Snip & Stitch: a simple and accessible correction for the pupil foreshortening error

Abstract

Modern eye trackers allow easy measurement of pupil size, a powerful physiological marker of cognitive processing. A major hindrance to using pupillometry is the so-called pupil foreshortening error (PFE). The PFE is the apparent change in pupil size that results from a change in the angle of the eye relative to the camera, despite the pupil remaining constant in reality. To prevent this, experimenters often rely on fixed gaze, limiting design flexibility. Existing correction methods necessitate specific geometric properties of the setup, often unavailable for existing datasets, and do not work for all tracker types. We here introduce ‘Snip & Stitch’, a straightforward, accessible, and easy-to-implement correction for the PFE. Specifically, Snip & Stitch corrects the PFE by ‘snipping’ pupil size changes during a saccade, and ‘stitching’ pupil size back by subtracting the difference between pre- and post-saccadic pupil size. Our results demonstrate that this simple method reduced the PFE by an estimated 71–81%. Snip & Stitch is openly available online. We recommend Snip & Stitch, especially for experiments in which participants make one saccade before pupil size is compared, and argue that the method can be easily applied to experiments in which participants make up to five saccades in sequence. Limitations and possible further improvements are discussed. Together, Snip & Stitch allows researchers to employ pupil size as an outcome measure in a wide range of tasks and setups.