Gazepath

Gazepath parses raw eye-tracking data into fixations and saccades to enable robust analysis across varying data quality and participant populations.

Key Features:

• Non-Parametric Speed-Based Parsing: Transforms raw gaze data into fixations and saccades on a trial basis using a non-parametric, speed-based approach.
• Dynamic Trial-Based Thresholding: Dynamically adjusts detection thresholds per trial or dataset to accommodate varying data characteristics and sampling rates.
• Adaptability Across Populations: Integrates methodologies from adult and infant eye-tracking literature to handle age-related differences in eye movement patterns.
• Quality Control Mechanisms: Implements controls to mitigate spurious correlations between fixation durations and data quality.
• Support for Noisy and Low-Sampling-Rate Data: Handles noisy datasets and low sampling rates such as Tobii recordings at 60 Hz.

Scientific Applications:

• Free-Viewing Data: Parses free-viewing data from infants and adults, accounting for individual differences and data quality and reported to outperform standard EyeLink parsing methods.
• High-Quality Reading Data: Detects precise fixations and saccades in adult reading datasets with high data quality for detailed reading behavior analysis.
• Noisy Infant Data: Robustly parses challenging infant datasets collected with Tobii eye-trackers at low sampling rates (60 Hz).

Methodology:

Uses a non-parametric speed-based parsing algorithm with trial-based threshold adjustment, integrates adult and infant parsing methodologies, and applies quality-control procedures to reduce spurious correlations between fixation durations and data quality.