Waveform reliability report: CMC, ICC / SEM / MDC curves

The integrated movement-waveform reliability summary. It combines the within-subject repeatability (per-subject Kadaba waveformCMC()) with the between-subject pointwise ICC curve and the frame-by-frame standard error of measurement and minimal detectable change (from sem() / mdc(), the same definitions as the scalar clinimetrics).

Usage

waveformReliability(x, model = "twoway", type = "agreement", confidence = 0.95)

Arguments

x: A [subject, trial, frame] array, or a list of trials x frames matrices (one per subject).
model, type: Passed to icc() for the pointwise ICC.
confidence: Confidence level for the MDC (default 0.95).

Value

a waveform_reliability object with per-subject cmc, the icc, sem, mdc curves (+ ICC CIs), and scalar summaries (mean_cmc, mean_icc, peak_mdc, ...). The SEM curve is the between-subject SD at each frame times sqrt(1 - ICC); the MDC curve is the 95% minimal detectable change in the measurement's own units.

References

Kadaba MP, et al. (1989) J Orthop Res 7:849-860; Weir JP (2005) J Strength Cond Res 19:231-240 (SEM/MDC).

Examples

set.seed(1)
subs <- lapply(1:10, function(s) {
  base <- s + 5 * sin(seq(0, 2 * pi, length.out = 101))
  t(sapply(1:3, function(k) base + rnorm(101, 0, 0.3)))
})
r <- waveformReliability(subs)
r$mean_cmc; r$mean_icc; r$peak_mdc
#> [1] 0.9961583
#> [1] 0.9896484
#> [1] 1.139571