement on non-trained tasks of working memory and cognitive functioning, i.e., Digit Span Test and Span Board. Furthermore, fMRI scans again indicated increased activation in prefrontal and parietal regions when subjects performed working memory dependent tasks. Thus, the impact of working memory training on neural activity was replicated.

Results from these experiments are important in that they provide initial evidence that behavioral changes following working memory training are associated with changes in brain activity in areas critical for working memory performance. The authors suggest that these changes in neural activity may provide a basis for the improved working memory capacity that subjects demonstrated on the tests. They note, however, that additional research is required to better understand the functions of those areas where training-induced increases in activity was observed. And, although studies that involve fMRI scans are often based on small samples, replicating these findings with a larger sample and in samples of individuals diagnosed with ADHD would also be important to pursue. Nonetheless, these are interesting findings and suggest a biological mechanism by which training may lead to real world improvements in cognitive functioning.

Nature Neuroscience 7:75-79

Link to abstract