OpenMReye Banner

Magnetic resonance-based eye tracking without camera

OpenMReye is a growing ecosystem of open computational tools for MR-based eye tracking for functional MRI research and clinical applications.

What is magnetic resonance-based eye tracking?

MR-based eye tracking describes a set of techniques that reconstruct gaze behavior directly from the MR-signal of the eyes, without requiring a camera. It works because gaze position and movement affect the MRI signal of the eyes and optic nerves, allowing to infer gaze-related variables directly from that raw signal.

By replacing camera hardware with smart, open-source software, OpenMReye enables even the smallest MRI facilities and hospitals to use (MR-based) eye tracking for a wide range of research and clinical applications.

OpenMReye Demo
Why would fMRI researchers use MR-based eye tracking?

Gaze behavior provides a powerful window into human cognition and a promising biomarker for many neurological conditions (e.g., Parkinson's Disease). Yet, most brain imaging studies miss this crucial measure, primarily due to the high costs (often >€50,000) and cumbersome use of MR-compatible cameras. Moreover, cameras impose experimental constraints (e.g., the eyes need to remain open).

While MR-based eye tracking comes with limitations compared to cameras (e.g., lower temporal resolution), it also has many advantages. For example, no MR-compatible hardware is needed, drastically reducing costs and setup time. Moreover, it works in many existing fMRI datasets, when the eyes are closed (e.g., sleep), and in patient groups for which cameras are difficult to calibrate (e.g., blind populations).

MR-based eye tracking complements camera-based systems by enabling research questions that are otherwise difficult or impossible to address.

Tools & Resources

DeepMReye

Decoding gaze position directly from eye-voxel patterns using deep neural networks. Our flagship model, extensively validated, widely applicable.

HMMeye

Eye voxel-based event segmentation using Hidden Markov Models. Perfect for naturalistic tasks (e.g., movie viewing) without gaze training labels.

MReyeMove

Coming soon

Reconstructing putative eye movements from eye-voxel patterns using simple multi-voxel pattern analysis. No training data required.

Eye Tracking Tips

Check out our tips for improving eye-tracking quality in neuroimaging experiments.

View Guide →

Contributors

Read more

  • DeepMReye:
    Frey, M.*, Nau, M.*†, & Doeller, C. F.† (2021).
    Magnetic resonance-based eye tracking using deep neural networks.
    Nature Neuroscience. https://doi.org/10.1038/s41593-021-00947-w
    * Shared-first author   † Shared-senior author
  • HMMeye:
    Nau, M., Greene, A., Tarder-Stoll, H., Lossio-Ventura, J. A., Pereira, F., Chen, J., Baldassano, C., & Baker, C. I. (2025). Neural and behavioral reinstatement jointly reflect retrieval of narrative events.
    Nature Communications 16, 7865. https://doi.org/10.1038/s41467-025-62375-9
  • MReyeMove:
    Nudelman, G., & Nau, M. (in prep). Brain-wide gaze-dependent activity during eyes-closed rest and sleep.
Honors

OSCAward 2025

Awarded by the Open Science Community Amsterdam for advancing inclusivity, reproducibility, and rigor in brain imaging.


FGB Open Science Award 2026

Awarded by the Faculty of Behavioral and Movement Sciences (FGB) at Vrije Universiteit Amsterdam for best Open Science practices.