Geometric eigenmode brain fingerprinting and its longitudinal associations with adolescent mental health and wellbeing

Abstract

AbstractBackground‘Brain fingerprinting’ research posits that individual uniqueness can be identified by structural and functional features that may also be linked to mental health outcomes. Global structural features of the brain can be succinctly and directly captured from magnetic resonance imaging (MRI) via the eigenmodes of the cortical surface - known asgeometric eigenmodes. This research investigates how the uniqueness of geometric eigenmodes changes across adolescence and their longitudinal relation to mental health and wellbeing.MethodsThe current study utilised n=613 MRI, self-report and demographic datasets from N=116 community-recruited adolescents enrolled in the Longitudinal Adolescent Brain Study (LABS), between the ages of 12-17 years. MPRAGE scans at each participant’s visit were used to derive 225 left-hemisphere geometric eigenmodes. Eigenmodes were clustered into 14 eigengroups and developmental trajectories of their uniqueness and longitudinal associations with mental wellbeing and psychological distress were examined.ResultsAll eigengroups become significantly more unique longitudinally, and higher mode (shorter wavelength) eigengroups were more unique than lower mode groups in adolescence. Less uniqueness in ‘eigengroup 6’ was significantly associated with higher psychological distress and lower mental wellbeing at concurrent and future timepoints.ConclusionGeometric eigengroup brain fingerprinting offers a novel way to examine neurodevelopment. This study provides evidence that eigengroups have distinct trajectories from adolescence to adulthood, consistent with other imaging studies demonstrating increasing uniqueness in this period. Importantly, they are associated with mental health state and thus may represent neurobiological markers for mental illness onset, building on previous LABS research demonstrating that the functional uniqueness of the ‘cognitive control network’ predicts psychological distress four months later.