The relationship between psychopathology and neurocognitive functioning in affective and psychotic disorders: A multimodal Bayesian network analysis

BACKGROUND: Affective and psychotic disorders share overlapping symptom constellations, environmental and genetic risk factors, and neurocognitive profiles. However, the nature of this association is not well understood, and accumulating evidence suggests a dimensional rather than categorical distinction between healthy and clinical populations. OBJECTIVE: This study investigated the relationships linking childhood trauma, positive and negative symptoms, depression symptoms, anxiety symptoms, and neurocognitive functioning such as verbal intelligence, executive functioning, and semantic processing within a unified transdiagnostic network. METHODS: We employed a partial correlation network and directed acyclic graph (DAG) analysis in a large psychiatric sample of 2444 participants, including 1364 patients with affective disorders or psychotic disorders, and 1080 healthy controls. Using self-report scales, clinician ratings, and neurocognitive tests, we analyzed risk and symptom clusters and developed a preliminary predictive model of these associations. RESULTS: Distinct, interconnected clusters of psychopathological symptoms emerged in the Gaussian Graphical Model (GGM), with strong connections between depression and anxiety clusters. Neurocognition and childhood trauma showed sparser associations with psychopathological symptoms. Nodes from self- and observer-ratings formed surprisingly strong bridges, especially concerning libido problems, lack of hobbies, social anhedonia, and loss of interest. DAG analyses indicated symptoms of low mood, reduced wellbeing, not feeling safe, and having difficulties initiating work or activities in general as potential predictors of downstream symptoms. CONCLUSION: Complex associations between psychopathological and neurocognitive functioning emerged, with (the lack of) self-reported happiness, and well-being as key nodes. Future research should validate the clinical utility by using longitudinal network analyses and experimental data.