Detect artifacts and add them in the events table of an eeg_lst

These functions search for artifacts on the signal table based on a threshold and a sliding window (when relevant), and annotate an event in the events table that spans from -lim to +lim. The signal table remains unchanged until eeg_events_to_NA().

eeg_artif_minmax(
  .data,
  ...,
  .threshold = 100,
  .direction = "above",
  .window = 0.2,
  .lim = c(-.window, .window),
  .unit = "s",
  .freq = NULL,
  .config = list()
)

eeg_artif_step(
  .data,
  ...,
  .threshold = 50,
  .window = 0.2,
  .lim = c(-.window, .window),
  .unit = "s",
  .freq = NULL,
  .config = list()
)

eeg_artif_amplitude(
  .data,
  ...,
  .threshold = c(-200, 200),
  .lim = c(-0.2, 0.2),
  .unit = "s",
  .freq = NULL,
  .config = list()
)

eeg_artif_peak(
  .data,
  ...,
  .threshold = 30,
  .window = 0.2,
  .lim = c(-.window, .window),
  .unit = "s",
  .freq = NULL,
  .config = list()
)

Arguments

.data

An eeg_lst object.

...

Channels to include. All the channels by default, but eye channels should be removed.

.threshold

Voltage threshold that indicates an artifact

.direction

Whether to look "above" or "below" the threshold.

.window

Sliding window length for the artifact detection (same unit as lim). This is the full width of the step function: this means that we are looking for a period of one voltage for half of the window immediately followed by a period of a different voltage (indicated by the threshold) for half of the window.

.lim

Vector with two values indicating the time before and after the artifact that will be included in events_tbl (by default the size the window before and afterwards).

.unit

"seconds" (or "s"), "milliseconds" (or "ms")

.freq

Vector with two values indicates whether to prefilter the signal prior to the artifact detection. (The filtering is not saved in the signal). For a low pass filter the first value should be NA, for a high-pass filter the second value should be NA.

.config

List with the configuration of the filter.

Value

An eeg_lst.

Details

eeg_artif_peak() is wrapper around pracma::findpeaks, .threshold is the minimum (absolute) height a peak has to have to be recognized as such and .window is the minimum distance peaks have to have to be counted.

eeg_artif_minmax() is also refered as a peak-to-peak artifact detector. It is less sensitive to drifts than eeg_artif_peak().

eeg_artif_minmax() and eeg_artif_step() can be used to detect blinks and horizontal eye movements in the electro-oculographic (V/HEOG) channels or large voltage jumps in other channels. See chapter 6 of Luck (2014).

For the EOG channels, a relatively low threshold (e.g., 30 µV) is recommended. For non EOG channels, a relatively high threshold (e.g., 100 µV) would be more appropriate.

References

Luck, S. J. (2014). An introduction to the event-related potential technique. MIT press.

See also

Other events functions: eeg_events_to_NA()

Examples

if (FALSE) {

# Artifacts are annotated in the events table:
faces_seg_artif <- faces_seg %>%
  eeg_artif_minmax(-HEOG, -VEOG, .threshold = 100, .window = 150, unit = "ms") %>%
  eeg_artif_step(-HEOG, -VEOG, .threshold = 50, .window = 200, unit = "ms")

# Signals with artifacts are turned into NA values:
faces_clean <- faces_seg_artif %>%
  eeg_events_to_NA(.type == "artifact", .entire_seg = TRUE, .drop_events = TRUE)
}