Fig. 7From: Successful classification of cocaine dependence using brain imaging: a generalizable machine learning approachRegions that is used to classify cocaine-dependent and controls participants with the best accuracy. 1500Â voxels in 30 clusters were identified. Figure shows sagittal sections of region-of-interests (ROIs) where 100Â % model, 89Â % LOO, and 88Â % 10xCV accuracies were obtained. Red identifies clusters of increased regional cerebral blood flow (rCBF) in cocaine-dependent participants relative to controls. Blue identifies clusters of decreased (rCBF) in cocaine-dependent participants relative to controls. Slice numbers are in MNI coordinates. MNI coordinates of each cluster and images in axial planes are provided in the Additional file 1 Back to article page