Full EEG preprocessing pipeline

Convenience wrapper that runs a complete EEG preprocessing pipeline in sequence: filtering, re-referencing, bad channel detection and interpolation, optional ICA, epoching, and artifact rejection. Each step can be toggled on or off.

Usage

eegPreprocess(
  x,
  filter = TRUE,
  lowcut = 0.1,
  highcut = 40,
  notch = NULL,
  rereference = TRUE,
  ref_type = "average",
  bad_channels = TRUE,
  interpolate = TRUE,
  ica = FALSE,
  epoch = FALSE,
  events = NULL,
  epoch_limits = c(-0.2, 0.8),
  baseline = c(-0.2, 0),
  artifact_reject = FALSE,
  threshold_uv = 100,
  assay_name = NULL,
  verbose = TRUE
)

Arguments

x

A PhysioExperiment object with continuous (2D) EEG data.

filter

Logical; apply frequency filtering (default: TRUE).

lowcut

Low cutoff frequency in Hz for filtering (default: 0.1).

highcut

High cutoff frequency in Hz for filtering (default: 40).

notch

Notch filter center frequency in Hz (NULL for none).

rereference

Logical; apply re-referencing (default: TRUE).

ref_type

Re-referencing type (default: "average").

bad_channels

Logical; detect and report bad channels (default: TRUE).

interpolate

Logical; interpolate detected bad channels (default: TRUE). Requires electrode positions in colData (apply eegMontage first or this will be skipped with a warning).

ica

Logical; apply ICA-based artifact removal (default: FALSE). Requires the eegICA function to be available.

epoch

Logical; epoch the data around events (default: FALSE).

events

Event data for epoching (data.frame or integer vector). Required if epoch = TRUE.

epoch_limits

Epoch window in seconds (default: c(-0.2, 0.8)).

baseline

Baseline window in seconds (default: c(-0.2, 0)).

artifact_reject

Logical; reject artifact epochs (default: FALSE). Only applies if epoch = TRUE.

threshold_uv

Amplitude threshold for artifact rejection (default: 100).

assay_name

Starting assay name. If NULL, uses defaultAssay(x).

verbose

Logical; print progress messages (default: TRUE).

Value

A PhysioExperiment object with processed data. The final assay name depends on which steps are enabled. Processing log is stored in metadata(x)$preprocess_log.

Examples

pe <- make_eeg(n_time = 10000, n_channels = 19, sr = 500)
pe <- eegMontage(pe, system = "10-20")
# filter + average-reference; bad-channel detection/interpolation disabled
# so the example does not depend on the montage's neighbour geometry
pe_proc <- eegPreprocess(pe, lowcut = 1, highcut = 40, notch = 50,
                         bad_channels = FALSE, interpolate = FALSE,
                         verbose = FALSE)
pe_proc
#> class: PhysioExperiment
#> dim: 10000 x 19 
#> assays(3): raw, filtered, rereferenced
#> samplingRate: 500 Hz
#> channels(19): Fp1, Fp2, F7, F3, Fz ...
#> colData names(5): label, type, pos_x, pos_y, pos_z
#> provenance: 1 steps