Disruption of the balance between excitation and inhibition in the prefrontal cortex is thought to underlie cognitive problems in schizophrenia. A computational model of spatial working memory showed that disinhibition—caused by perturbing NMDA receptors on interneurons—broadens the tuning of memory-related neural activity, leading to more variable and less precise stored information and a reduced ability to filter out distractions. This prediction was tested with behavioral data from humans given ketamine, which induces disinhibition, and ketamine increased errors as the model predicted. The model also showed that restoring excitation-inhibition balance could reverse these working memory deficits, pointing to new experimental approaches for studying memory problems in schizophrenia.
Glutamate signaling through NMDA receptors is essential for brain computations that support cognition, and its disruption may contribute to schizophrenia. Using ketamine, an NMDA receptor antagonist, the study found that the normal anticorrelation between the default-mode and task-positive brain systems was disrupted during a working memory task. The degree of this disruption predicted task performance and produced schizophrenia-like symptoms. A computational model suggests that cortical disinhibition underlies this effect, linking glutamate's role in large-scale brain organization to cognition and psychiatric symptoms.