The Common Structure of the Major Psychoses: More Similarities Than Differences in the Network Structures of Schizophrenia, Schizoaffective Disorder, and Psychotic Bipolar Disorder.

Abstract

There has been a century-long debate about whether the major psychoses (eg, bipolar disorder, schizophrenia, and schizoaffective disorder) are one disorder with various manifestations or different disease entities. Traditional approaches using dimensional models have not provided decisive findings. Here, we address this question by examining the network constellation of affective and psychotic syndromes. Comparable symptom data of 1882 patients with psychotic bipolar disorder, schizoaffective disorders, and schizophrenia were extracted from three datasets: B-SNIP 1, B-SNIP2, and PARDIP. Twenty-six items from the Positive and Negative Syndrome Scale, YMRS, and the Montgomery-Asberg Depression Rating Scale were selected for the analysis using a principled approach to eliminate overlapping/redundant items. Gaussian graphical models were estimated and assessed for stability, and their communities were identified using bootstrapped exploratory graph analysis. The structures and global densities of the networks were compared with network comparison tests. The network structures were highly similar (r >. 80) across diagnostic groups. For all diagnoses, manic symptoms were more connected with positive symptoms while depressive symptoms were more linked with negative symptoms. The depressive and negative symptoms were the strongest indicators of depressive and psychotic communities. Theoretically interesting variability in network edge weights between symptoms was found relating to thought disorder and pessimistic thinking. The same broad structure of psychopathology underlies the symptom expressions of bipolar disorder, schizoaffective disorder, and schizophrenia. Future studies should build on the present finding by comparing specific inter-relations between symptoms in the different diagnostic groups using methods capable of detecting causality.