Model-Based Separation, Detection, and Classification of Eye Movements.

Area of Science:

• Neuroscience
• Ophthalmology
• Biomedical Engineering

Background:

• Accurate eye movement analysis is crucial for understanding neurological disorders and visual perception.
• Existing methods often struggle to differentiate between various eye movement types and are susceptible to artifacts.

Purpose of the Study:

• To develop a physiologically motivated framework for model-based separation, detection, and classification (MBSDC) of eye movements.
• To independently analyze saccades, smooth pursuit, and fixational eye movements using a mechanistic oculomotor model.

Main Methods:

• Extended an oculomotor model with neural controller and blink artifact models.
• Employed Kalman smoothing and sparse input estimation to derive kinematic and neural signals from eye position data.
• Utilized estimated signals for detecting saccade endpoints and classifying eye movements.

Main Results:

• Achieved approximately 50% reduction in velocity profile reconstruction error compared to traditional filtering methods on simulated data.
• Demonstrated accurate signal separation for smooth pursuit eye movements in human subjects.
• Showcased striking correlation between estimated neural controller signals and actual recordings in non-human primates.

Conclusions:

• The MBSDC framework facilitates the analysis of diverse eye movement recordings.
• Offers a physiologically grounded method for studying motor commands and potentially identifying novel digital biomarkers.