Functional network connectivity deviation is associated with transdiagnostic symptomatology.

Abstract

Background: Large comorbidity and heterogeneity within psychiatric populations have prompted the use of transdiagnostic methods to account for this variability in brain-phenotype associations. Normative modelling offers a way to map individual deviations in brain functioning with respect to a large reference population. This study aims to explore brain-phenotype associations, using normative modelling to compute individual deviation scores of brain functioning, and relating them to different levels of psychopathology within a naturalistic patient sample.

Methods: We applied normative modelling to estimate individual deviations in brain functional connectivity in a naturalistic sample (N=309) comprising both patients and healthy controls. We examined the association between the resulting neural deviation scores and levels of psychopathology, including traditional diagnostic categories, transdiagnostic symptom profiles, and cognition measures using sparse canonical correlation analysis (sCCA) RESULTS: We successfully estimated normative models using data from the MIND-Set study, and found significantly more extreme deviation scores in the patient as compared to the control population. We found a significant association (R_c=0.16, R² = 2.56%, p=0.021) between neural deviations scores and transdiagnostic symptom profiles, aligning with four Research Domain Criteria (RDoC) domains: negative valence, cognition, arousal/inhibition and social systems.

Conclusions: With the use of normative modelling, we could detect differences in functional brain connectivity in patients as compared to controls, even in a highly heterogeneous and comorbid patient sample. Additionally, transdiagnostic approaches, like those embodied in the RDoC framework, are more accurate in uncovering shared neurobiological mechanisms than traditional diagnostic categories or cognitive measures.