Compute EEG band power per epoch

Estimates power spectral density (PSD) using Welch's method and integrates power within standard EEG frequency bands (delta, theta, alpha, sigma, beta, gamma) for each epoch and each specified channel. Mirrors the band-power feature extraction used by YASA's staging pipeline.

Usage

compute_band_power(
  psg,
  channels = NULL,
  bands = list(delta = c(0.5, 4), theta = c(4, 8), alpha = c(8, 13), sigma = c(13, 16),
    beta = c(16, 30), gamma = c(30, 40)),
  relative = FALSE,
  window_s = 4,
  overlap = 0.5
)

Arguments

psg

An mrpheus_psg object from prepare_psg().

channels

Character vector. EEG channel labels. If NULL (default), all non-bad EEG channels are used.

bands

Named list of length-2 numeric vectors defining frequency bands. Default:

list(delta = c(0.5, 4), theta = c(4, 8), alpha = c(8, 13),
     sigma = c(13, 16), beta = c(16, 30), gamma = c(30, 40))

relative

Logical. If TRUE, return relative band power (band / total power in 0.5–40 Hz). Default FALSE.

window_s

Numeric. Welch window length in seconds. Default 4.

overlap

Numeric in [0, 1). Fractional overlap between Welch windows. Default 0.5.

Value

A tibble with columns epoch, channel, one column per band, and total_power. Units are µV²/Hz (or dimensionless if relative = TRUE).

Examples

if (FALSE) { # \dontrun{
bp <- compute_band_power(psg)
bp <- compute_band_power(psg, channels = "EEG Fpz-Cz", relative = TRUE)
} # }