Posturographic center-of-pressure sway metrics (Prieto 1996)

Computes the full Prieto et al. (1996) battery of postural-steadiness measures from a center-of-pressure (CoP) time series, together with the non-linear complexity descriptors used by schema_balance. Distances are reported in the units of the supplied CoP (typically mm or m); frequencies in Hz; areas in CoP-units squared; velocities in CoP-units per second.

Usage

swayMetrics(
  cop,
  sampling_rate,
  ap = NULL,
  ml = NULL,
  detrend = c("mean", "none"),
  entropy = TRUE,
  m = 2L,
  r = 0.2,
  base_of_support = NULL
)

Arguments

cop: Center-of-pressure data. A data.frame from calculateCOP() (columns cop_x = ML, cop_y = AP), a two-column matrix/data.frame (columns interpreted as ml, ap unless named), or NULL when supplying ap/ml directly.
sampling_rate: Sampling rate in Hz.
ap, ml: Optional anteroposterior / mediolateral CoP vectors (override cop).
detrend: Either "mean" (subtract the mean CoP, the Prieto convention) or "none".
entropy: Logical; compute per-axis sampleEntropy() and DFA-$\alpha$ (default TRUE).
m, r: Sample-entropy embedding dimension and tolerance (tolerance is a fraction of each axis SD).
base_of_support: Optional numeric c(ap = , ml = ) giving the half extent of the base of support (in CoP units) about the mean CoP, used for time-to-boundary. NULL (default) returns NA for the TTB measures.

Value

A sway_metrics object (a named list) with the Prieto measures and, in $metrics, a one-row data.frame aligned to the schema_balance metric names.

Details

The measures are defined relative to the mean CoP. With $RD_n$ the resultant distance of sample $n$ from the mean CoP and $T$ the trial duration:

Path length (total excursion, TOTEX): the summed point-to-point distance; mean velocity is TOTEX / T.
RMS distance (RDIST) and mean distance (MDIST), resultant and per-axis.
95% confidence circle area (AREA-CC) $= \pi (MDIST + z\, s_{RD})^2$ with $z = 1.645$.
95% confidence ellipse area (AREA-CE) $= 2\pi F_{0.05[2,n-2]} \sqrt{s_{AP}^2 s_{ML}^2 - s_{AP,ML}^2}$.
Sway area (AREA-SW), area swept per unit time.
Mean frequency (MFREQ), the equivalent rotational frequency $MVELO / (2\pi\, MDIST)$.
Fractal dimension (FD-CC) and frequency-domain measures (total power, 50%/95% power frequency, centroidal frequency, frequency dispersion).
Non-linear sample entropy and DFA-$\alpha$ per axis, and optional time-to-boundary if a base of support is supplied.

References

Prieto TE, Myklebust JB, Hoffmann RG, Lovett EG, Myklebust BM (1996). "Measures of postural steadiness." IEEE Trans Biomed Eng 43(9):956-966.

Examples

t <- seq(0, 30, by = 0.01)
cop <- data.frame(cop_x = 2 * sin(2 * pi * 0.2 * t),
                  cop_y = 3 * cos(2 * pi * 0.2 * t))
sm <- swayMetrics(cop, sampling_rate = 100)
sm$path_length
#> [1] 95.19201
sm$mean_velocity
#> [1] 3.173067