Table 3 Results from all analyses.

From: Early detection and tracking of bulbar changes in ALS via frequent and remote speech analysis

Analysis

Estimate (standard error)

p-valuec

Mixed-effects models for comparisons between healthy controls and ALS with normal bulbar function

 Articulatory precision

−0.62 (0.29)a

0.0236

 Speaking rate

−0.77 (0.28)a

0.009

Mixed-effects models for comparisons between healthy controls and all ALS

 Articulatory precision

−0.82 (0.28)a

0.005

 Speaking rate

−1.01 (0.25)a

<0.001

Growth curve model for articulatory precision

 Fixed-effects parameters

 Intercept

9.94 (0.37)

 

 Slope for nonbulbar-onset participants (change per month)

−0.03 (0.02)b

0.025

 Difference in slope between bulbar and non-bulbar -onset participants

−0.12 (0.03)b

<0.001

 Random effects parameters

 Intercept standard deviation

2.54

 

 Slope standard deviation

0.10

 

 Correlation between intercepts and slopes

−0.67

 

 Residual standard deviation

0.33

 

Growth curve model for speaking rate

 Fixed-effects parameters

 Intercept

4.21 (0.25)

 

 Slope for non-bulbar -onset participants (change per month)

−0.003 (0.007)b

0.960

 Difference in slope between bulbar and non-bulbar -onset participants

−0.05 (0.01)b

<0.001

 Random effects parameters

 Intercept standard deviation

1.57

 

 Slope standard deviation

0.04

 

 Correlation between intercepts and slopes

−0.70

 

 Residual standard deviation

0.25

 
  1. aFor group comparison analyses, the metrics were standardized to have mean 0 and standard deviation 1 for easier interpretation of the coefficients16.
  2. bFor growth curve models, the slopes are scaled to the amount of change per month. p-values are provided for slopes; no p-values are available for intercepts or random effects parameters.
  3. cp-values calculated using χ2 likelihood ratio tests.
Back to article page