INHIBITORY CONTROL MEG EVENT-RELATED FIELDS AS PREDICTIVE BIOMARKERS FOR BINGE DRINKING IN ADOLESCENTS

Adolescent binge drinking (BD) is a major public health concern due to its high prevalence and neurodevelopmental consequences. As little is known about neural markers preceding alcohol initiation, this longitudinal study investigated whether baseline magnetoencephalography (MEG) event-related fields (ERFs) recorded during inhibitory control predict later BD. Eighty-one alcohol naïve adolescents initially completed a Go/No-Go task during MEG acquisition and completed self-report measures of impulsivity, sensation seeking and executive functioning. After two years, 44 participants remained eligible and were classified as controls (CN, n=20) or binge drinkers (BD, n=24) based on reported habits of alcohol consumption. Behavioral analyses showed no group differences in accuracy in the inhibition task, impulsivity (BIS-11), or executive functioning (BRIEF-SR), although the BD group reported higher sensation seeking (SSS-V). Cluster-based permutation analyses identified significant group differences in both the M200 (180–260ms) and M300 (310–510ms) components. In both cases, future BD adolescents exhibited larger amplitudes, with sensors localized to left medial and dorsolateral prefrontal areas. These differences were moderate predictors in logistic regression models, with M200 activity achieving 77.3% accuracy and M300 activity reaching 70.5% accuracy. Association between ERF and future alcohol use were not influenced by biological sex. The results converge with prior evidence of left prefrontal hyperactivation in adolescent BD and developmental studies showing enhanced recruitment of control networks during adolescence. Findings support the hypothesis that atypical prefrontal executive engagement may represent a vulnerability profile that precedes alcohol use and may contribute to the emergence of BD.

Classification performance of logistic regression models predicting alcohol use based on ERF activity (M200 and M300 clusters)
Model	ACC (95% CI)	SEN	SPE	PPV	NPV​​​​​
M200 (original)	77.3 (62.2–88.5)	80.0	75.0	72.7	81.8
M200 (cross-validation)	72.7 (57.2–85.0)	75.0	70.8	68.2	77.3
M300 (original)	70.5 (54.8–83.2)	70.0	70.8	66.7	73.9
M300 (cross-validation)	68.2 (52.4–81.4)	70.0	66.7	63.6	72.7
Note. ACC = accuracy; SEN = sensitivity; SPE = specificity; PPV = positive predictive value; NPV = negative predictive value. Accuracy values are reported with their 95% confidence intervals (CI).