I’m presenting a poster at the 2012 Society for Neuroscience meeting, on computational modeling results describing working memory deficits induced by schizophrenia-related disinhibition. My poster is titled: Effects of disinhibition in a cortical working memory circuit model with relevance to schizophrenia. It will be presented on the morning of Tuesday, October 16. (I also contributed computational modeling to another related poster: NMDA receptor function in large-scale anti-correlated neural systems: implications for cognition and schizophrenia.) My abstract is below:
Effects of disinhibition in a cortical working memory circuit model with relevance to schizophrenia
J. D. Murray, A. Anticevic, M. Ichinose, M. Gancsos, P. R. Corlett, J. H. Krystal, X.-J. Wang
Excitation-inhibition balance (E/I balance) is a fundamental property of cortical microcircuitry. In the prefrontal cortex (PFC), E/I balance is critical for persistent activity supporting working memory (WM), and disruption of E/I balance in PFC is hypothesized to be associated with WM deficits in mental disorders such as schizophrenia (SCZ). Specifically, NMDA receptor hypofunction on interneurons may contribute to cognitive dysfunction in SCZ. This deficit hinders the recruitment of feedback inhibition by local pyramidal-cell activity, resulting in cortical disinhibition. However, it is poorly understood how synapse-level disinhibition may lead to behavioral deficits in WM. To study this, we incorporated synaptic disinhibition into a computational model of spatial WM, implemented with spiking neurons and realistic synaptic conductances.
We found that WM function is highly sensitive to E/I balance. Even a small reduction in inhibition profoundly degrades neural activity and behavior. At the neural level, disinhibition broadens tuning of persistent activity patterns, and thereby decreases the population signal-to-noise ratio. At the behavioral level, we characterize two primary deficits: i) behavioral variability, due to noise-induced drifts of the population firing pattern over time, is increased by disinhibition, degrading WM precision; ii) resistance to incoming distractors is impaired by disinhibition, due to increased overlap in target/distractor neural representations. We show, as proof-of-principle, that neural and behavioral degradations can be ameliorated by treatments that restore E/I balance via compensatory mechanisms. To test the core prediction of a broadened WM representation under disinhibition, we examined behavior of human subjects performing a spatial WM match/non-match task. Ketamine administration, a pharmacological model of SCZ hypothesized to induce cortical disinhibition through NMDA receptor antagonism on interneurons, increased errors as predicted by the model.
Our findings detail the importance of E/I balance for cortical microcircuits underlying cognitive function. The model makes specific neural and behavioral predictions following synaptic disinhibition, which can be tested under pharmacological manipulation or in disease states. We propose important features of task design for sensitive probing of these effects.