Automated iPad-based gaze detection in the NIH Baby Toolbox® norming study

Area of Science:

• Developmental Psychology
• Pediatric Neurodevelopment
• Human-Computer Interaction

Background:

• The NIH Baby Toolbox® is a comprehensive iPad-based assessment for children 1-42 months.
• Existing developmental assessments often require motor or verbal responses.
• Gaze paradigms are crucial in developmental research for non-invasive cognitive assessment.

Purpose of the Study:

• To describe the development and validation of gaze technology for the Baby Toolbox.
• To assess the feasibility and reliability of iPad-based eye gaze data collection in infants.
• To establish normative data for a novel developmental assessment tool.

Main Methods:

• Development and validation of gaze technology integrated into the Baby Toolbox.
• Utilized iPad's native augmented reality features for real-time eye gaze data collection.
• Norming study included 1079 infants aged 6-23 months.

Main Results:

• Strong calibration success was observed across diverse demographic groups (age, sex, race/ethnicity, parental education, eye color).
• Some group differences in calibration emerged for race/ethnicity and parental education.
• Reliability against hand-coding showed moderate to strong results, with individual variability.

Conclusions:

• The Baby Toolbox gaze technology represents a novel and effective method for developmental assessment.
• This technology offers a non-invasive approach to measuring infant attention and cognitive processes.
• Findings support the use of this technology in large-scale developmental studies.