State-Specific Segmental Abnormalities of White Matter Structure and Function in Pediatric Bipolar Disorders.

OBJECTIVE

White matter (WM), as the critical infrastructure for neural communication, is implicated in pediatric bipolar disorder (PBD) pathophysiology. However, the state-specific abnormalities in the segmental structural and functional properties of adolescent fiber tracts remain unclear.

METHODS

Processed fMRI and DTI data from 19 manic-phase PBD (MPBD), 20 euthymic-phase PBD (EPBD), and 17 matched healthy controls (HCs) were analyzed for structural and functional alterations across 100 segments of 8 WM tracts using Automated Fiber Quantification, with six machine learning classifiers subsequently employed for disease classification.

RESULTS

Tract-based analysis revealed predominant microstructural alterations in the left corticospinal tract (CST), characterized by concurrent increases in FA and CL alongside decreased RD, particularly in the posterior internal capsule and mid-portion of the tract. Functional ALFF abnormalities correlated with manic symptoms and episode frequency, suggesting affective dysregulation in emotion-regulation pathways (including anterior thalamic radiation and inferior fronto-occipital fasciculus). The random forest classifier, trained on a multimodal feature set combining DTI-derived tensor metrics, functional ALFF values, and clinical information, achieved 84% accuracy in distinguishing MPBD patients from HCs.

CONCLUSIONS

This multimodal study reveals segmental abnormalities in the structure and function of WM tracts across different mood episodes in PBD, identifying a structure-function desynchronization pattern: motor pathway exhibits enhanced structural integrity with preserved function, whereas emotion-regulation pathways display functional hyperactivity without structural alterations. These alterations correlate with cognitive-executive impairment and enable clinical subtype differentiation, providing novel insights into PBD neuropathology and supporting neuroimaging-based diagnostic development.