Grey matter morphology as a predictor of individual response to cognitive behavioural therapy in depression - findings from a preliminary machine learning study.

Major depression is a disabling illness, with many treatment options, but no reliable predictors of treatment response. In the present study, we utilised a longitudinal design to test if pre-treatment grey-matter morphology could predict individual response to computerized guided-self-help cognitive-behavioural therapy for major depression. Using an advanced machine learning technique, in a sample of 43 un-medicated patients, we found that grey matter morphology predicted response to cognitive-behavioural therapy with 73.92% (p=0.001) accuracy and a proportionate reduction in uncertainty of 82%. We provide preliminary evidence that pre-treatment grey matter morphology is a biomarker of response to cognitive-behavioural therapy.