Skip to contents

Multitaper Power Spectral Density for EEG

Source: R/eeg-timefreq.R
eegMultitaper.Rd

Computes the multitaper power spectral density (PSD) estimate for multi-channel EEG data using Discrete Prolate Spheroidal Sequences (DPSS, Slepian tapers). The multitaper method provides an optimal bias-variance trade-off for spectral estimation compared to single-window methods.

Usage

eegMultitaper(
  x,
  bandwidth = 4,
  n_tapers = NULL,
  assay_name = NULL,
  output_assay = "multitaper_psd"
)

Arguments

x

A PhysioExperiment object with EEG data (2D: time x channels).

bandwidth

Time-half-bandwidth product (NW) controlling the spectral concentration of the tapers (default: 4). Higher values give smoother but lower-resolution estimates.

n_tapers

Number of DPSS tapers to use. If NULL, defaults to floor(2 * bandwidth) - 1.

assay_name

Name of the input assay. If NULL, the default assay is used.

output_assay

Name of the assay to store PSD results (default: "multitaper_psd").

Value

Modified PhysioExperiment with:

  • PSD matrix (frequencies x channels) in output_assay

  • Frequency vector and taper parameters in metadata(x)$multitaper, a list containing frequencies (numeric vector), bandwidth (numeric NW parameter), and n_tapers (integer)

References

Tallon-Baudry, C., et al. (1997). Oscillatory gamma-band activity during conscious perception. Trends in Cognitive Sciences, 3(4), 151-162.

Thomson, D. J. (1982). Spectrum estimation and harmonic analysis. Proceedings of the IEEE, 70(9), 1055-1096.

See also

eegMorletWavelet(), eegSTFT(), eegPlotSpectrogram()

Examples

if (FALSE) { # \dontrun{
pe <- make_eeg(n_time = 5000, n_channels = 19, sr = 500)
pe_mt <- eegMultitaper(pe, bandwidth = 4)
psd <- SummarizedExperiment::assay(pe_mt, "multitaper_psd")
dim(psd)  # frequencies x channels
} # }