Longitudinal Network Analysis Reveals Interactive Change of Schizophrenia Symptoms during Acute Antipsychotic Treatment

Abstract

Background and Hypothesis: Complex schizophrenia symptoms were recently conceptualized as interactive symptoms within a network system. However, it remains unknown how a schizophrenia network changed during acute antipsychotic treatment. The present study aimed to evaluate the interactive change of schizophrenia symptoms under seven antipsychotics from individual time series. Study Design: Data on 3030 schizophrenia patients were taken from a multicenter randomized clinical trial and used to estimate the partial correlation cross-sectional networks and longitudinal random slope networks based on mul-tivariate multilevel model. Thirty symptoms assessed by The Positive and Negative Syndrome Scale clustered the networks. Study Results: Five stable communities were detected in cross-sectional networks and random slope networks that describe symptoms change over time. Delusions, emotional withdrawal, and lack of spontaneity and fow of conversation featured as central symptoms, and conceptual disorganization, hostility, uncooperativeness, and diff-culty in abstract thinking featured as bridge symptoms, all showing high centrality in the random slope network. Acute antipsychotic treatment changed the network structure (M-test = 0.116, P