Dynamic Causal Modeling of P300 Processing in Schizophrenia

Shilai Zhang,Lin Jiang,Guangying Wang,Dezhong Yao,Peng Xu,Fali Li

DOI: https://doi.org/10.1109/civemsa58715.2024.10586544

2024-01-01

Abstract:Brain network disconnectivity, resulting in aberrant functional integration, underlies cognitive deficits and clinical symptoms in schizophrenia (SCZ). In a P300 experiment, we investigate the intrinsic (self-)connectivity of the frontoparietal cortical hierarchy by employing dynamic causal models (DCM) to estimate how evoked activity arises from the dynamic outcomes of coupled neural populations. EEG recordings were performed in 147 SCZ patients, 181 unaffected relatives (SCZRs), and 200 health controls (HCs). Of the eight network models, an optimal neural coupling model was identified separately on the individual model and the grand average model. Notably, the winning model included changes in connectivity at all three levels and self-inhibitory changes. In the individual model, information flow transmitted between the left frontal to left parietal lobes of SCZ patients was significantly attenuated in contrast to their relatives and HCs. While no significant difference was found between the control group and relatives. As for the grand-average model, patients showed increased voluntary inhibition across the left and right superior frontal gyrus, i.e., a decrease in cortical excitability, in different task conditions, when compared to controls. Relatives showed a state consistent with controls in the right upper frontal, a patient-like reduction in excitability in the left frontal, and a reversal of normal synaptic gain changes in the target tone relative to the standard tone. Overall, these findings shed light on the pathophysiological mechanisms of schizophrenia from a dynamic causal model perspective.