Skip to main content
. 2024 Jun 9;14:13229. doi: 10.1038/s41598-024-63946-4

Table 3.

Estimation of clinical characteristics using gait spatio-temporal parameters.

Task Clinical characteristic Accuracy (SD across folds) [range across folds] Sensitivity (SD across folds) [range across folds] Specificity (SD across folds) [range across folds] F1-score (SD across folds) [range across folds]
Gait FoG (Binary) 0.98 (0.16) [0.33–1.00] 0.98 (0.22) [0.00–1.00] 0.98 (0.11) [0.50–1.00] 0.98 (0.22) [0.00–1.00]
MDS-UPDRS FoG Score 0.90 (0.12) [0.60–1.00] 0.84 (0.30) [0.00–1.00] 0.97 (0.08) [0.75–1.00] 0.87 (0.30) [0.00–1.00]
MDS-UPDRS Gait Score 0.92 (0.13) [0.50–1.00] 0.93 (0.26) [0.00–1.00] 0.97 (0.09) [0.67–1.00] 0.94 (0.26) [0.00–1.00]
Presence of Dystonia 0.88 (0.25) [0.33–1.00] 0.90 (0.38) [0.00–1.00] 0.93 (0.19) [0.50–1.00] 0.90 (0.38) [0.00–1.00]

Classification performance (accuracy, sensitivity, specificity, and F-1 score) achieved by the ML-based estimation algorithms using gait spatio-temporal parameters. Prediction of presence/absence of freezing of gait (FoG) on clinical examination, MDS-UPDRS FoG score (item 3.11), MDS-UPDRS gait score (item 3.10), and presence/absence of dystonia on clinical examination are shown.

SD, standard deviation.