Brain state dynamics and working memory in patients with schizophrenia and unaffected siblings.

Abstract

Working memory (WM) deficits are a key feature of schizophrenia and are also seen in unaffected siblings. These deficits might arise  from disrupted transitions from one brain state to another. Using a robust algorithm called the Bayesian Switching Dynamical System (BSDS), we studied hidden brain states and their transitions during a WM task in people with schizophrenia. We used BSDS to identify brain states based on regions of interest (ROIs) within the default mode network and the frontoparietal network in 161 patients with schizophrenia, 37 unaffected siblings, and 96 healthy controls during N-back (0, 2, and resting fixation) tasks. We estimated group differences in the properties of brain states and studied the influence of WM performance and clinical characteristics on them using General Linear Models. We identified 4 brain states underlying the WM task: high-demand, low-demand, fixation, and non-dominant states. Compared with controls and siblings, patients showed reduced occupancy and lifetime of high-demand state during the "2-back," reduced lifetime of low-demand state during the "0-back," but increased occupancy and lifetime of fixation state during both task periods. Aberrant high-demand state mediated the association between WM performance and negative symptoms. Compared with controls and patients, siblings showed increased occupancy of high-demand and reduced fixation state during the resting fixation condition; this putative compensatory process correlated with better WM performance. Latent brain states of intrinsic connectivity that represent internal mental processes affect WM performance, influencing the expression of negative symptoms in schizophrenia and cognitive resilience in unaffected siblings.