Predicting Future Alcohol Use via Connectome-Behavior Mapping
Poster No:
1097
Submission Type:
Abstract Submission
Authors:
James Kim1, Qingyu Zhao2, Mert Sabuncu3, Amy Kuceyeski4
Institutions:
1Cornell University, Ithaca, NY, 2Weill Cornell Medicine, New York, NY, 3Cornell Tech, Weill Cornell Medicine, New York, NY, 4Weill Cornell Medicine, Ithaca, NY
First Author:
Co-Author(s):
Introduction:
Alcohol use remains a significant social issue, often starting in adolescence. A 2023 survey reported alarming rates among adolescents, with 20.1% of 8th graders and 52.8% of high school seniors having consumed alcohol (Miech et al., 2024). These statistics underscore the need for predictive modeling to identify individuals vulnerable to heavy alcohol use for early intervention strategies.
While heavy alcohol use affects both brain function and structure (Yang et al., 2023; Park et al., 2023), limited research exists on how structural connectivity (SC) predicts heavy drinking before its onset, with most studies focusing on functional connectivity's (FC) link to alcohol abuse (McKenna et al., 2021). This study explored neurological vulnerability to heavy alcohol use by combining both SC and FC measures.
While heavy alcohol use affects both brain function and structure (Yang et al., 2023; Park et al., 2023), limited research exists on how structural connectivity (SC) predicts heavy drinking before its onset, with most studies focusing on functional connectivity's (FC) link to alcohol abuse (McKenna et al., 2021). This study explored neurological vulnerability to heavy alcohol use by combining both SC and FC measures.
Methods:
Based on the 7-year longitudinal data from the National Consortium on Alcohol and Neurodevelopment in Adolescence (NCANDA), we analyzed 653 subjects (315 males, 338 females, aged 12–21 years at baseline) who were no-to-low drinkers (Youth-Adjusted Cahalan=0) at their baseline visits (Zhao et al., 2020). We predicted who would engage in heavy drinking after baseline (Cahalan>1). This classification was implemented using six logistic regression models trained on different data types: SC, FC, FC with global signal regression, demographics (age, sex, race, baseline alcohol severity, site, and scanner model), an ensemble meta-learner, and a simple averaging ensemble. The model loop included 1,000 iterations with a 5-fold train/test split. Within each training set, an additional 5-fold split determined the optimal L2 regularization parameter for each model. For each data type, we trained and evaluated models on all subjects, males only, and females only.
Results:
The SC (all-subjects) model outperformed the FC model in predicting future heavy alcohol use, achieving a balanced accuracy of 0.585 ± 0.013 and an AUC of 0.634 ± 0.013, compared to FC's 0.523 ± 0.009 and 0.532 ± 0.012, respectively. Demographics alone had an AUC of 0.650 ± 0.006 and adding demographics to SC via ensembling did not enhance model performance. Haufe-transformed weights (Haufe et al., 2013) indicated that stronger SC was associated with having future problematic drinking, whereas stronger FC was negatively associated. Interestingly, female-only models performed better than male-only models, achieving the following balanced accuracies (SC: 0.531 ± 0.022; FC: 0.547 ± 0.018) and AUCs (SC: 0.551 ± 0.024; FC: 0.570 ± 0.020). Male-only models had lower balanced accuracies (SC: 0.503 ± 0.011; FC: 0.506 ± 0.006) and mixed AUC results (SC: 0.569 ± 0.017; FC: 0.494 ± 0.022).
Key regions associated with SC across all sexes included the inferior temporal gyrus, superior temporal gyrus, and insula, while FC was linked to the caudate nucleus, angular gyrus, putamen, and middle temporal gyrus. Network-level importance analysis identified the ventral attention network as most influential for SC and the default mode networks for FC.
Key regions associated with SC across all sexes included the inferior temporal gyrus, superior temporal gyrus, and insula, while FC was linked to the caudate nucleus, angular gyrus, putamen, and middle temporal gyrus. Network-level importance analysis identified the ventral attention network as most influential for SC and the default mode networks for FC.
