REMoDNaV

REMoDNaV detects and classifies eye movement events in gaze recordings from static and dynamic visual stimuli to enable robust analysis of saccades, post-saccadic oscillations, fixations, and smooth pursuit under temporally varying noise conditions.

Key Features:

• Algorithmic basis: Builds upon the adaptive Nyström & Holmqvist algorithm (Nyström & Holmqvist, 2010).
• Stimulus versatility: Operates on both static and dynamic stimulation environments, including prolonged recordings.
• Comprehensive event detection: Classifies saccades, post-saccadic oscillations, fixations, and smooth pursuit movements.
• Robustness to noise: Handles temporally varying noise levels, including conditions encountered during magnetic resonance imaging (MRI) recordings.
• Continuous-data processing: Processes data without requiring a trial structure and supports continuous gaze trajectories.
• Implementation: Implemented in Python.

Scientific Applications:

• Validation across datasets: Applied to manually annotated gaze trajectories for static images, moving dots, short video sequences, and feature-length movies.
• MRI-compatible eye-tracking: Suitable for analyzing recordings obtained under suboptimal lighting and noise conditions in MRI scanners.
• Natural viewing analysis: Supports analysis of prolonged dynamic recordings to capture biologically plausible eye-movement behavior.
• Event-level eye-tracking studies: Enables detailed classification and analysis of eye movements across diverse experimental paradigms.

Methodology:

Applies a velocity-based approach and extends the adaptive Nyström & Holmqvist algorithm (Nyström & Holmqvist, 2010) to classify eye-movement events.