- Parts labeled as a and c correspond to results before the feature selection process, whereas b and d represent results with the set features selected above. Number of patients assigned to each cluster are shown, and their previous clinical PPA diagnosis is represented by the column labeled PPA. Values 1, 2 and 3 in the first column correspond to nonfluent/agrammatic, semantic and logopenic PPA, respectively, whereas HC represents healthy controls. Kn specifies the number of clusters, where n=0,1,..,N and N is a value between 4 and 8
