Table 2 Stepwise backward selection regression model to predict changes in hypokinetic and hyperkinetic symptoms of PD, ordered by relative importance
From: Digital speech biomarkers can measure acute effects of levodopa in Parkinson’s disease
Estimate | p value | |
|---|---|---|
HYPOKINETIC FEATURES MODEL (dependent variable: MDS-UPDRS-III (no tremor) change) | ||
INTERCEPT | − 16.84165 | <0.001 *** |
sdF0 (text) change | −11.80804 | <0.001 *** |
Int kurt (text) change | 7.84045 | <0.01 ** |
RST (text) change | −0.11860 | <0.001 *** |
LTAS skew (text) change | 1.78596 | <0.05 * |
Int sd (text) change | −4.26142 | <0.05 * |
NSR (text) change | 3.27815 | 0.0573 |
LTAS mean (text) change | 0.04163 | 0.1914 |
VOT (ddk) change | −0.32521 | 0.1679 |
Residual std. error: 7.118 (42 df); Multiple R2: 0.60; Adj. R2: 0.52; F-statistic: 7.9; p-value: < 0.001 | ||
HYPERKINETIC FEATURES MODEL (dependent variable: Marconi (axial subscore) change) | ||
INTERCEPT | 2.479927 | <0.001 *** |
MPT (text) change | 0.189965 | <0.05 * |
Int sd (text) change | 0.685253 | 0.0662 |
Int sd (phon) change | -0.627082 | 0.0702 |
LTAS sd (text) change | 0.007723 | <0.05 * |
LTAS skew (text) change | 0.992226 | <0.05 * |
LTAS kurt (text) change | -0.009829 | 0.1262 |
Residual std. error: 2.078 (44 df); Multiple R2: 0.37; Adj. R2: 0.29; F-statistic: 4.3; p-value: < 0.01 | ||
