Linking Individual Differences in Personalized Functional Network Topography to Psychopathology in Youth

Abstract

The spatial layout of large-scale functional brain networks differs between individuals and is particularly variable in association cortex that has been implicated in a broad range of psychiatric disorders. However, it remains unknown whether this variation in functional topography is related to major dimensions of psychopathology in youth. Capitalizing on a large sample with 27-minutes of high-quality functional MRI data (n=790, ages 8-23 years) and advances in machine learning, we examined associations between functional topography and four correlated dimensions of psychopathology (fear, psychosis, externalizing, anxious-misery) as well as an overall psychopathology factor. We found that functional topography significantly predicted individual differences in dimensions of psychopathology, driven mainly by robust associations between topography and overall psychopathology. Reduced cortical representations of association networks were among the most important features of the model. Our results emphasize the value of considering systematic differences in functional neuroanatomy for personalized diagnostics and therapeutics in psychiatry.

Multimodal neural correlates of childhood psychopathology

“…Moreover, our findings recapitulated the hierarchy within psychopathology (Kotov et al, 2017;Lahey et al, 2017), with the p factor explaining the highest covariance across multiple imaging features, and distinct structural and functional signatures related to internalizing, externalizing and neurodevelopmental dimensions. Previous research had found that disentangling the variation due to the p factor from other dimensions could be challenging, with sometimes few morphological or connectivity changes related to second-order dimensions left after adjusting for the p factor (e.g., (Cui et al, 2021;Parkes et al, 2021)). By using PLS, which derives orthogonal components (McIntosh and Lobaugh, 2004;McIntosh and Mišić, 2013), we ensured that the neural substrates associated with LCs 2-5 were independent from those associated with the p factor (i.e., LC1).…”