Remove EOG artefacts using ICA — correct_eog

Decomposes the EEG into independent components via FastICA, identifies components whose time courses correlate with the EOG channels above threshold, subtracts their contribution from the EEG, and returns a cleaned mrpheus_psg. Mirrors MNE's ICA EOG rejection pipeline with default settings matching the companion Python notebook (n_components = min(6, n_eeg), threshold = 0.35).

Usage

correct_eog_ica(
  psg,
  eog_channels = NULL,
  eeg_channels = NULL,
  n_components = NULL,
  threshold = 0.35,
  fun = "logcosh",
  verbose = TRUE
)

Arguments

psg: An mrpheus_psg object from prepare_psg().
eog_channels: Character vector. EOG channel labels used as the artifact reference. If NULL (default), all "EOG" channels are used.
eeg_channels: Character vector. EEG channels to decompose and clean. If NULL (default), all non-bad "EEG" channels are used.
n_components: Integer or NULL. Number of ICA components. NULL (default) uses min(6L, n_eeg_channels), matching the notebook.
threshold: Numeric. Absolute Pearson correlation threshold above which a component is flagged as EOG-related. Default 0.35.
fun: Character. Contrast function passed to fastICA::fastICA(). "logcosh" (default) or "exp". The R method is used internally for robustness across channel counts; this is slower than the C backend but avoids matrix-conformality errors on small channel sets.
verbose: Logical. Print progress messages. Default TRUE.

Value

A new mrpheus_psg with cleaned EEG signals and re-segmented epochs. EOG and other channels are unchanged.

Details

Note: FastICA (this implementation) and MNE's default InfoMax ICA find different decompositions; the identified components and final signal will therefore differ from MNE output, but artifact removal performance is comparable.

Examples

if (FALSE) { # \dontrun{
rec   <- read_edf("psg.edf")
psg   <- prepare_psg(rec) |> preprocess_psg()
clean <- correct_eog_ica(psg)

# Stricter threshold
clean <- correct_eog_ica(psg, threshold = 0.5)
} # }