·SC and FC (all-subjects) model coefficients corresponding to each brain region (AAL atlas).
·SC and FC (all-subjects) model coefficients corresponding to each Yeo network were averaged and visualized to examine network influences on future heavy drinking.
Conclusions:
Adolescent alcohol use is shaped by a complex interplay of social, psychological, and neurobiological factors. Our findings suggest that white matter connections (SC) may be crucial to problematic alcohol use, perhaps more so than brain co-activity patterns (FC). While the predictive value at an individual level is limited, SC models exceeded random guessing in predicting future alcohol consumption patterns and were more accurate for females than males.
A previous study linked changes in cortical thickness and surface area in regions such as the insula, frontal pole, and precuneus to predispositions for alcohol use (Baranger et al., 2023). This aligns with our results, highlighting that both cortical changes and white matter connections contribute to predisposition for heavy drinking. By pinpointing regional and network-level influences on future problematic drinking, this work advances the identification of potential targets for early clinical intervention.
A previous study linked changes in cortical thickness and surface area in regions such as the insula, frontal pole, and precuneus to predispositions for alcohol use (Baranger et al., 2023). This aligns with our results, highlighting that both cortical changes and white matter connections contribute to predisposition for heavy drinking. By pinpointing regional and network-level influences on future problematic drinking, this work advances the identification of potential targets for early clinical intervention.
Disorders of the Nervous System:
Psychiatric (eg. Depression, Anxiety, Schizophrenia) 2
Lifespan Development:
Early life, Adolescence, Aging
Modeling and Analysis Methods:
Classification and Predictive Modeling 1
Connectivity (eg. functional, effective, structural)
fMRI Connectivity and Network Modeling
Keywords:
Addictions
Data analysis
Development
FUNCTIONAL MRI
Machine Learning
Modeling
Psychiatric Disorders
Statistical Methods
STRUCTURAL MRI
White Matter
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Are you Internal Review Board (IRB) certified? Please note: Failure to have IRB, if applicable will lead to automatic rejection of abstract.
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Provide references using APA citation style.
Haufe, S., Meinecke, F., Görgen, K., Dähne, S., Haynes, J.-D., Blankertz, B., & Bießmann, F. (2013, November 15). On the interpretation of weight vectors of linear models in multivariate neuroimaging. NeuroImage. https://www.sciencedirect.com/science/article/pii/S1053811913010914
McKenna, B. S., Anthenelli, R. M., Smith, T. L., & Schuckit, M. A. (2021, November 11). Low versus high level of response to alcohol affects amygdala functional connectivity during processing of emotional stimuli. Alcoholism, clinical and experimental research. https://pubmed.ncbi.nlm.nih.gov/35064942/
Miech, R. A., Johnston, L. D., Patrick, M. E., & O’Malley, P. M. (2024, May). National Survey Results on Drug Use, 1975-2023: Overview and Detailed Results for Secondary School Students. Monitoring the Future Monograph Series. https://monitoringthefuture.org/wp-content/uploads/2024/01/mtfoverview2024.pdf
Park, S.-E., Jeon, Y.-J., & Baek, H.-M. (2023, June 12). Functional and structural brain abnormalities and clinical characteristics of male patients with alcohol dependence. Brain sciences. https://pmc.ncbi.nlm.nih.gov/articles/PMC10296436/
Yang, W., Han, J., Luo, J., Tang, F., Fan, L., Du, Y., Yang, L., Zhang, J., Zhang, H., & Liu, J. (2023, December 30). Connectome-based predictive modelling can predict follow-up craving after abstinence in individuals with opioid use disorders. General Psychiatry. https://doi.org/10.1136/gpsych-2023-101304
Zhao, Q., Sullivan, E. V., & Honnorat, N. (2020). Association of heavy drinking with deviant fiber tract development in Frontal Brain Systems in adolescents. JAMA psychiatry. https://pubmed.ncbi.nlm.nih.gov/33377940/