Table 2 Estimation of MDS-UPDRS scores using sensor-based data features and the presence of dystonia.

From: Motor assessment of X-linked dystonia parkinsonism via machine-learning-based analysis of wearable sensor data

 

Task

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]

MDS-UPDRS

Finger-to-nose (Item 3.16)

0.75 (0.18) [0.55–1.00]

0.79 (0.28) [0.31–1.00]

0.85 (0.13) [0.66–1.00]

0.78 (0.27) [0.31–1.00]

Hand pronation/supination (Item 3.6)

0.63 (0.17) [0.50–1.00]

0.56 (0.17) [0.00–1.00]

0.88 (0.12) [0.67–1.00]

0.56 (0.35) [0.00–1.00]

Leg agility (Item 3.8)

0.81 (0.18) [0.38–1.00]

0.65 (0.28) [0.07–1.00]

0.85 (0.14) [0.53–1.00]

0.65 (0.28) [0.07– 1.00]

Toe tapping (Item 3.7)

0.64 (0.19) [0.50–1.00]

0.67 (0.37) [0.00–1.00]

0.88 (0.12) [0.67–1.00]

0.66 (0.37) [0.00–1.00]

Gait (Item 3.10)

0.78 (0.12) [0.55–1.00]

0.79 (0.23) [0.11–1.00]

0.92 (0.08) [0.70–1.00]

0.79 (0.23) [0.11–1.00]

DYSTONIA

Finger-to-nose (Item 3.16)

0.91 (0.09) [0.72–1.00]

0.83 (0.14) [0.58–1.00]

0.91 (0.07) [0.79–1.00]

0.85 (0.15) [0.58–1.00]

Hand pronation/supination (Item 3.6)

0.94 (0.14) [0.50–1.00]

0.91 (0.21) [0.25–1.00]

0.94 (0.11) [0.62–1.00]

0.92 (0.21) [0.25–1.00]

Leg agility (Item 3.8)

0.74 (0.25) [0.33–1.00]

0.74 (0.38) [0.00–1.00]

0.85 (0.19) [0.50–1.00]

0.73 (0.38) [0.00–1.00]

Toe-tapping (Item 3.7)

0.80 (0.24) [0.37–1.00]

0.80 (0.37) [0.05–1.00]

0.87 (0.18) [0.52–1.00]

0.80 (0.37) [0.00–1.00]

Heel-toe alternate movement (Dystonia-provoking)

0.85 (0.21) [0.41–1.00]

0.87 (0.32) [0.11–1.00]

0.92 (0.16) [0.56–1.00]

0.87 (0.32) [0.11–1.00]

Heel walking (Dystonia-provoking)

0.76 (0.21) [0.33–1.00]

0.77 (0.31) [0.00–1.00]

0.86 (0.16) [0.50–1.00]

0.79 (0.31) [0.00–1.00]

Toe walking (Dystonia-provoking)

0.60 (0.21) [0.54 -1.00]

0.58 (0.38) [0.00–1.00]

0.76 (0.12) [0.66–1.00]

0.58 (0.38) [0.00–1.00]

Straight walking (Item 3.10)

0.66 (0.14) [0.59–1.00]

0.73 (0.21) [0.39–1.00]

0.80 (0.10) [0.69–1.00]

0.71 (0.20) [0.39–1.00]

  1. Classification performance (accuracy, sensitivity, specificity, and F-1 score) achieved by the ML-based estimation algorithms for the estimation of MDS-UPDRS scores using sensor-based data features (upper section) with corresponding MDS-UPDRS Item number, and the presence/absence of dystonia (lower section) clinical scores in each of the considered motor tasks. MDS-UPDRS item numbers are provided in parenthesis. Additional dystonia-provoking tasks were also used for the dystonia analysis.
  2. SD, standard deviation.
Back to article page