Motion-corrected eye tracking improves gaze accuracy during visual fMRI experiments

Abstract

Abstract Human eye movements are essential for understanding cognition, yet achieving high-precision eye tracking in functional Magnetic Resonance Imaging (fMRI) remains challenging. Even slight head shifts from the initial calibration position can introduce drift in eye tracking data, leading to substantial gaze inaccuracies. To address this, we present Motion-Corrected Eye Tracking (MoCET), which corrects drift using head motion parameters derived from fMRI preprocessing. MoCET requires no additional hardware and can be applied retrospectively to existing datasets. We show that it outperforms conventional detrending methods with respect to accuracy of gaze estimation and offers higher spatial and temporal precision compared to magnetic resonance-based eye tracking approaches. By overcoming a key limitation in integrating eye tracking with fMRI, MoCET enables precise investigations of naturalistic vision and cognition in fMRI research.