Schizophrenia is characterized by impairments in working memory (WM). Neurobiological studies, including nonhuman experimental studies, have led to the hypothesis that a deficit in dopamine transmission through D1 receptors in the prefrontal cortex (PFC) of patients with schizophrenia might be associated with the impairments of WM. However, there are limitations to measuring the processes of the complex WM system in the patients who are not medicated. These limitations cause difficulty of clarifying in more detail the mechanism which relates the impairments in WM to the decreased dopaminergic modulation via D1 receptors.
On the other hand, a computational modeling approach that is biologically plausible has helped form a theoretical framework of the impairments of WM in schizophrenia. In this study, we computationally investigated the relation between WM performance and dopaminergic modulation via D1 receptors using a leaky integrate-and-fire model of prefrontal cortex. Our results showed that decreased dopaminergic modulation unitarily caused reduced capacity and imprecision in WM. The imprecision of WM had characteristics of random drift. The decline of N-methyl-D-aspartate (NMDA) receptor function is the main factor of the reduced capacity. The decreases in both the ratio of NMDA-channel conductance to AMPA-channel conductance in pyramidal cells and the firings of interneurons are the main factors of the imprecision. Based on the mechanism, we theoretically suggest that combination therapy with antipsychotic drugs and drugs enhancing NMDA receptor function should be effective for the treatment of the impairments of WM in schizophrenia.