A biophysical model of multiple-item working memory: a computational and neuroimaging study.

Biophysically based computational models have successfully accounted for the persistent neural activity underlying the maintenance of single items of information in working memory. The aim of the present study was to extend previous models in order to retain multiple items, in agreement with the observed human storage capacity. This was done by implementing cellular mechanisms known to occur during the childhood development of working memory, such as an increased synaptic strength and improved contrast and specificity of the neural response. Our computational study shows that these mechanisms are sufficient to create a neural network which can store information about multiple items through sustained neural activity. Furthermore, by using functional magnetic resonance imaging, we found that the information-activity curve predicted by the model corresponds to that in the human posterior parietal cortex during performance of working memory tasks, which is consistent with previous studies of brain activity related to working memory capacity in humans.