A novel machine learning-based method to quantify the effect of transcranial direct current stimulation on opioid users

Comprehensive Summary

Kazemzadeh et al. examined whether electroencephalography combined with machine learning could be used to objectively evaluate the effects of transcranial direct current stimulation in individuals with opioid use disorder. Do accomplish this, thirty six male patients undergoing methadone maintenance therapy were randomized to receive active or sham tDCS, with EEG recordings collected before and after stimulation and compared to healthy controls. Using targeted EEG channel selection and a support vector machine classifier, the model distinguished opioid users from healthy individuals with 94.3% accuracy. Following active tDCS, EEG patterns shifted toward those observed in healthy controls and were accompanied by reductions in craving and improvements in psychological and biological measures, while no meaningful changes were observed in the control group that received placebo stimulation.

Outcomes and Implications

These findings suggest that EEG based machine learning may serve as a noninvasive, objective tool for assessing both opioid addiction and response to neuromodulation therapies. By providing quantifiable neural markers of craving reduction, this approach could complement traditional clinical assessments that heavily rely on self report or biological testing with known limitations. The results also support the use of tDCS as a potential adjunctive intervention targeting prefrontal circuits involved in craving and executive control. With further validation, this framework has to potential to help improve monitoring of treatment response and guide individualized care strategies for patients with opioid use disorder.