Could patterns of somatic tinnitus modulation help predict treatment response?
The question in plain language
Many people can temporarily change their tinnitus — its loudness, pitch, or character — by moving or tensing the jaw, neck, or head, or by pressing on certain muscles. This is common: studies report that roughly 56–80% of people with tinnitus show some modulation on formal maneuver testing.
Researchers have already asked the obvious question: does simply HAVING this ability predict who will respond to treatments that engage the body–hearing connection (bimodal stimulation devices, cervical or jaw-focused physiotherapy, related neuromodulation)? So far, that simple yes/no characteristic has generally not been a useful predictor.
A narrower possibility remains open: the number of maneuvers that change the tinnitus, the direction of the change (louder vs quieter), the size of the change, and how consistently the same response occurs may carry more information than mere presence. This remains unproven.
Evidence for the refined question
- Shore/Jones 2023 (bisensory device RCT): in a supplementary analysis, participants for whom MORE maneuvers changed their tinnitus showed greater questionnaire (TFI) improvement during active treatment — and not during control or washout phases. Important caveats: this was a post-hoc correlation (r ≈ 0.3), inside a trial that only enrolled people who could modulate, so it cannot say anything about non-modulators.
- Rocha & Sanchez 2012 (trigger-point therapy RCT): the DIRECTION of modulation at baseline mattered — people whose tinnitus temporarily decreased when a trigger point was pressed were more likely to have lasting relief after treatment (p=0.002).
- Related phenotype signals: in cervical physiotherapy studies, self-reported co-variation of tinnitus with neck complaints — and history-based somatic-association scores — have predicted response better than maneuver testing itself.
These are associations, not demonstrations of prediction. None comes from a study designed and powered to test the question.
Evidence against the simple version
The binary form of this idea — "can modulate" vs "cannot" — has been tested directly and has generally not held up:
- van der Wal 2020 (N=101, multidisciplinary orofacial treatment): explicitly expected stronger somatic influence to mean larger treatment effects; ability to modulate was not identified as a prognostic indicator. Shorter tinnitus duration, painful jaw-muscle palpation and other characteristics predicted outcome instead.
- Spencer 2022 (N=29, bimodal auditory + electrical stimulation): baseline modulation ability did not predict outcome — though with only six modulators, the test was severely underpowered.
- Lee 2020 (N=81, retrospective clinic series): modulation status was unrelated to improvement; only age predicted outcome.
- Low 2017 (N=27, electroacupuncture): no overall outcome difference between somatic and non-somatic groups (a post-hoc signal favored people whose maneuver responses were consistent).
- Vanneste 2010 (N=240, electrical stimulation at C2): every participant was selected for somatosensory modulation — yet only 17.9% responded. Being a modulator clearly does not guarantee response to a somatosensory treatment.
With modulation present in roughly 56–80% of patients, the simple yes/no trait appears too common and too weakly discriminating to work as a biomarker by itself. That is a finding about the simple version — it does not settle the refined, graded question above.
An unanswered Lenire/TENT-A1 question
The first large Lenire trial (TENT-A1, 326 participants) collected baseline somatic-modulation data using a standardized 25-maneuver assessment performed by a physiotherapist, and its published protocol prespecified an efficacy analysis by somatic-tinnitus status, to be reported in a subsequent publication. In our literature audit (through September 2026), we did not identify a published report of the prespecified somatic-subgroup analysis. The existing dataset may therefore contain information relevant to this question.
Refined hypothesis (technical wording)
In interventions engaging auditory–somatosensory integration, graded features of baseline somatic modulation — including the number of effective maneuvers, direction of induced change, magnitude of change, and test–retest consistency — may predict treatment response more effectively than the simple presence or absence of modulation.
This has not yet been established in an adequately powered, prespecified prospective test.
How researchers could test this
Lowest-cost approach: analyze existing datasets that already collected standardized somatic-modulation measurements — prioritizing TENT-A1, if the relevant baseline data and outcomes are available.
Next step: add a standardized baseline modulation battery to ongoing bimodal or somatosensory tinnitus trials, and prespecify the interaction analysis before outcomes are known.
Definitive test: a prospective controlled study enrolling participants across the full modulation spectrum, with a standardized baseline maneuver battery, a prespecified modulation × treatment interaction, and a percept-level tinnitus outcome — not only distress/severity questionnaires.
What would disprove this?
The refined hypothesis would be weakened or rejected if: graded modulation features fail to predict response in adequately powered studies; the same apparent relationship appears in matched control groups (indicating a non-specific effect); test–retest modulation measurements prove too unreliable to carry a signal; stronger baseline predictors fully account for the association; or independent datasets do not reproduce it. The hypothesis is falsifiable, and several of these tests are inexpensive.
Why this question matters
If a simple baseline test could identify who is more likely to respond to a particular tinnitus treatment, future trials could become more targeted and less likely to dilute subgroup effects. But current evidence is not strong enough to use somatic modulation for clinical treatment selection.
Somatic-modulation characteristics are being investigated as potential response predictors, but they are not established tools for selecting tinnitus treatment. Nothing on this page is a treatment recommendation.
Novelty and precedent
This is not a new idea in its broad form. Researchers have proposed somatic testing for treatment selection since at least 2007 (Levine; later Won 2013), and binary somatic-modulation predictors have been tested before, with mostly null results. The remaining research question concerns whether more detailed, graded modulation characteristics carry predictive information.
Sources
- van der Wal AC, et al. Prognostic indicators for positive treatment outcome after multidisciplinary orofacial treatment in patients with somatosensory tinnitus. Front Neurosci 2020. PMID 33041758
- Spencer S, et al. Bimodal (auditory and electrical) stimulation for tinnitus — feasibility and predictor analysis. Front Neurosci 2022. PMID 36090280
- Lee HY, et al. Retrospective analysis of somatic modulation and treatment outcome in a tinnitus clinic. J Int Adv Otol 2020. PMID 32784160
- Low WK, et al. Electroacupuncture in somatic vs non-somatic tinnitus. Med Acupunct 2017. PMID 28736590
- Vanneste S, et al. Transcutaneous electrical nerve stimulation (C2) in 240 patients selected for somatosensory modulation. Exp Brain Res 2010. PMID 20505927
- Jones GR, …, Shore SE. Bisensory stimulation randomized clinical trial (incl. supplementary eFigure 7). JAMA Netw Open 2023. PMID 37266943
- Rocha CB, Sanchez TG. Myofascial trigger-point deactivation RCT with modulation-direction analysis. Braz J Otorhinolaryngol 2012. PMID 23306563
- D'Arcy S, et al. TENT-A1 trial protocol (prespecified somatic-subtype analysis). BMJ Open 2017. PMID 29074518
- Levine RA, et al. Somatic testing and treatment-subgroup proposal. Prog Brain Res 2007. PMID 17956783
- Won JY, et al. Modulation-characteristics candidate-selection proposal. Audiol Neurootol 2013. PMID 23881235
Summaries are original to Tinnitus Evidence, written from the sources above during an adversarial precedent audit (September 2026); they are not publisher abstracts.