braindecode.visualization.layer_grad_cam

braindecode.visualization.layer_grad_cam(model, x, target, layer)

Layer GradCAM via captum: class-discriminative localization at layer.

Captum’s captum.attr.LayerGradCam returns the cam at the layer’s spatial dims; we then bilinear-resample to the input’s last two dimensions via captum.attr.LayerAttribution.interpolate().