Skip to contents

Detect non-wear and clipping time in the raw accelerometer data

detect_nonwear_clipping.Rd

Detects periods in which the accelerometer has not been worn or the accelerometer signal was stuck very close to the dynamic range limit of the accelerometer.

Usage

detect_nonwear_clipping(data = c(), windowsizes = c(5, 900, 3600), sf = 100,
                        clipthres = 7.5, sdcriter = 0.013, racriter = 0.05,
                        nonwear_approach = "2013",
                        params_rawdata = c())

Arguments

data

Matrix with raw accelerometer data for X, Y, and Z axes.

windowsizes

Numeric vector of length three, with short, long epoch and window size in seconds.

sf

Sample frequency in Hertz.

clipthres

Threschold to detect clipping in _g_ units. Usually 0.5 _g_ below the dynamic range of the accelerometer.

sdcriter

Criteria to define non-wear time, defined as the estimated noise measured in the raw accelerometer data.

racriter

Absolute criteria below which the absolute range of the accelerations should be to define non-wear time.

nonwear_approach

Whether to use the traditional version of the non-wear detection algorithm (nonwear_approach = "2013", default) or the new version (nonwear_approach = "2023"). The 2013 version would use the longsize window (windowsizes[3], one hour as default) to check the conditions for nonwear identification and would flag as nonwear the mediumsize window (windowsizes[2], 15 min as default) in the middle. The 2023 version differs in which it would flag as nonwear the full longsize window. For the 2013 method the longsize window is centered in the centre of the mediumsize window, while in the 2023 method the longsizewindow is aligned with its left edge to the left edge of the mediumsize window.

params_rawdata

See details in GGIR.

Value

List containing the next numeric vectors:

  • NWav (non-wear score), from 1 to 3 indicating the number of axes that met the non wear criteria.

  • CWav (clipping score), binary, 0-1 indicating non-clipping and clipping, respectively.

  • nmin is the minimum numebr of windows in this block of data.

Each number in these vectors represent a long epoch duration (i.e., ws2, 900 seconds by default).

Author

Vincent T van Hees <v.vanhees@accelting.com> Jairo Hidalgo Migueles <j.h.migueles@accelting.com>

References

  • van Hees et al. 2011, doi: 10.1371/journal.pone.0022922.

  • van Hees et al. 2013, doi: 10.1371/journal.pone.0061691 (in supplementary material).

Examples

if (FALSE) {
    detect_nonwear_clipping(data = data, windowsizes = c(900, 3600), sf = sf,
                            clipthres = clipthres, sdcriter = sdcriter, 
                            racriter = racriter, nonwear_approach = "old")
  }