Skip to contents

Prepare binned pupil time-course data for GAMM modelling with mgcv::bam(). The function aggregates processed sample-level pupil data into subject-by- condition-by-time-bin rows and creates an AR.start indicator for autoregressive GAMM models.

Usage

prepare_gazepoint_pupil_gamm_data(
  data,
  pupil_col = NULL,
  time_col = "time",
  subject_col = "subject",
  condition_col = "condition",
  x_col = NULL,
  y_col = NULL,
  group_cols = c("subject", "condition"),
  bin_width_ms = 50,
  time_window = NULL,
  min_valid_samples = 1,
  missing_condition_label = "all_data"
)

Arguments

data

A Gazepoint sample-level data frame, usually after pupil preprocessing, interpolation, baseline correction, and optional smoothing.

pupil_col

Name of the pupil column to aggregate. If NULL, the function tries common processed pupil columns such as pupil_smoothed, pupil_baseline_corrected, pupil_interpolated, pupil_clean, and pupil_for_preprocessing.

time_col

Name of the time column in milliseconds. If the requested column is not available, the function tries common alternatives.

subject_col

Name of the subject column. If unavailable, the function tries common participant identifiers.

condition_col

Name of the condition column. If unavailable or entirely missing, a single condition label is used.

x_col

Optional gaze x-coordinate column. If NULL, common x-coordinate columns are auto-detected when available.

y_col

Optional gaze y-coordinate column. If NULL, common y-coordinate columns are auto-detected when available.

group_cols

Columns defining independent time series before binning. Defaults to c("subject", "condition").

bin_width_ms

Width of time bins in milliseconds.

time_window

Optional numeric vector of length 2 giving the time window to retain before binning.

min_valid_samples

Minimum number of valid pupil samples required for a bin to be retained.

missing_condition_label

Label used when condition values are missing or when no usable condition column is available.

Value

A tibble with binned pupil time-course data for GAMM modelling.