28. SHARED AND DISTINCT CORTICAL MORPHOMETRIC SIMILARITY ALTERATIONS IN SMOKING AND ALCOHOL USE CORRELATE WITH NEUROTRANSMITTER AND TRANSCRIPTOMIC SIGNATURES

Background Smoking and alcohol use are major modifiable risk factors for brain health and frequently co-occur. However, their shared and distinct associations with multi-modal cortical morphometric similarity networks (MSNs) remain unclear. This study aimed to investigate the common and distinct cortical MSN alteration patterns associated with smoking and alcohol use, as well as the potential biological mechanisms underlying MSN alterations. Methods We included 39,875 UK Biobank participants with FreeSurfer-processed T1- and T2-weighted MRI. Cortical MSNs were constructed using eight structural and microstructural features. Associations of smoking and alcohol use with global and regional MSNs were examined using linear regression and ANCOVA, adjusting for age, sex, imaging site, education, estimated total intracranial volume, Townsend deprivation index, with smoking and alcohol use mutually adjusted for each other. Robustness was assessed using Bonferroni correction and 1,000 bootstrap iterations. Regional effect maps were linked to Allen Human Brain Atlas gene expression using PLS regression and neurotransmitter maps using JuSpace. Results Both smoking (β = −0.088, 95% CI: −0.109 to −0.068) and alcohol consumption frequency (β = −0.037, 95% CI: −0.038 to −0.019) were associated with lower global MSNs values. Ever smoking was associated with 16 regional MSNs alterations, whereas alcohol consumption frequency was associated with 80 regional alterations. These findings were robust across 1,000 bootstrap iterations. JuSpace analyses showed significant alcohol-related associations with opioid, dopaminergic, serotonergic, and cholinergic maps, particularly MU and MOR opioid receptor maps, while smoking showed no significant associations after FDR correction. PLS analyses identified significant transcriptomic associations for both smoking- and alcohol-related MSN8 alterations (smoking: r = 0.450, p_perm = 0.009; alcohol: r = 0.604, p_perm < 0.001). Enrichment analyses were mainly driven by negatively associated genes for both exposures, whereas positively associated genes showed no significant enrichment. Smoking-related negative genes implicated cholesterol biosynthesis, postsynaptic neurotransmitter signaling, NMDA receptor-related processes, and synaptic/dendritic components, while alcohol-related negative genes additionally implicated focal adhesion, actomyosin organization, and myofibril/muscle contraction-related pathways. Discussion Smoking and alcohol use were both associated with lower global MSNs, suggesting shared reductions in cortical morphometric similarity. However, smoking showed more regionally selective cortical alterations, whereas alcohol use showed broader regional effects and stronger neurotransmitter-related spatial associations, particularly with opioid-related systems. Transcriptomic analyses suggested that both exposures may involve negatively associated genes related to synaptic signaling, NMDA receptor-related processes, and cholesterol biosynthesis, with alcohol additionally implicating focal adhesion and actomyosin/myofibril-related pathways. These findings highlight both shared and exposure-specific neurobiological mechanisms linking smoking and alcohol use to cortical MSNs alterations.