A functional model for studying common trends across trial time in eye tracking experiments.

Area of Science:

• Neuroscience
• Data Science
• Developmental Psychology

Background:

• Eye tracking (ET) experiments collect detailed gaze trajectory data.
• Traditional ET analysis often simplifies data into summary statistics, losing temporal information.
• There is a need for methods that preserve the rich temporal dynamics of gaze behavior.

Purpose of the Study:

• Introduce functional data analysis (FDA) for analyzing eye tracking data.
• Develop novel functional outcomes, or "viewing profiles," to capture gaze trends over trial time.
• Apply FDA to identify group differences in visual exploration patterns.

Main Methods:

• Utilized functional data analysis (FDA) for eye tracking data.
• Introduced "viewing profiles" as a novel functional outcome measure.
• Employed functional principal components analysis to model mean and variation of viewing profiles.

Main Results:

• The proposed FDA approach successfully captured temporal gaze trends lost in traditional summaries.
• Functional principal components analysis effectively modeled gaze profile variations across subjects.
• Significant differences were found in the consistency of looking at faces early in trial time between children with autism and typically developing children.

Conclusions:

• Functional data analysis provides a powerful new framework for analyzing eye tracking data.
• Viewing profiles offer a more comprehensive understanding of gaze behavior dynamics.
• The FDA approach highlights distinct visual exploration patterns in children with autism, particularly in early face engagement.