Fig. 7

The line plots for clusters that were discovered by applying K-Means over (left) \(t_{\rm pre }\) and (right) \(t_{\rm post }\). The outliers were excluded from the data. The depicted clusters are discovered over the complete TBI data rather than bootstraped samples (they are used only for individual runs). The bold lines represent the centroids of the clusters, while thin dotted lines depict the patients. The colours show which clusters from \(t_{\rm pre }\) and \(t_{\rm post }\) are related to each other. (Color figure online)