Neurostructural Heterogeneity in Youths With Internalizing Symptoms

Abstract

Internalizing disorders such as anxiety and depression are the most common psychiatric disorders, frequently begin in youth, and exhibit marked heterogeneity in treatment response and clinical course. It is increasingly recognized that symptom-based classification approaches to internalizing disorders do not align with underlying neurobiology. An alternative to classifying psychopathology based on clinical symptoms is to identify neurobiologically-informed subtypes based on brain imaging data. We used a recently developed semi-supervised machine learning method (HYDRA) to delineate patterns of neurobiological heterogeneity within youth with internalizing symptoms using structural imaging data collected at 3T from a large communitybased sample of 1,141 youth. Using volume and cortical thickness, cross-validation methods indicated a highly stable solution (ARI=.66; permutation-based pfdr < .001) and identified two subtypes of internalizing youth. Subtype 1, defined by smaller brain volumes and reduced cortical thickness, was marked by impaired cognitive performance and higher levels of psychopathology than both Subtype 2 and typically developing youth. Using resting-state fMRI and diffusion images not considered during clustering, we found that Subtype 1 also showed reduced amplitudes of low-frequency fluctuations in fronto-limbic regions at rest, as well as reduced fractional anisotropy in white matter tracts such as the parahippocampal cingulum bundle and the uncinate fasciculus. In contrast, Subtype 2 showed intact cognitive performance, greater volume, cortical thickness, and amplitudes during rest compared to Subtype 1 and typically developing youth, despite still showing clinically significant levels of psychopathology.Identification of biologically-grounded subtypes of internalizing disorders may assist in targeting early interventions and assessing longitudinal prognosis..

Identification of transdiagnostic psychiatric disorder subtypes using unsupervised learning

“…The wealth of available data and advances in machine learning intensified efforts to redefine disorder categories using datadriven methods. Previous studies stratified psychiatric disorders mostly by clustering single domains (e.g., psychometry [6][7][8][9][10], neuroimaging [11][12][13][14][15][16], biochemical markers [17], or genetics [18,19]) or by analyzing patients from a single diagnosis (e.g., major depressive disorder (MDD) [5,7,11,[18][19][20][21][22] or schizophrenia (SCZ) [23][24][25][26][27][28]). Previous transdiagnostic clustering studies support the existence of diagnostically mixed subtypes across two [29][30][31] or more disorders [32][33][34][35].…”

Development and Application of Novel Performance Validity Metrics for Computerized Neurocognitive Batteries

“…There were few associations between our validity metrics and clinical diagnoses of depression, PTSD, anxiety, psychosis spectrum disorders, or externalizing disorders, supporting the potential validity of our approach. This feature is especially relevant considering neurocognitive performance differences in these conditions from prior work (e.g., Barzilay et al, 2019;Gur et al, 2014;Kaczkurkin et al, 2020;Service et al, 2020). However, there were diagnoses with reduced validity metrics that should be addressed.…”

Mental health in the UK Biobank: a roadmap to self-report measures and neuroimaging correlates

“…A second potential reason for limited effect sizes (even with the use of multivariate methods like CCA) is between-subject heterogeneity. A first potential type of heterogeneity is diversity in symptoms, such that two patients with depression may present with largely non-overlapping symptom profiles [Drysdale et al, 2017; Feczko et al, 2019; Feczko and Fair, 2020; Kaczkurkin et al, 2020]. A second potential type of heterogeneity is diversity in psychophysiological disease mechanisms.…”