Schedule of Events | Search Abstracts | Invited Symposia | Symposia | Poster Sessions | Data Blitz

Neuroimaging Biomarkers for Predicting Substance Use Trajectories in Adolescents: Insights from the ABCD Study

Poster Session C - Sunday, March 8, 2026, 5:00 – 7:00 pm PDT, Fairview/Kitsilano Ballroom

Alex Xinyue Teng¹ (), Rebecca Elliott¹, Tim Cootes¹; ¹University of Manchester

Introduction: Early identification of individuals at risk for escalating substance use is crucial for preventative interventions. This study uses neuroimaging data from the Adolescent Brain Cognitive Development (ABCD) study to explore early brain markers predictive of substance use escalations. Methods: Adolescents with no substance use at baseline (ages 9-10) were followed for four years, and categorised into low-risk, mild escalation, and severe escalation groups based on their consumption patterns for alcohol, nicotine and other substances. Multimodal neuroimaging data, including T1-weighted, resting-state fMRI, and diffusion tensor imaging, were analysed using one-way ANOVA tests with family-wise FDR corrections to control for multiple comparisons. Topological analyses of brain networks were conducted to assess network properties. Logistic regression models with 10-fold cross-validation were also employed to predict substance use status based on neuroimaging features. Results: The cohort included 319 participants (130 low-risk, 96 mild, 93 severe) with no demographic differences at baseline. Morphological alterations in posterior occipital regions and widespread microstructural changes were found in escalation individuals. Severe escalation was further distinguished by altered tensor diffusivity in the left cingulate and subparietal regions. Network analyses indicated varied clustering, local efficiency and small-worldness in escalation, suggesting impaired segregation and integration. Changes in betweenness and connectivity were observed within auditory and cingulo-parietal networks as well. Moreover, a combined model incorporating structural and functional markers achieved 72% accuracy in classifying escalation. Conclusion: Early-occurring neuroimaging, particularly within occipital, sensorimotor, and auditory systems, may underlie vulnerability to escalating substance use and serve as potential biomarkers for targeted early interventions.

Topic Area: METHODS: Neuroimaging