There is no objective test for tinnitus yet: diagnosis still relies on questionnaires, though a 2025 study of 47 people with tinnitus and sound sensitivity and 50 without found pupil and facial responses to sound predicted questionnaire scores. Dozens of blood, saliva and urine markers are also under study.
Biomarker research asks a foundational question: can tinnitus be measured objectively at all, rather than relying entirely on what a patient reports? AI and machine learning in tinnitus care covers a closely related line of work.
Why this gap matters so much
Tinnitus has no blood test or scan that confirms it — diagnosis and severity assessment rely almost entirely on subjective questionnaires like the Tinnitus Handicap Inventory, described in how tinnitus is measured. This isn't just an inconvenience: without an objective measure, it's difficult to run properly blinded, placebo-controlled treatment trials, since researchers have no way to independently verify a treatment's effect beyond what patients report — and patient-reported outcomes are exactly the kind of measure most vulnerable to placebo response and expectation effects, a theme what happens if you do nothing puts numbers on.
A promising 2025 finding: pupils and faces
A study published in Science Translational Medicine involved 97 adults: 47 with tinnitus and sound sensitivity, and 50 without. It found that pupil dilation and subtle, involuntary facial movements — measured by video analysis while participants listened to pleasant, neutral and distressing sounds — predicted individual scores on the Tinnitus Handicap Inventory and on a hyperacusis questionnaire. People with severe tinnitus showed unusually wide pupil dilation across all sound types, while their facial movement responses were blunted compared to people without tinnitus, who instead showed more normal, exaggerated facial responses specifically to unpleasant sounds. The researchers describe this as evidence of body-wide threat-evaluation systems operating outside their normal range in severe tinnitus — meaning the nervous system is treating ordinary sounds as threats more broadly than just within auditory brain regions.
The same study measured central auditory gain, the brain's amplification of incoming sound that the leading model of tinnitus centres on. It was raised in the group with tinnitus and sound sensitivity, but it did not track how severe their symptoms were. The measures that did were the body's emotional responses to sound — in the authors' words, signatures of disrupted affective sound processing — rather than how much the auditory system amplifies it.
The broader biomarker landscape
Beyond the pupil and face finding, a 2021 systematic review of 49 studies identified 58 candidate biomarkers across blood, saliva, and urine samples, spanning eight categories including metabolic, hemostatic, inflammatory, endocrine, immunological, neurologic, and oxidative markers — reflecting how many different body systems have been investigated for a tinnitus signature. Separately, machine learning is being applied to brain-activity and hearing-test data to tell people with tinnitus from people without it, so far with accuracy figures from research samples rather than from new patients in a clinic.
Why this remains research, not a clinical tool
Despite the promise, significant limitations remain across this body of work: low spatial resolution in some measurement techniques, small sample sizes, inconsistent methodology between studies, and — critically — a lack of validation for using any of these markers to diagnose tinnitus in an individual patient rather than distinguishing group averages in a research setting. Standardized protocols, larger validation studies, and demonstrated clinical utility in prospective trials are all still needed before any of these become something your own audiologist could actually use.
Why this would matter so much if it succeeds
An objective biomarker wouldn't just be a diagnostic convenience — it would fundamentally change what's possible in tinnitus treatment research. Trials could be properly blinded and placebo-controlled in ways that are currently much harder to achieve, treatments could be evaluated against a measurable outcome rather than relying entirely on self-report, and patients with normal hearing tests but real tinnitus (covered in the article on hidden hearing loss) could potentially get objective confirmation of what they're experiencing.
What this means practically
- No blood test, scan, or objective biomarker for tinnitus exists in clinical practice today — anyone claiming otherwise is ahead of where the actual research stands
- The 2025 pupil and facial-movement findings are promising as a research direction, not yet a diagnostic tool available anywhere
- This research area is worth watching over the coming years, given how directly an objective measure would improve the quality of future treatment trials
- Current tinnitus diagnosis and management continues to rely on subjective, questionnaire-based approaches — that isn't expected to change imminently