Brain connectivity pattern predicts treatment response in cocaine use disorder
Researchers reporting in Nature Mental Health have identified a resting-state brain connectivity signature that both distinguishes cocaine use disorder from healthy controls and helps forecast who is likely to benefit from repetitive transcranial magnetic stimulation (rTMS), a non-invasive magnetic therapy aimed at reducing cravings. Psychosocial care still leaves many people cycling through compulsive use and relapse, so clinicians have sought objective biological markers rather than relying only on interviews and self-report.
A team led by bioengineers Kanhao Zhao and Yu Zhang of Lehigh University trained a machine-learning model on resting-state functional MRI from a discovery cohort of 129 people in Mexico—71 with cocaine use disorder and 58 healthy controls—who lay still in the scanner with no task. The model separated the groups with 83 percent accuracy. Patients showed stronger coupling between visual and attention networks and between circuits for goal-directed behavior and sensory attention, and weaker coupling between the default mode network (internal thought) and the limbic network (emotion).
Applied to an independent replication set of 82 patients and 81 controls scanned in the United States and Mexico, the same signature classified people at 70 percent accuracy, indicating a generalizable feature of the disorder rather than a local artifact. The researchers then tested whether that diagnostic pattern could predict rTMS outcomes in a smaller study of 45 patients. Clinical trials already suggest rTMS can lower cocaine cravings for some people; linking those responses to a measurable connectivity marker is meant to support more targeted diagnosis and treatment selection.
Continue in the app — vote & join in ➔