Skip to main content
. 2021 Jul 14;12:669076. doi: 10.3389/fneur.2021.669076

Figure 2.

Figure 2

Frequently selected features. The frequency of selection for each cortical feature used to train the SVM model using BC+CC+DC and proposed Enet-subset feature selection method was plotted onto a cortical mesh surface. The top 60 features were selected in all 100 iterations and sorted according to the frequency of its selection during the 100 iterations. Cortical regions outlined in green are bilateral while those outlined in black are unilateral.