Symptom networks in manic and mixed presentations of bipolar disorder: Mapping the crossroads.

BACKGROUND: Mixed states in bipolar disorder (BD) remain a diagnostic and nosological challenge, with current classifications differing in their ability to capture the overlap between manic and depressive symptomatology. METHODS: We used baseline data from 1384 patients in a manic or mixed episode of BD according to DSM-IV criteria across five randomized clinical trials. We modelled mood symptom networks using the Young Mania Rating Scale (YMRS) and Montgomery-Åsberg Depression Rating Scale (MADRS). We identified mood symptom clusters and compared symptom networks between DSM-IV manic and mixed episodes, and between DSM-IV mixed episodes and DSM-5 manic episodes with mixed features. RESULTS: The joint YMRS-MADRS mood symptom network revealed strong within-domain coherence (i.e., strong connectivity among manic symptoms and among depressive symptoms). There were positive and negative associations between manic and depressive nodes, indicating that these domains both co-occur and inhibit each other. Depressive symptoms, especially sadness and anhedonia items, showed greater overall connectivity and centrality than manic symptoms. Cluster analyses identified three main domains - manic, depressive, and a mixed activation-dysphoria cluster -plus a smaller insomnia-related cluster. DSM-IV manic and mixed episodes differed significantly in structure and global connectivity, whereas DSM-5 distinctions were subtler and reflected increased depressive burden within manic presentations. CONCLUSIONS: Manic and depressive symptoms formed distinct yet interconnected clusters within a shared affective network. An activation-dysphoria cluster linked both domains, capturing core elements of mixed presentations. DSM-IV groups showed clearly differentiated network structures, whereas DSM-5 mixed features reflected variations within the manic spectrum rather than a distinct configuration.