Figure 1.

Flow-chart of the study. We obtained clinical data, MRI and MEP metrics from the test cohort. The test cohort was followed for two years, collecting clinical information (disability and relapses). Single variables were tested for predicting disease activity (new relapses or increase in the disability scales EDSS or MSFC) outcomes and predictive models were developed using computational classifiers, after performing an attribute selection of the most informative variables. The different classifiers were tested in the test cohort using a 10-fold cross-validation. From the different variables and classifiers, the NNet using EDSS at baseline and CMCT for predicting the EDSS range two years later was selected for further development because of its high performance. Validation was carried out in a second prospective cohort for whom EDSS at baseline and two years later and CMCT were available. Finally, we calculated the diagnostic accuracy of the NNet using the 10-fold cross-validation method in the overall population (test and validation cohorts).