Researchers at Mass General Brigham report that they have identified measurable biomarkers of tinnitus by tracking pupil dilation and subtle, involuntary facial movements that align with how distressing the condition feels to a person. Reported in Science Translational Medicine, the work could make placebo-controlled tinnitus treatment trials far more feasible by adding objective measures to a field that has largely relied on self-reported questionnaires.
“Imagine if cancer severity were determined by giving patients a questionnaire — this is the state of affairs for some common neurological disorders like tinnitus,” said corresponding author Daniel Polley, PhD, vice chair for basic science research and director of the Eaton-Peabody Laboratories at Mass Eye and Ear, part of the Mass General Brigham system. He said the team was surprised to find that sound-triggered facial movements not only occurred, but turned out to be one of the most informative indicators of tinnitus-related distress observed so far.
Tinnitus is typically described as persistent phantom sounds—such as ringing, buzzing, or clicking—heard in the absence of an external source. It affects about 12% of the general population and roughly a quarter of adults aged 65 and older. While many people can adapt and find it more annoying than debilitating, researchers estimate that around 15% experience symptoms severe enough to disrupt sleep, mental health, and everyday functioning. Until now, clinicians have had no objective method to reliably distinguish milder cases from disabling ones.
Alongside standard assessments of hearing and auditory brain function, the research team focused on the sympathetic nervous system—the body’s “fight, flight, or freeze” response—to look for outward signs of distress that may be easily overlooked. Pupil dilation is known to reflect arousal, and tiny involuntary facial movements can provide insight into how the brain evaluates potential threats.
The scientists hypothesized that people with debilitating tinnitus may remain in a chronic state of vigilance, reacting to ordinary sounds as though they are dangerous. To test this hypothesis, they recruited 97 participants with normal hearing: 47 people with varying levels of tinnitus and sound sensitivity, and 50 control participants without tinnitus.
Participants were video-recorded while listening to a range of sounds categorized as pleasant, neutral, or unpleasant and distressing, such as coughing fits, yelling, or a baby crying. Using AI-based analysis, the team detected rapid, subtle facial movements—small involuntary movements in areas such as the cheeks, eyebrows, and nostrils—that were associated with participants’ reported levels of tinnitus distress. Combining facial movement data with pupil dilation further improved the ability to predict severity.
People with severe tinnitus showed unusually large pupil dilation across all sound types, whether pleasant, neutral, or unpleasant, while their facial responses to those sounds were comparatively muted. By contrast, participants without tinnitus—or with less bothersome tinnitus—tended to show stronger pupil and facial reactions primarily to the most unpleasant sounds. The same measures also helped predict questionnaire scores related to hyperacusis (reduced tolerance to everyday sounds), although the predictions were less accurate than for tinnitus distress.
Polley said the approach could be practical because it does not depend on advanced brain imaging. If adapted for consumer-grade devices, similar measurements could be used in hearing clinics, in clinical trials as objective endpoints, and even by individuals outside medical settings.
The researchers noted a key limitation: to demonstrate the method clearly, they excluded many people with co-existing issues such as hearing loss, older age, or mental health conditions—factors often linked to more complex or severe tinnitus. Future studies will aim to include those higher-risk groups to see how well the biomarkers generalize.
The team is now exploring how these biomarkers could guide the development and testing of new treatments, including therapies that combine neural stimulation with immersive software environments intended to eliminate or substantially reduce the perceived loudness of tinnitus.
“These biomarkers get to the root of the distress,” Polley said. While brain imaging may reveal hyperactive regions associated with tinnitus, he argued, these outward signs reflect broader threat-evaluation systems operating outside their normal range—helping explain why the symptoms can be so disruptive.