Eye tracking identifies biomarkers in α-synucleinopathies versus progressive supranuclear palsy

Mahboubeh Habibi, Wolfgang H. Oertel, Brian J. White, Donald C. Brien, Brian C. Coe, Heidi C. Riek, Julia E. Perkins, Rachel Yep, Laurent Itti, Lars Timmermann, Christoph Best, Elisabeth Sittig, Annette Janzen, Douglas P. MunozJournal of Neurology 2022 · DOI

This study explores how the way we look at the world—specifically our eye movements and pupil reactions—can serve as a "window" into the brain to differentiate between similar neurodegenerative diseases and even catch early warning signs before major symptoms appear.

TL;DR

Researchers used a simple 10-minute "Free Viewing" video task to identify distinct oculomotor and pupillary biomarkers that distinguish Parkinson’s disease and Multiple System Atrophy from Progressive Supranuclear Palsy. Crucially, they found that patients with REM sleep behavior disorder (RBD) already exhibit subtle eye-movement abnormalities, marking a potential prodromal biomarker for future neurodegeneration.

What Problem Does This Paper Address?

Clinicians often struggle to differentiate between various movement disorders in their early stages. Parkinson’s disease (PD) and Multiple System Atrophy (MSA) are $\alpha$-synucleinopathies ($\alpha$SYN), while Progressive Supranuclear Palsy (PSP) is a tauopathy. Though they share motor symptoms, their underlying pathology and progression differ significantly.

Furthermore, there is a pressing need for "prodromal" biomarkers—indicators that appear before the onset of classic motor symptoms. Isolated REM sleep behavior disorder (RBD) is a known precursor; up to 85% of these patients eventually convert to a manifest $\alpha$SYN like PD. The authors sought to determine if a non-invasive, naturalistic eye-tracking task could provide the "fingerprints" needed to tell these diseases apart and identify early-stage changes in RBD.

Methods

The researchers moved away from traditional, rigid "look-at-the-dot" tests in favor of a Free Viewing (FV) paradigm.

  • The Task: Participants watched 10 minutes of short, 2–5 second video clips (animals, nature, city scenes). They were given no instructions other than to watch the screen.
  • The Tech: A high-speed video-based eye tracker (Eyelink 1000 Plus) recorded eye position and pupil size at 500 Hz.
  • Analysis: The team analyzed "macro-saccades" (large jumps in gaze), "micro-saccades" (tiny fixational movements), and pupil constriction/dilation triggered by the sudden luminance changes when video clips switched. They also mapped "center bias"—the tendency to stare at the middle of the screen.

Datasets

The study involved a total of 232 participants:

  • Patient Groups: 46 with RBD, 27 with PD, 17 with MSA, and 10 with PSP.
  • Control Group: 132 healthy, age-matched individuals (CTRL).
  • Clinical Metrics: Patients were also assessed using standard scales like the UPDRS-III (motor function), MoCA (cognition), and BDI-II (depression) to correlate eye data with disease severity.

Key Insights and Ideas

  • The "Tunnel Vision" of Neurodegeneration: All patient groups exhibited a significant "center bias." They spent more time fixating on the center of the screen and less time exploring the periphery compared to healthy controls. This suggests a reduced ability or motivation to "harvest" visual information from the environment.
  • The RBD "Aha" Moment: RBD patients showed discrete saccade abnormalities (reduced frequency and amplitude) that were intermediate between healthy controls and PD/MSA patients. This confirms that oculomotor circuits are affected long before the "manifest" stage of the disease.
  • Opposite Pupil Profiles: One of the most striking findings was the divergence in pupil control. When screen brightness changed, MSA patients showed exaggerated pupil responses, while PSP patients showed attenuated (smaller) responses. This points to different types of autonomic and brainstem dysfunction.
  • Verticality as a Differentiator: While all patients had reduced vertical saccade rates, the drop was most extreme in PSP, confirming the "vertical gaze palsy" characteristic of the disease even in an unstructured viewing environment.

Key Results

  • Saccade Frequency: Macro-saccade rates were significantly lower in patients (RBD: 1.74/s; PD: 1.51/s; MSA: 1.49/s; PSP: 1.13/s) than in controls (1.87–1.94/s).
  • Motor Correlation: Saccade frequency, amplitude, and the "rebound" speed of eye movements after a clip change were all negatively correlated with motor symptom severity (UPDRS-III scores, $\rho \approx -0.3$ to $-0.4$).
  • Pupil Dynamics: MSA patients had significantly greater pupil constriction/dilation than PSP patients ($P < 0.05$), providing a clear physiological divider between the two.
  • Fixation Duration: Because they made fewer saccades, PSP and PD patients had significantly longer fixation durations (staring at one spot longer) than controls.

Limitations and Future Work

The authors acknowledge a few hurdles:

  1. Sample Size: The PSP group was relatively small (n=10), which is common for this rarer condition but limits statistical power.
  2. Cross-sectional Design: The study shows that RBD patients have abnormalities, but it doesn't yet track them over time. Longitudinal studies are needed to see if specific eye-tracking "signatures" can predict when an RBD patient will convert to PD or MSA.
  3. Lack of Targets: Because the task was "free viewing," the researchers couldn't measure "saccade dysmetria" (how accurately the eye hits a specific target), which is a staple of clinical neurology.

Despite these, the study proves that a simple, 10-minute "passive" test—essentially just watching TV—can yield powerful biomarkers for complex brain disorders.

Summary updated

Abstract

RBD patients already have discrete but less pronounced saccade abnormalities than PD and MSA patients. Vertical gaze palsy and altered pupil control differentiate PSP from αSYN.
BibTeX
@article{habibi2022eye,
  author = {Mahboubeh Habibi and Wolfgang H. Oertel and Brian J. White and Donald C. Brien and Brian C. Coe and Heidi C. Riek and Julia E. Perkins and Rachel Yep and Laurent Itti and Lars Timmermann and Christoph Best and Elisabeth Sittig and Annette Janzen and Douglas P. Munoz},
  title = {{Eye tracking identifies biomarkers in α-synucleinopathies versus progressive supranuclear palsy}},
  journal = {Journal of Neurology},
  year = {2022},
  doi = {10.1007/s00415-022-11136-5},
}
Figure