Figure 3 shows our spatial explainability results and the location of each electrode. Interestingly, the major

/Spatial explainability identifies left hemisphere and central electrodes as most important; channel dropout DA reduces model sensitivity to channel perturbation