TMS for Tinnitus: rTMS Protocols, Evidence, and Who Responds Best

At a Glance

Tinnitus — the perception of sound in the absence of an external acoustic source — affects roughly 10–15 % of adults worldwide. For about 1–2 %, it is severely disabling: disrupting sleep, concentration and emotional regulation with the same insistence as chronic pain. Standard audiological management (hearing aids, sound enrichment, CBT) addresses the distress but rarely reduces the perceived loudness itself.

Repetitive transcranial magnetic stimulation has emerged over the past two decades as the most studied brain-directed intervention for tinnitus, targeting the cortical hyperactivity that drives the phantom percept. The evidence base is now large enough to generate meta-analyses, though it still carries real limitations: small samples, heterogeneous protocols, and a signal that fades at six months. This article reviews the neuroscience, the protocols, the evidence hierarchy, and the clinical factors that predict who actually benefits.

Why TMS Targets the Auditory Cortex

The leading mechanistic account of chronic tinnitus is thalamo-cortical dysrhythmia (TCD): hearing loss or cochlear injury reduces afferent input to the auditory thalamus, which then slips into abnormal oscillatory firing. That dysrhythmic thalamic output reorganises the primary auditory cortex (A1, located in Heschl’s gyrus of the temporal lobe), generating self-sustained hyperactivity that the brain interprets as sound. Supporting evidence comes from functional imaging showing elevated spontaneous activity in A1 that correlates with tinnitus loudness, and from cortical mapping studies demonstrating tonotopic reorganisation at the frequencies lost to hearing damage.

The rationale for inhibitory rTMS follows directly: 1 Hz stimulation applied over A1 suppresses cortical excitability through long-term depression (LTD)-like mechanisms, temporarily reducing the pathological hyperactivity. Early case reports in 2003 showed measurable tinnitus suppression, and the field has built on that signal ever since (PMID 33934235).

The limitation of a pure-auditory-cortex model is that tinnitus also has strong attentional, emotional and mnemonic components. Limbic and prefrontal networks gate how much distress the signal causes, which is why patients with the same loudness rating differ wildly in disability. This observation is what motivated multisite protocols that target the dorsolateral prefrontal cortex (DLPFC) in addition to the auditory cortex.

rTMS Protocols: Standard, Multisite, and Emerging

Standard 1 Hz Temporoparietal Protocol

The workhorse protocol positions the coil over the left temporal cortex, typically using neuronavigation or the standard 10–20 EEG landmark (T3 or CP5), and delivers 1 Hz (inhibitory) stimulation at 100–110 % of resting motor threshold. Most published studies use 1,500–2,000 pulses per session over 10–15 consecutive daily sessions. Fourteen of sixteen RCTs included in a 2025 meta-analysis used this 1 Hz approach (PMID 40303608).

A right-sided variant has been tested when the dominant tinnitus percept lateralises to the right, but left-sided stimulation remains more studied and is the clinical default.

Multisite Sequential Protocols

Recognising the limbic-attentional component of tinnitus, several groups have added DLPFC stimulation before the auditory cortex component. In the approach reviewed by Marder et al., each session begins with either excitatory stimulation (10 Hz or intermittent theta-burst) to the left DLPFC or inhibitory stimulation to the right DLPFC, followed immediately by 1 Hz inhibitory stimulation over the auditory cortex (PMID 35665054). The rationale is that prefrontal priming modulates the top-down gating of tinnitus distress while the auditory cortex component addresses the bottom-up signal.

A systematic review of 22 studies found that sequential protocols using 1 Hz at the left DLPFC appeared superior to single-site auditory-cortex rTMS in the short term. For patients with comorbid depression — a common and undertreated pairing — sequential protocols showed particularly meaningful improvements in the tinnitus-only component, with TFI reductions near 28 % in the non-depressed subgroup.

Theta-Burst Stimulation

Continuous theta-burst stimulation (cTBS) produces LTD-like inhibition with far fewer pulses (600 pulses in 40 seconds) than 1 Hz protocols. Practical advantages include shorter session times and potentially stronger effects per pulse. Head-to-head data against 1 Hz for tinnitus remain limited; cTBS currently serves mainly as a protocol variant when session time is constrained.

What the Evidence Actually Shows

The 2025 Meta-Analysis

The most comprehensive evidence synthesis to date pooled 16 RCTs with 1,105 participants (600 rTMS, 505 sham). Active rTMS produced a statistically significant improvement on the Tinnitus Handicap Inventory: mean difference −11.54 (95 % CI −17.32 to −5.77, P < 0.00001). Visual Analogue Scale scores also favoured active rTMS (MD −1.85, P < 0.0001). However, the Tinnitus Questionnaire and loudness matching did not reach significance, and heterogeneity across studies was high (PMID 40303608).

Crucially, follow-up analysis separated short-term (up to 1 month post-treatment) from long-term (6 months) outcomes. At one month, the benefit remained significant (MD −10.98, P = 0.0008). At six months, it was not (MD −4.26, P = 0.17). This trajectory — meaningful short-term gain, uncertain long-term durability — is the central tension in the rTMS-for-tinnitus literature.

