Fig. 4.

Rapid-eye movement sleep behaviour disorder (RBD) detection performance using automated sleep staging. For this figure B1, C2, and A3 represent automated sleep staging using electrooculogram (EOG), both EOG and electromyogram (EMG), and electrocardiogram (ECG), EOG, and EMG features, respectively. While D1, E1, and D2 represent metrics derived from EMG, ECG, and both EMG and ECG sensors for RBD detection, used in combination with automated sleep staging. The Z3D1 combination represents the results from our previous study that included electroencephalogram (EEG), EOG, and EMG features for automated sleep staging and EMG metrics for RBD detection. From this figure we can observe how non-EEG sensors (C2E1, C2D2, and A3D2) can achieve fully automated RBD detection similar to EEG sensors (Z3D1).