Brain-predicted age associates with psychopathology dimensions in youth

Abstract

Background: This study aims to investigate whether dimensional constructs of psychopathology relate to advanced, attenuated or normal patterns of brain development, and to determine whether these constructs share common neurodevelopmental profiles.Methods: Psychiatric symptom ratings from 9312 youths (8-21 years) were parsed into 7 independent dimensions of clinical psychopathology representing conduct, anxiety, obsessivecompulsive, attention, depression, bipolar, and psychosis symptoms. Using a subset of this cohort with structural MRI (n=1313), a normative model of brain morphology was established and the model was then applied to predict the age of youth with clinical symptoms. We investigated whether the deviation of brain-predicted age from true chronological age, called the brain age gap, explained individual variation in each psychopathology dimension.Results: Individual variation in the brain age gap significantly associated with clinical dimensions representing psychosis (t=3.16, p=0.0016), obsessive-compulsive symptoms (t=2.5, p=0.01), and general psychopathology (t=4.08, p<0.0001). Greater symptom severity along these dimensions was associated with brain morphology that appeared older than expected for typically developing youth of the same age. Psychopathology dimensions clustered into two modules based on shared brain loci where putative accelerated neurodevelopment was most prominent. Patterns of morphological development were accelerated in frontal cortices for depression, psychosis and conduct symptoms (Module I), whereas acceleration was most evident in subcortex and insula for the remaining dimensions (Module II). Conclusions:Our findings suggest that advanced brain development, particularly in frontal cortex and subcortical nuclei, underpins clinical psychosis and obsessive-compulsive symptoms in youth.Psychopathology dimensions share common neural substrates, despite representing clinically independent symptom profiles.

Individual variation underlying brain age estimates in typical development

“…As with previous studies in this cohort, we did not find strong evidence of association between brain age delta and cognitive performance, after accounting for age and sex [Ball et al, 2017;Khundrakpam et al, 2015]. Other studies have reported small to moderate associations between brain age and measures of cognitive performance during typical development [Erus et al, 2015;Lewis et al, 2018], or associations between brain age estimates and symptoms of psychopathology or neurodevelopmental disorders [Cropley et al, 2020;Tunç et al, 2019]. Our findings do not preclude the detection of brain age-behaviour associations in typically-developing samples, others have employed multi-modal approaches to successfully predict brain age from brain structure and function, additionally using measures of tissue microstructure from diffusion MRI and subcortical metrics rather than just cortical thickness and area [Erus et al, 2015;Lewis et al, 2019;Liem et al, 2017].…”

“…In children with autism spectrum disorder, a negative brain age gap (reflecting that the brain was predicted to be younger than expected for chronological age) was related to greater disorder symptom severity [Tunç et al, 2019]. In a non-typically developing sample of children and adolescents, a positive brain age gap was associated with greater psychopathology symptom severity [Cropley et al, 2020].…”

Neighborhood disadvantage and longitudinal brain-predicted-age trajectory during adolescence

“…The present study included 1313 participants (mean age 14.5 years, SD = 3.43, 659 F) for whom a structural magnetic resonance imaging (MRI) scan was acquired. This sample was obtained after excluding participants on the basis of image quality, medical conditions, and missing data (n = 285; for details see Cropley et al, 2020 ). Structural MRI scans from this final PNC sample were used to train the normative brain age model described in subsequent sections.…”

“…Using data from the PNC sample, a support vector regression (SVR) model (using 10-fold cross-validation) was used to predict each individual’s age (in the PNC sample) based on the 111 measures listed above (i.e., the normative model; Cropley et al, 2020 ). All measures were standardized.…”

“…All measures were standardized. Model performance within the PNC sample was quantified in a previous study (see Cropley et al, 2020 and Section 3 .2 for model performance within the PNC sample). The present study used the previously trained normative model to predict the age of the OADS sample at each time point.…”