Sensitivity and specificity from the ROC analysis.

Note- EZ-MAP, “one vs. many” T, and Z-scores were thresholded at α1 = 0.05 (2-tails); α2 = 0.05 (corrected for multiple comparisons). In the ROC analysis, the optimal cut-off number (all abnormal voxels for discrimination of patients and controls) is the point closest to the top-left corner of the ROC curve. Sensitivity and specificity are not applicable because the FWER-control method was not significantly powered in the ROC analysis.