accelerometry (version 3.1.2)

bouts: Physical Activity Bout Detection

Description

Identify bouts of physical activity based on a vector of accelerometer count values.

Usage

bouts(counts, weartime = NULL, bout_length = 10L, thresh_lower = 0L,
  thresh_upper = 100000L, tol = 0L, tol_lower = 0L, tol_upper = 100000L,
  nci = FALSE, days_distinct = FALSE)

Arguments

counts

Integer vector with accelerometer count values.

weartime

Integer vector with 1's for wear time minutes and 0's for non-wear time minutes.

bout_length

Integer value specifying minimum length of an activity bout.

thresh_lower

Integer value specifying lower bound for count values to be included for the intensity level.

thresh_upper

Integer value specifying upper bound for count values to be included for the intensity level.

tol

Integer value specifying number of minutes with count values outside of [thresh_lower, thresh_upper] to allow during an activity bout.

tol_lower

Integer value specifying lower cut-off for count values outside of intensity range during an activity bout.

tol_upper

Integer value specifying upper cut-off for count values outside of intensity range during an activity bout.

nci

Logical value for whether to use algorithm from NCI's SAS programs. See Details.

days_distinct

Logical value for whether to treat each day of data as distinct, i.e. identify non-wear time and activity bouts for day 1, then day 2, etc. If FALSE, algorithm is applied to full monitoring period continuously. If protocol has participants remove accelerometer for sleep, strongly recommend setting to FALSE to capture non-wear periods that start between 11 pm and midnight. Function assumes that first 1440 data points are day 1, next 1440 are day 2, and so on.

Value

Integer vector with 1's for minutes that are part of an activity bout and 0's for minutes that are not.

Details

If nci = FALSE, the algorithm uses a moving window to go through every possible interval of length bout_length in counts. Any interval in which all counts are >= tol_lower and <= tol_upper, and no more than tol counts are less than thresh_lower or greater than thresh_upper, is classified as an activity bout.

If nci = TRUE, activity bouts are classified according to the algorithm used in the NCI's SAS programs. Briefly, this algorithm defines an activity bout as an interval of length bout_length that starts with a count value in [thresh_lower, thresh_upper] and has no more than tol counts outside of that range. If these criteria are met, the bout continues until there are (tol + 1) consecutive minutes outside of [thresh_lower, thresh_upper]. The parameters tol_lower and tol_upper are not used.

If the user allows for a tolerance (e.g. tol = 2) and does not use the NCI algorithm (i.e. nci = FALSE), specifying a non-zero value for tol_lower is highly recommended. Otherwise the algorithm will tend to classify minutes immediately before and after an activity bout as being part of the bout.

Specifying thresh_lower while using an arbitrarily large value for thresh_upper is generally recommended. Specifying both of these parameters can be overly restrictive in that the algorithm may miss bouts of activity in which counts are consistently high, but not exclusively in one intensity range.

References

National Cancer Institute. Risk factor monitoring and methods: SAS programs for analyzing NHANES 2003-2004 accelerometer data. Available at: http://riskfactor.cancer.gov/tools/nhanes_pam. Accessed Aug. 19, 2018.

Acknowledgment: This material is based upon work supported by the National Science Foundation Graduate Research Fellowship under Grant No. DGE-0940903.

Examples

# NOT RUN {
# Load accelerometer data for first 5 participants in NHANES 2003-2004
data(unidata)

# Get data from ID number 21005
counts.part1 <- unidata[unidata[, "seqn"] == 21005, "paxinten"]

# Identify periods of valid wear time
wear.part1 <- weartime(counts = counts.part1)

# Identify moderate-to-vigorous activity bouts
mvpa.bouts <- bouts(counts = counts.part1, weartime = wear.part1, 
                    thresh_lower = 2020)


# }