J. Lucas McKay 11/2/2021

This document is based on data from the file GaitSignatureQuality.csv provided by Michael Rosenberg, Ph.D. This .csv file is not tracked by github.

We powered the study based on preliminary gait quality data of N = 8 stroke survivors and N = 5 able-bodied comparison subjects. Across groups, based on 1000 Welch Independent Samples t-tests simulated based on preliminary data with a differential effect size of Cohen’s d = 1.3, a sample size of N = 80 stroke survivors and N = 20 able-bodied individuals will have >99% power to resolve a difference between stroke and able-bodied. Within the stroke group, based on contrasts between gait quality measures obtained at enforced speed = 5 and self-selected speed of Cohen’s d = 0.8, a paired-sample t-test will have >95% power to resolve an improvement in gait quality accompanying increased gait speed. Further, based on observed contrasts between enforced speed = 5 and enforced speed = 2 of Cohen’s d = 0.5, a paired-sample t-test will have >85% power to resolve differential improvements in gait quality between enforced speeds 5 and 2. Tests of all pairwise combinations of speeds will have >80% power to resolve at least 5/10 pairwise contrasts between speeds after Bonferroni correction for multiple comparisons (n = 10, alpha = 0.005).

The primary multi-panel data figure is saved as panels.eps. The arguments in the following line can be changed to alter its size and resolution as needed.

ggsave("panels.eps",plt,dpi=300,width=6.5,height=3,units="in")