This repository holds a few R functions used for the pre-processing of SR Research Eyelink EDF files. If you should ever use these functions, you may acknowledge the publication where they were first used:
- Schweitzer, R., & Rolfs, M. (2020). Intra-saccadic motion streaks as cues to linking object locations across saccades. Journal of vision, 20(4), 17.
In order to use those functions, just source them.
source("~/your_folder/convert_edf_to_asc.R")
source("~/your_folder/extract_trial_from_edf.R")
source("~/your_folder/scan_asc_trial_bounds.R")
source("~/your_folder/interpolate_time.R")Some edf2asc function must be installed on your system. The function in this repository calls wine to run the edf2asc.exe to convert a specified EDF file to ASC.
saccade_edf_file <- convert_edf_to_asc(file_path = path_to_edf, edf_file = saccade_edf_filename,
options = '-y -t -failsafe')You should now have a ASC files that you can read in as a text file.
saccade_asc_file <- readLines(con=saccade_edf_file)Provided that you have sent messages to your Eyelink, such as "Trial_Start_1" and "Trial_End_1", to demarcate trials, you can now parse the text file for those lines that contain trial data. The function can assert consistency (assert_consistency=TRUE), assuming that the trial counter starts at 1. If that is not the case, you may specify a trialCntr_addon, i.e., the counter of the first trial in the EDF file.
saccade_trial_bounds <- scan_asc_trial_bounds(saccade_asc_file,
trialCntr_addon = 0,
assert_consistency = TRUE)
head(saccade_trial_bounds)
paste("Found", nrow(saccade_trial_bounds), "consistent trial boundaries in EDF.")You can now extract properly formatted data of a specified trial, let's say trial 1. To do that, you specify the boundaries of the trial from the previous step and supply the matching text file. If you tracked monocularly, always specify dom_eye = 0 and normal_msg_length = 5. If you tracked binocularly, you may choose dom_eye = 0 or 1, but specify normal_msg_length = 8, so that the eyelink message and data rows can be properly recognized. Also be sure to have a matching msg_separator (default: "\t"), so that messages can be parsed. If you have sent custom messages during your trials, you may specify them in msg_of_interest and their timestamps will be returned if they were found (if not, you'll receive a warning). Finally, col_names are the column names of the returned data.frame. Update: There is a new parameter called use_end_of_string that deals with cases in which you have defined messages called "EVENT_sac" and "EVENT_sacend" (Thanks, Kai). If you then parsed for "EVENT_sac", you would find two matches. If you have messages like that and want to avoid such a situation, you may set use_end_of_string to TRUE (default: FALSE).
trial_edf <- extract_trial_from_edf_nosearch(trial_i_start = saccade_trial_bounds$trial_i_starts[1],
trial_i_end = saccade_trial_bounds$trial_i_ends[1],
asc_file = saccade_asc_file,
dom_eye = 0,
normal_msg_length = 5,
used_mode_1440 = FALSE,
use_end_of_string = FALSE,
msg_of_interest = c("EVENT_cueOn", "EVENT_timeSaccade",
"EVENT_timeMoveStarted", "EVENT_timeMoveFinished",
"EVENT_boundaryCross"),
col_names = c("sac_t_raw", "sac_x_raw", "sac_y_raw", "sac_pupil") )Finally, if you have sampled online data during your experiment and thus have both the retrieval timestamps of the eyelink and the experiment PC (below the vectors t.eye and t, respectively), then you can remap eyelink timestamps (such as those in trial_edf$msg or trial_edf$eye_raw$sac_t_raw below) to the time on your experiment PC.
interpolated_pc_timestamps <- interpolate_pc_from_eyelink(eyelink_time = t.eye,
pc_time = t,
eyelink_timestamps = trial_edf$msg )
interpolated_pc_time <- interpolate_pc_from_eyelink(eyelink_time = t.eye,
pc_time = t,
eyelink_timestamps = trial_edf$eye_raw$sac_t_raw,
use_names = FALSE )