Fig. 4: The saliency maps of the discriminative biomarkers learned by the GCNs models for the FAF task. | npj Digital Medicine

Fig. 4: The saliency maps of the discriminative biomarkers learned by the GCNs models for the FAF task.

From: An interpretable model based on graph learning for diagnosis of Parkinson’s disease with voice-related EEG

Fig. 4

The numbers ranging from 1 to 64 correspond to the electrodes according to the international EEG 10-20 system. Panels a–d show the distribution of the contribution network for the PCC+GCNs model, the PCC + GSP-GCNs model, the PLV+GCNs model, and the PLV + GSP-GCNs model, respectively. Nodes with a redder hue indicate greater contributions to the model, and edges with a blacker hue indicate higher contributions.

Back to article page