Fig. 4

Overview of seven standardized motion capture trials, visualizing the difference between the manual FOG segmentation by the clinician and the automated FOG segmentation by the MS-GCN. The x-axis denotes the number of samples (at a sampling frequency of 100 hz). The color gradient visualizes the overlap or discrepancy between the model and experts’ annotations. The model annotations were derived from the test set, i.e., subjects that the model had never seen