Table 6.
Accuracy (%) of alternative models. TUG-only and cogTUG-only models used features derived from the first trial, the second trial, and the mean of corresponding features from both trials. TUG-duration and cogTUG-duration models used the durations of TUG and cogTUG tasks. The numbers in parentheses represent the difference in accuracy between the respective model and the corresponding all-task (standard) model (same row).
All Tasks | TUG-Only | cogTUG-Only | TUG-Duration | cogTUG-Duration | |
---|---|---|---|---|---|
All PD vs. controls | 92.6 | 90.1 (−2.5) | 89.4 (−3.2) | 83.3 (−9.3) | 84.0 (−8.6) |
Mild PD (H&Y ≤ 2) vs. controls | 89.4 | 77.9 (−11.7) | 87.2 (−2.4) | 78.7 (−10.7) | 77.0 (−12.4) |