Analysis Of Multiple Sclerosis Data

After testing the NPC on the toy problems showed above, we analysed a quantitative MRI data set from 43 multiple sclerosis patients. The results showed to a reasonable splitting of the data into two subgroups according to the individual patients’ disease grade. The first group consist of 21 subjects with an average EDSS of 0.64±1.5 and an average age of 38±8 years. The second group consists of 15 points classified as noise. It contains primarily older patients with a mean age of 46±12 years and a higher EDSS score of 2.78±2.3.

The very good performance of the algorithm on such highly correlated non-spherical datasets, typical for MRI derived image features, shows that Nuclear Potential Clustering is a valuable tool for automated data and image analysis, not only in the MRI domain.