tDCS shows craving reduction in male methadone patients; machine learning algorithm achieves 94% diagnostic accuracy

This randomized controlled trial enrolled 36 male patients undergoing methadone maintenance treatment for opioid addiction, along with 24 healthy controls. Patients were assigned to one of three tDCS protocols: Group A (left anodal/right cathodal), Group B (right anodal/left cathodal), or Group C (sham stimulation). The study examined both diagnostic classification using EEG-based machine learning and craving outcomes following tDCS.

The machine learning algorithm achieved 94.30% diagnostic accuracy in distinguishing patients with opioid addiction from healthy controls. Regarding craving, both active tDCS groups (Groups A and B) showed significant reductions compared to the sham group, though specific effect sizes, absolute numbers, and statistical details were not reported in the abstract.

Safety and tolerability data were not reported. Key limitations include the small, male-only sample of methadone patients, which limits generalizability to broader opioid-addicted populations including women. The absence of reported effect sizes, confidence intervals, and detailed statistical methods in the abstract requires cautious interpretation of the craving reduction findings.

For clinical practice, these results suggest tDCS may warrant further investigation as an adjunctive intervention for craving in opioid addiction, but the evidence remains preliminary. The high diagnostic accuracy of the EEG-based algorithm is notable but requires validation in larger, more diverse populations before clinical application could be considered.