Machine Learning-Driven fMRI Analysis for Objective Craving Prediction

Abstract

BackgroundCraving is a fundamental aspect of substance use disorder (SUD), traditionally assessed through subjective self-report measures. To develop more objective assessments, we created a brain-based marker to predict craving based on machine learning approaches using functional magnetic resonance imaging (fMRI) drug cue reactivity data from 69 participants with methamphetamine use disorders. MethodsTo predict craving intensity, our analysis demonstrated that utilizing principal component analysis (PCA) and linear regression outperformed other models in terms of Root Mean Squared Error (RMSE). Employing a 5-fold cross-validation strategy with a 20% holdout set, we established the reliability of the model. Additionally, the model successfully classified high and low craving levels and distinguished cue types (neutral vs. drug) based on fMRI data. ResultsThe model achieved an RMSE of 0.983 {+/-} 0.026 (standard deviation), with strong generalization evidenced by an out-of-sample RMSE of 0.985 and statistical significance (p < 0.026; effect size (Cliffs Delta) = 0.715; statistical power = 0.639). Key neurobiological signatures included the Parahippocampal Gyrus, Superior Temporal Gyrus, Medioventral Occipital Cortex, and Amygdala (positively associated with craving), as well as the Inferior Temporal Gyrus (negatively associated). Classification of high vs. low craving levels yielded an AUC-ROC of 0.684 {+/-} 0.084 (out-of-sample AUC-ROC = 0.714), with significant separation (p < 0.04; Cliffs Delta = 0.831). Additionally, classification of cue types (neutral vs. drug) achieved an AUC-ROC of 0.692 {+/-} 0.090 (out-of-sample AUC-ROC = 0.693), with p < 0.002, Cliffs Delta = 0.896, and statistical power = 0.800, highlighting the robustness of the model. ConclusionThese findings underscore the potential of neuroimaging and machine learning to provide objective, data-driven insights into the neural mechanisms underlying subjective experience of craving and to inform future clinical applications in SUD.