TMS has not been shown to help tinnitus in a lasting way: earlier trials found a modest benefit that was gone by six months, and a 2026 meta-analysis of 8 trials found no significant effect.
Transcranial magnetic stimulation (TMS) uses pulsed magnetic fields, delivered through a coil held near the scalp, to non-invasively influence electrical activity in a targeted area of the brain. For tinnitus, it's aimed at the auditory cortex specifically, based on the theory that overactive neural firing there is part of what generates the perception of tinnitus.
How it's done
Repetitive TMS (rTMS) — repeated pulses delivered in a session, as opposed to single pulses used for some diagnostic purposes — is the form studied for tinnitus. A typical protocol involves multiple sessions over one to several weeks, with the coil positioned over the auditory cortex (usually on the temporal or temporoparietal region), delivered by a trained clinician. It's non-invasive and doesn't require surgery or implantation, distinguishing it from more involved neuromodulation approaches.
What the evidence shows — and where it gets murkier
TMS for tinnitus has been studied fairly extensively, and the honest summary is: a modest short-term benefit in earlier analyses, no demonstrated durability, and a null result in the newest.
- A meta-analysis of 29 randomized trials (1,228 patients) found rTMS produced significant improvement compared to sham (fake) stimulation on standard tinnitus severity measures.
- A more recent meta-analysis of 16 randomized trials (1,105 patients) confirmed rTMS outperformed sham treatment on tinnitus handicap scores, and found a positive effect on tinnitus impact at one month — but did not find a significant benefit sustained at six months, and found no significant immediate effect on some of the other tinnitus questionnaires used across the studies.
- A separate trial specifically comparing active versus sham TMS found the active treatment group had significant improvement in both symptom severity and standardized handicap scores, while the sham group showed no improvement (and in one measure, slightly worsened) — supporting a real, not purely placebo, effect in that trial specifically.
Researchers reviewing the field describe a consistent pattern across studies: short-term efficacy shows up fairly reliably, but questions remain about how long benefits last, and there's substantial variability between studies in the specific TMS equipment, coil placement, pulse frequency, and treatment protocol used — differences that may explain some of the inconsistency in results, but that also mean "TMS for tinnitus" isn't yet a single standardized treatment the way, say, a specific medication dose would be.
The cleanest negative result this treatment has
A 2026 systematic review and meta-analysis in The Laryngoscope pooled randomised trials of neuromodulation for chronic tinnitus published between 2015 and 2025. For repetitive TMS it found 8 randomised trials, 432 participants, and a pooled effect of:
From the sourceSMD −0.15 (95% CI −0.37 to 0.07), I² = 0%.
Both halves of that matter. The interval comfortably crosses zero, so no effect was demonstrated. And I² = 0% means the eight trials did not disagree with each other — this is not the familiar picture of a promising treatment obscured by noisy studies. It is eight trials pointing at the same place, and that place is nothing.
That is a different kind of negative from the one the same paper reported for tDCS, where the interval also crossed zero but the trials disagreed (I² = 51%). A result with high heterogeneity leaves room for the possibility that the treatment works in some population the pooling obscured. A result with zero heterogeneity and a null estimate leaves considerably less.
It does not settle the question by itself — earlier meta-analyses reported short-term benefit, and the review's own recommendation is for standardised, adequately powered trials with harmonised protocols. It is why the treatment comparison still grades TMS as Limited rather than Evidence against. But anyone being offered TMS for tinnitus should know this analysis exists, and that the AAO-HNS guideline recommends against it for routine use.
The same paper found one modality that did hold up, and it was not this one: bimodal stimulation, with reductions often of 10–20 points on the THI or TFI and durability to twelve months.
What Cochrane said, and how well it has aged
The Cochrane review of rTMS for tinnitus, published in 2011, found five trials comprising 233 participants — and notes that each study used a different device delivering different waveforms at different frequencies, which is most of the problem with summarising this field. The trials were small but generally at low risk of bias.
On quality of life, exactly one study showed a statistically significant improvement in Tinnitus Handicap Inventory scores, at four months, for low-frequency rTMS against sham. Two other studies testing the same frequency showed no significant improvement. The review adds a caution that generalises well beyond this topic: that single positive finding should be read in the context of the many different variables the study authors recorded at many different points in time — which is to say, with that many chances to find something, finding one thing is not surprising.
Its conclusion: there is very limited support for the use of low-frequency rTMS, resting on a single study, a single outcome measure, and a single point in time. On safety, rTMS appeared safe in the short term, with insufficient data to say anything about the long term.
Fifteen years on, that has aged well. Everything since has pointed the same way — a modest, inconsistent, short-lived effect that depends heavily on how and where the stimulation is applied. The 2025 meta-analysis below finds a benefit at one month and none at six; the 2026 null reported above finds essentially nothing. Cochrane's "very limited support" remains an accurate description.
The effect that shows up at one month and is gone by six
A 2025 meta-analysis in Frontiers in Neuroscience pooled 16 randomised controlled trials and 1,105 patients with chronic tinnitus, comparing rTMS against sham. Its results are more specific than either a yes or a no, and the specificity is the useful part.
Where it found an effect: rTMS beat sham on the Tinnitus Handicap Inventory and on visual analogue scale ratings, with a positive effect on THI at one month.
Where it did not:
- At six months, the effect was gone. The analysis did not find a positive long-term effect.
- No significant immediate effect on the Tinnitus Questionnaire or on loudness matching.
So of four ways of measuring, two moved and two did not, and the one that moved clearly had faded by half a year.
That is a coherent picture rather than a contradictory one, and it fits the null reported above: a real but modest short-term effect, measured differently by different instruments, that does not persist. A treatment whose benefit has disappeared by six months is a different proposition from one that works, particularly for a condition measured in years — and it is a large part of why this sits at a C rather than higher.
The authors' own conclusion is that rTMS has some efficacy and that more RCTs with larger samples and longer follow-up are needed. That is a request for evidence, not a finding of durable benefit.
If you are considering it
TMS works through a different mechanism from sound-based or psychological approaches — modulating brain activity directly rather than working through the ear or through thought patterns — which is why it keeps being tried when sound therapy, TRT or CBT have not provided enough relief. Set expectations by the evidence rather than the mechanism: at most a modest, short-lived improvement, from protocols that still vary between providers and research centres. It is generally well tolerated, but it needs specialised equipment and a trained provider, typically at an academic medical centre or specialised clinic rather than a general audiology practice.
Where the professional guidelines land
The American Academy of Otolaryngology–Head and Neck Surgery's clinical practice guideline on tinnitus issues a recommendation against transcranial magnetic stimulation for routine treatment of persistent, bothersome tinnitus. The 2024 VA/DoD tinnitus guideline is softer: it finds insufficient evidence to recommend for or against repetitive TMS.
The AAO-HNS guideline dates from 2014 and research has continued since, which is why this article covers the newer work rather than stopping at the guideline. But a professional body having examined the evidence and recommended against it is the baseline any newer result has to improve on, and it belongs in the picture before deciding whether to pursue TMS privately. Neurofeedback occupies a similar position — an approach with a real research literature, and a wide gap between what is being studied and what is being sold.