The Laryngoscope Neuromodulation Review

A 2024 systematic review and meta-analysis in Otolaryngology–Head & Neck Surgery (19 RCTs, N = 1,186) reached a partly discordant conclusion: rTMS showed significant long-term benefit on THI (−8.6, 95 % CI −11.5 to −5.7) but not short-term benefit, in contrast to the 2025 Frontiers meta-analysis (PMID 38353342). The discrepancy likely reflects different study selection, endpoint definitions and follow-up windows. Both reviews converge on a moderate effect size; the disagreement is about when that effect is most robust.

Interpreting “Clinically Significant”

A minimal clinically important difference (MCID) for the THI is approximately 7 points. The pooled effect of −11.54 exceeds that threshold, but the wide confidence interval means that a substantial proportion of treated patients experience smaller changes. At the individual level, roughly 40–60 % of patients in RCTs are classified as responders, defined as a ≥ 7-point THI reduction. The remaining 40–60 % are non-responders — a proportion that motivates rigorous patient selection.

Who Responds Best

Patient selection is probably the single most modifiable factor in rTMS outcomes for tinnitus. The consistent predictors of better response across multiple datasets:

Shorter tinnitus duration. Patients with symptoms for less than two years respond more reliably than those with decade-long histories. Cortical reorganisation deepens over time, and once large-scale tonotopic remapping is established, temporary suppression of A1 hyperactivity may not undo the structural changes (PMID 17903575).

Normal or near-normal hearing thresholds. Tinnitus driven primarily by central dysrhythmia rather than profound cochlear degeneration is a better target. Patients with severe sensorineural hearing loss show blunted responses, possibly because the auditory cortex hyperactivity in that group is a consequence of ongoing deafferentation rather than a reversible functional state.

Tonal character. Patients with a clearly tonal percept (a pure-tone pitch they can match to an external reference) respond better than those with complex noise or narrowband tinnitus, possibly because tonal tinnitus maps more cleanly to a focal cortical region that can be suppressed.

Laterality. Results are mixed, but several studies find that left or bilateral tinnitus responds better to left-sided stimulation than purely right-sided tinnitus does. When right dominance is present, contralateral (right hemisphere) targeting or bilateral protocols warrant consideration.

Absence of severe depression. Comorbid depression substantially reduces response rates in single-site auditory-cortex protocols. Sequential DLPFC-first protocols partially mitigate this, but severe treatment-resistant depression should be addressed in its own right before tinnitus-focused rTMS is initiated.

Safety Profile and What to Expect

Adverse events in controlled trials are both mild and comparable to sham: headache is the most common (7–9 % across both groups in the 2025 meta-analysis), and there is no difference in event rates between active rTMS and sham (OR 0.89, P = 0.67). Transient dizziness, nausea or brief discomfort at the coil site are occasionally reported and uniformly self-limiting. The absence of serious adverse events across more than 1,100 participants in RCTs is reassuring.

What patients should be told at intake:

1. The experience during stimulation is a tapping or clicking sensation at the scalp; most patients find it tolerable after the first session.
2. Some patients notice a temporary worsening in the first 1–2 sessions before improvement, consistent with the phasic nature of LTD induction.
3. Realistic expectations: a 20–40 % reduction in tinnitus annoyance is a good outcome; complete cessation is uncommon.
4. Maintenance sessions (every 4–8 weeks) may extend the benefit beyond the initial course. Optimal maintenance schedules are not yet established by RCT data.
5. Patients with cochlear implants, metallic intracranial implants, a history of seizures, or cranial surgery over the target region are excluded.

Integrating rTMS into a Tinnitus Workup

Before pursuing rTMS, a full audiological assessment, tinnitus characterisation (pitch matching, minimum masking level, THI baseline), and a qEEG or functional brain mapping are worth obtaining. Tinnitus with pulsatile features, asymmetric hearing loss, or a sudden onset requires vascular and imaging workup before any intervention.

At our centre, candidates who are most likely to benefit from rTMS share three characteristics: chronic but not decades-old tinnitus, an intact tonal percept, and a THI score above 38 (moderate or worse). Those with concurrent high-frequency hearing loss receive a hearing aid fitting in parallel — amplification alone sometimes reduces tinnitus distress substantially, and that gain should be captured regardless of neuromodulation.

For patients where tinnitus arises in the context of inflammatory or infectious pathology — reactivated Epstein-Barr virus, Lyme disease affecting the auditory nerve, or mould-related neuroinflammation — the neuromodulation component is secondary to treating the upstream cause. Attempting cortical suppression in an actively inflamed system is both less effective and logistically premature.

Related Articles

• TMS for Depression: Protocols, Evidence, and Patient Selection — the larger evidence base for auditory-adjacent cortical targets
• qEEG Brain Mapping: What It Shows and How We Use It — why individualized target mapping improves outcomes across TMS indications
• Neurofeedback for ADHD: Evidence, Protocols, and Patient Selection — complementary neuromodulation in patients with attentional comorbidities
• PBM for Brain Fog and Cognitive Decline — combining photobiomodulation with rTMS for neuroinflammatory presentations
• Vagus Nerve Stimulation for Inflammation and Chronic Illness — non-invasive VNS as an adjunct in patients with systemic inflammatory drivers of tinnitus