Frequently Asked Questions

Does EyeSort label saccades?

No. EyeSort assigns codes to fixation events. Saccade events provide incoming/outgoing direction criteria for selecting those fixations.

Does EyeSort support pixel-drawn regions or arbitrary images?

No. Version 1.0 reconstructs named text regions from literal strings, one horizontal offset, and pixels per character.

Can every word be selected as its own region?

Not automatically in the label GUI. EyeSort computes word index/text metadata, but labeling operates on user-defined region columns. You may define word-sized regions only when the region schema remains consistent and interpretable.

Does EyeSort save my result?

Yes. Apply All & Finish auto-saves <base>_processed.set/.fdt to the Step 1 output directory. There is no Save Labeled Dataset menu item.

Can I change readable event types later?

Yes. eyesort_full_code remains canonical, while Modify Event Code Format rewrites EEG.event.type. original_type supports restoration when available.

Can I use ERPLAB or deconvolution?

Yes, downstream. EyeSort can generate a BINLISTER BDF and its event metadata can serve as model predictors. It does not epoch, average, or correct temporal overlap.

Can I batch process?

Yes. Inputs must come from one directory; outputs go to the required output directory and files are processed one at a time.

How do I reproduce a run?

Save the text-IA config and full label queue, export the History Script bundle, preserve the IA file, and report versions, conflict policy, criteria, and post-label counts.

See known limitations and issue reporting.