Detection of eye movements from fMRI data

Area of Science:

• Neuroimaging
• Cognitive Neuroscience
• Biomedical Engineering

Background:

• Eye movements in awake humans vary with behavioral state and task.
• These movements complicate functional magnetic resonance imaging (fMRI) by causing signal variance in adjacent brain regions and confounding task-related activation.
• Existing methods for tracking eye movements in fMRI require expensive and specialized equipment.

Purpose of the Study:

• To describe a novel method for detecting eye movements directly from fMRI data.
• To validate this method against a standard infrared pupil tracking system.
• To assess the utility of this technique for improving fMRI data analysis.

Main Methods:

• Developed a method to detect eye movements by analyzing MR time series data from the vitreous of the eye.
• Correlated MR signal changes in the eye with simultaneous measurements from an infrared pupil tracking system.
• Applied the method to 10 subjects performing fixation and eye movement tasks.

Main Results:

• MR time series data from the eye vitreous showed significantly greater variance during eye movements compared to central fixation (average SD 99.7 vs. 75.6, P = 0.001).
• The MR-based detection of eye movements correlated well with the infrared pupil tracking system.
• The method successfully identified eye movements without requiring additional hardware within the scanner.

Conclusions:

• Eye movements can be reliably detected directly from fMRI data by analyzing signals from the eye's vitreous.
• This technique offers a cost-effective and non-invasive alternative to external eye-tracking systems for fMRI.
• The method has significant potential for enhancing the accuracy of fMRI studies, particularly in retrospective and meta-analyses of brain activity.