A novel blink detection method based on pupillometry noise.

Area of Science:

• Cognitive Science
• Neuroscience
• Psychology

Background:

• Pupillometry, measuring pupil size, is a key indicator of cognitive load and arousal.
• Eyetracking devices record pupil size but lose data during blinks, creating missing values.
• Blinks introduce artifactual eyelid signals that can distort pupil size data if not removed.

Purpose of the Study:

• To develop a novel algorithm for accurate detection of blink onset and offset in pupillometry data.
• To address limitations of existing methods, such as failure to remove eyelid signals and significant data loss.

Main Methods:

• The novel algorithm identifies blink onset and offset based on characteristic fluctuations (noise) in pupil size data.
• It distinguishes blink-related fluctuations from the eyelid signal.
• Algorithm performance was validated against three common blink detection methods and human raters.

Main Results:

• The proposed algorithm demonstrates superior effectiveness in detecting blink onset and offset compared to existing methods.
• It successfully differentiates blink artifacts from genuine pupil size changes.
• The algorithm minimizes data loss associated with blink removal.

Conclusions:

• This novel blink detection algorithm provides a more accurate and less data-intensive approach for pupillometry.
• It enhances the reliability of pupil size as an index of cognitive load and arousal.
• The algorithm's code is publicly available for researchers.