Altered MD-PFC thalamocortical network dynamics in an NMDAR antagonist animal model of cognitive control deficits in Schizoprenia

Abstract

Schizophrenia (SZ) is debilitating neuropsychiatric disorder and afflicted patients exhibit an array of symptoms including hallucinations, delusions, flat affect, and deficits in cognitive functioning. Even though the functional outcome of patients with SZ is correlated with the severity of cognitive deficits, current therapeutics do not effectively improve cognitive deficits. However, there is still not a clear understanding of the pathophysiology underlying cognitive deficits, making it difficult to develop therapeutics target at the improvement of cognition in patients. The mediodorsal nucleus of the thalamus (MD) has been thought to play a role in cognitive control behaviors alongside the prefrontal cortex (PFC). This is largely due to the reciprocal connections between the MD and PFC, reports of MD neurons exhibiting similar task-related activity patterns during cognitive behaviors, and that lesions in the MD produce deficits in cognition similar to PFC lesions. Also, in patients with SZ the MD and PFC exhibit correlated reduced activation during cognitive control performance and are function disconnected from each other. In order to further our understanding of the pathophysiology in the MD-PFC thalamocortical network, I trained monkeys to perform a cognitive control task that exposes cognitive control deficits in patients with SZ. First, I relate thalamocortical distributed processing and functional coupling to cognitive control by recording in the MD and PFC simultaneously in the healthy state. Then, I characterize the effects of NMDA receptor (NMDAR) blockade on thalamocortical distributed processing and functional coupling underlying cognitive deficits. I found that MD neurons represent cognitive control state similarly to PFC neurons, but that PFC state neurons contain more information early in the trial, while MD state neurons contain more information late in the trial, MD neurons are more involved in response selection than PFC neurons and neurons in the MD and PFC transmit state and response information reciprocally during cognitive control performance. Following NMDAR blockade, there were less MD and PFC neurons recruited to encode state and response information and the neurons that did encode task information were delayed in their recruitment. Additionally, representation of task state was decreased at the single neuron and population levels in both the MD and PFC, but to a stronger degree in the PFC than the MD, and these changes in state representation predicted task failure on a trial-by-trial basis more strongly in the MD than the PFC. Lastly, NMDAR blockade strongly attenuated transmission of information in local circuits in and between the MD and PFC. Overall, I am the first to characterize MD-PFC cellular level network dynamics underlying cognitive control and the effect of NMDAR blockade on this thalamocortical network that results in pathophysiology and SZ-like cognitive deficits.