Right hippocampal volume and attenuated psychotic symptoms distinguish subgroups of youth with common patterns of temporoparietal effective connectivity

Abstract

Associations between patterns of brain connectivity and brain structural features have transdiagnostic relevance to psychopathology. There is considerable evidence for disruptions to the hippocampus and temporoparietal brain systems in psychotic disorders. The present study examines structure–function relationships—specifically, the relation of hippocampal volume with patterns of temporoparietal effective connectivity—in youth at clinical high-risk for psychosis (CHR-P) and healthy controls (HCs). Participants at CHR-P and HCs completed clinical symptom measures and magnetic resonance imaging at baseline (n = 388, 42.5% female, age = 19.8 ± 4.2) in the second cohort of the North American Prodrome Longitudinal Study. Group Iterative Multiple Model Estimation established a common functional network of temporoparietal effective connectivity in the full sample. Next, supervised (assuming the CHR-P and HC groups represent classes) and unsupervised (data-driven) clustering procedures interrogated effective connectivity parameters relevant to subsamples. Mean difference tests by unsupervised cluster membership determined whether clusters formed based on temporoparietal effective connectivity differed in hippocampal volume. Clusters were also compared on representation of CHR-P and positive and negative symptom totals to determine whether unsupervised clustering recovered clinical features. Unsupervised clustering generated two clusters that differed significantly in right hippocampal volume and attenuated positive and negative symptoms with small-to-medium effect sizes. The cluster demonstrating reduced right hippocampal volume reported greater symptoms. Unsupervised clustering did not recover diagnostic groups. Findings are consistent with literature indicating the transdiagnostic relevance of hippocampal volume to organizational properties of brain functional networks and patterns of temporoparietal connectivity.