Skip to contents

Compute Power Spectra With Optional Subject-Level Aggregation

Source: R/plotting.R
get_power_spectra.Rd

Computes power spectral density estimates using spectrum, optionally aggregated across subjects or posterior predictive samples. All arguments intended for the underlying spectral estimator should be supplied through spectrum.args.

Usage

get_power_spectra(data, by.postn = FALSE, spectrum.args = list())

Arguments

data

A data frame with reaction time data. Must contain subjects and rt, and for posterior predictive data optionally postn and trials.

by.postn

Logical. If TRUE, compute a separate spectrum for each posterior predictive draw and each posterior sample index.

spectrum.args

A named list of arguments passed directly to spectrum. These override the defaults internally used in this function. Useful for customizing smoothing spans, detrending, tapering, and so on. Defaults: list(spans=c(3, 5), detrend=FALSE, demean=TRUE, log=FALSE, taper=0). By default, we run spectrum without log, and log-transform while plotting

Value

Either a data frame with columns freq and power, or (if by.postn = TRUE) a list with frequency vector and a matrix of spectra across posterior samples.

Details

The function organizes the data by subject (and optionally posterior sample index), computes spectra individually, interpolates spectra to a common frequency grid if needed, and averages them appropriately.