System and Data Requirements

Software

EyeSort 1.0 requires MATLAB with a working EEGLAB installation. It is distributed from the project’s GitHub releases rather than the EEGLAB plugin manager. ERPLAB is needed only for a downstream BINLISTER workflow; a deconvolution toolbox is likewise optional and downstream.

Dataset contract

Use continuous EEGLAB .set datasets whose eye events are already synchronized into EEG.event. Every selected dataset must come from the same input directory, and Step 1 requires an existing output directory even for one file.

At minimum, events must provide:

• a fixation event type and fixation X-position field;
• a saccade event type plus start-X and end-X fields if direction filters will be used;
• trial start and end codes;
• condition and item trigger codes that contain usable numeric identifiers;
• optional stimulus-window start/end codes when only part of each trial should be mapped.

The defaults match common EYE-EEG output: R_fixation, fix_avgpos_x, R_saccade, sac_startpos_x, and sac_endpos_x. All five names can be changed.

Interest-area file contract

The IA file is tab-delimited, with one row for each condition–item stimulus and columns for:

• numeric condition and item keys;
• one or more descriptive condition columns used in BDF text;
• the same ordered set of named text-region columns on relevant rows;
• optional covariates to import later.

Region cells contain literal displayed text. Spaces count toward geometry. Measure the sentence offset and pixels per character from the actual experiment display.

Applicability checks

Confirm single-line stimuli, fixed-width text, uniform character width, and a consistent region schema. EyeSort uses X coordinates only. Handle blinks, missing eye data, synchronization, artifact correction, and fixation detection upstream.

Next: install EyeSort or prepare your inputs.