Simulating brain dynamics in healthy subjects and schizophrenic patients: a first exploratory study.

Abstract Background: Schizophrenia is a severe psychiatric disorder marked by positive, negative and cognitive symptoms. A particularly challenging subset of patients are those with treatment resistant schizophrenia (TRS). TRS is characterized by a failure to respond to standard antipsychotic therapy, highlighting critical need for deeper insights into it's neurobiology. Objectives: This study aimed to investigate neural dynamics in TRS, non-TRS and healthy controls (HC) using The Virtual Brain (TVB) neuroinformatics platform. Differences in excitation/inhibition profiles between patient groups were assessed both at the whole-brain level and within the Default Mode Network (DMN). Methods: Data from 40 subjects were analyzed across three groups (TRS, non-TRS and HC). Functional and structural connectomes built using magnetic resonance imaging data were extracted and fed to the TVB to build personalized models of brain dynamics. Brain simulations used the Wong-Wang neural mass model to extract neurophysiological indices. Results: Significant group differences emerged from whole-brain simulations. In particular, SZ patients showed elevated recurrent excitation (w+), Kolomogorov-Smirnov (KS) distance and model cost function compared to HC. When comparing SZ subgroups, non-TRS individuals exhibited significantly higher model cost function than HC, whereas TRS patients did not differ significantly from either group. No signifficant differences emerged from the DMN-specific simulations. Conclusion: This study demonstrates the utility of TVB for capturing disrupted neural dynamics in SZ. Our findings suggest that computational modeling can be a useful research tool in neuroscience with potential for exploring complex brain disorders in silico. Keywords: Schizophrenia, TRS, The Virtual Brain, Multiscale Brain modeling, MRI