Predicting Treatment Response in Female Adolescents With Non-Suicidal Self-Injury Using Neurophysiological Biomarkers and Machine Learning

Comprehensive Summary

This study looked at whether quantitative electroencephalography (qEEG) features, combined with clinical data, could predict treatment outcomes in female adolescents with non-suicidal self-injury (NSSI). Through the analysis of clinical and EEG data from 104 female adolescent inpatients with repetitive NSSI, clinical outcomes were assessed using pre- and post-admission scores on the Health of the Nation Outcome Scales, Clinical Global Impression-Severity, World Health Organization Disability Assessment Schedule, and Global Assessment of Functioning, and performance was interpreted through the SHapley Additive exPlanations analysis. All of the models showed substantial predictive performance, with the key qEEG predictors being relative low-beta power at Pz, absolute theta power at Fp1, and the delta-to-beta ratio at Cz. Pre-admission clinical severity largely contributed to prediction accuracy. The qEEG features, combined with machine learning, can effectively predict treatment response in adolescents with NSSI, which supports the idea that their use can demonstrate neurophysiological biomarkers for individualized care.

Outcomes and Implications

These results demonstrate that, along with baseline clinical scores, specific qEEG features do contribute to treatment outcome prediction. These may reflect neurophysiological dysfunctions relevant to NSSI, also associated with impaired cognitive function and emotions regulation. Given their non-invasive and scalable nature of these methods, EEG biomarkers prove to be promising for early intervention and longitudinal monitoring in clinical settings.