Skip to main content
. 2021 May 27;3(3):fcab115. doi: 10.1093/braincomms/fcab115

Figure 1.

Figure 1

Performance of the classifier. Our SVM ensemble yielded an area under the curve of 0.87 for the prediction of individual therapeutic success (left). That is, on the basis of a whole-brain parcellation of grey matter, the machine learning algorithm predicted individual response to rTMS in tinnitus patients with an accuracy of 83.5% (right).