Predictive performance achieved for each MDS-UPDRS time point and each imaging feature type, computed through leave-one-out cross-validation using the default workflow (“Repli.”).

Similar Items

• Predictive features learned by the best performing models to predict MDS-UPDRS score at each time point for the original study (left—extracted from [6]) and the <i>default workflow</i> (right) using fALFF.
  by: Elodie Germani (20765922)
  Published: (2025)
• Predictive features learned by the best performing models to predict MDS-UPDRS score at each time point for the original study (left—extracted from [6]) and the <i>default workflow</i> (right) using ReHo.
  by: Elodie Germani (20765922)
  Published: (2025)
• Distribution of MDS-UPDRS scores reported in the original paper’s cohort (<i>top</i>: S1 Fig extracted from [6]), the replication cohort (<i>middle</i>) and the closest-to-original cohort (<i>bottom</i>).
  by: Elodie Germani (20765922)
  Published: (2025)
• Performance of models trained for prediction at each time point, using fALFF or ReHo, with variations in the workflow.
  by: Elodie Germani (20765922)
  Published: (2025)
• Performance reported using different model selection and evaluation methods.
  by: Elodie Germani (20765922)
  Published: (2025)