McVeighFcn: McVeighFcn

In vancebee/McVeighFcn: Implements the McVeigh algorithm to detect sleeping time in accelerometer data.

Description

McVeighFcn is used to process minute-level accelerometer data from devices worn over 24 hours with the McVeigh algorithm to identify instances of in-bed versus sleep time. The McVeigh specification for determining non-wear time can be used or a non-wear vector calculated via some other methodology can be input.

Usage

McVeighFcn(timevec, cpmvec, nwvec= NULL, rollwin = 30,
slthresh = 88, slperiod = 180, wkperiod = 10,
prslthresh1 = 89, prslthresh2 = 50, prslcount = 4,
prwkthresh1 = 91, prwkthresh2 = 200, prwkcount = 3,
spurthresh = 2.5*60, nwperiod = 90, nwctsgt0 = 3,
nwmaxct = 50,    vhighpct = 0.9, highpct = 0.8,
modpct = 0.5,fairtime = 14*60,  validday = 10*60)

Arguments

timevec	vector detailing the time of each observation. Must be minute-level data. The specified format is <e2><80><9c>POSIX<e2><80><9d>. When timevec is a <e2><80><9c>POSIX<e2><80><9d> variable, an error will be returned if successive measurements are not separated by one minute. If timevec is not a <e2><80><9c>POSIX<e2><80><9d> variable, rows will be assumed to spaced one-minute apart and the calculation of validday and validinbed variables will not be possible.
cpmvec	vector summarizing activity level in cpm at each observation point. Must be same length as timevec.
nwvec	optional vector specifying non-wear observations. If left as NULL, nonwear observations will be calculated via McVeigh<e2><80><99>s algorithm.
rollwin	number of minutes on either side of an observation to be used to calculate moving/rolling average. If the entire size of rollwin is not available (e.g., observation 2, with rollwin = 5), the number number of measurements available are used (e.g, observations 1:8 in the example above).
slthresh	threshold defining low versus high activity.
slperiod	consecutive time (in minutes) of rolling average low activity require to define approximate into-bed times.
wkperiod	consecutive time (in minutes) of rolling average high activity require to define approximate out-of-bed times.
prslthresh1	threshold for individual cpm measurements when calculating precise into-bed times.
prslthresh2	along with prslcount, used to assess 10-minute intervals to calculate precise into-bed times.
prslcount	threshold for number of cpm observations below prslthresh2 when calculating precise into-bed times.
prwkthresh1	threshold for individual cpm measurements when calculating precise out-of-bed times.
prwkthresh2	Along with prwkcount, used to assess 10-minute intervals to calculate precise out-of-bed times.
prwkcount	threshold for number of cpm observations below prwkthresh2 when calculating precise out-of-bed times.
spurthresh	time in minutes between consecutive in-bed periods used to classify spurious data. When the interval between 2 in-bed periods < spurthresh, a new in-bed period corresponding to this interval is created.
nwperiod	length of interval in minutes considered for non-wear calculations.
nwctsgt0	number non-zero measurements within nwperiod allowed for non-wear classification.
nwmaxct	threshold for maximum value allowed when classifying non-wear intervals.
vhighpct	in-bed period with a proportion <e2><89><a5> vhighpct of non-wear measurements is considered a non-wear in-bed period.
highpct	in-bed period with a proportion <e2><89><a5> highpct of non-wear measurements and remaining proportion <e2><89><a4> slperiod is considered a non-wear in-bed period.
modpct	in-bed period with a proportion <e2><89><a4> modpct of non-wear measurements that is less than <e2><89><a4> fairtime is considered a wear in-bed period.
fairtime	maximum time allowed for wear in-bed period.
validday	number of minutes of waking wear required for a day to be considered valid.

Value

McVeighFcn returns a data frame with the following variables:

time	the input timevec vector.
cpm	the input cpmvec vector.
nw	either the input nwvec vector or the calculated non-wear vector.
inbed	binary variable indicating whether observation is an in-bed observation.
wakingwear	binary variable indicating whether observation is a waking wear observation.
validday	binary variable indicating whether observation occurred on a valid day. If timevec is not of type <e2><80><9c>POSIX<e2><80><9d>, NA will be returned for this variable.
validinbed	binary variable indicating whether observation is an inbed observation from a valid day. If timevec is not of type <e2><80><9c>POSIX<e2><80><9d>, NA will be returned for this variable.

Author(s)

Vincent Berarardi,Assistant Professor, Chapman University, Department of Psychology; John Bellettiere, Post Doctoral Research Fellow, University of California San Diego

References

McVeigh, J. A., Winkler, E. A., Healy, G. N., Slater, J., Eastwood, P. R., & Straker, L. M. (2016). Validity of an automated algorithm to identify waking and in-bed wear time in hip-worn accelerometer data collected with a 24 h wear protocol in young adults. Physiological Measurement, 37(10), 1636.

Examples

##---- Should be DIRECTLY executable !! ----
##-- ==>  Define data, use random,
##--	or do  help(data=index)  for the standard data sets.

## The function is currently defined as
function (x)
{
  }

vancebee/McVeighFcn documentation built on May 24, 2019, 7:50 a.m.