TMS Research: The Landmark Trials and What They Showed

The evidence for TMS in depression rests on a handful of sham-controlled trials, starting with O'Reardon 2007 and the NIMH-funded George 2010 study, extended by Levkovitz 2015 for deep TMS, Blumberger 2018 for theta-burst stimulation and Cole 2022 for accelerated SAINT treatment. In these trials, active TMS beat sham, but by a margin much smaller than clinic marketing suggests: active-versus-sham differences in remission were typically about 9 to 18 percentage points. Several well-designed trials were negative, and they matter as much as the positive ones.

This page summarizes each landmark study with its size, design, how it defined success and what it actually found, followed by the meta-analyses that pool them.

How to read a TMS trial

Three questions sort out most of the confusion around TMS numbers:

• Was there a sham (placebo) arm? Sham-treated patients improve substantially; a meta-analysis of 61 trials found a within-group sham effect size of about 0.8. Without a sham arm, improvement cannot be attributed to the magnetic stimulation itself.
• Which patients were counted? Intent-to-treat (ITT) analyses count everyone randomized; per-protocol or completer analyses count only those who finished, which usually inflates results.
• How was success defined? Response and remission cutoffs differ between scales and studies. The definitions are explained in response vs remission.

The landmark trials at a glance

The trials that established TMS (2007-2010)

O'Reardon et al., 2007

This was the pivotal NeuroStar trial. It randomized 301 medication-free outpatients who had not benefited from prior treatment to active or sham 10 Hz TMS, using 3,000 pulses per session at 120% of motor threshold, five days a week for 4 to 6 weeks. Its primary endpoint, MADRS change at week 4, favored active TMS only after a post hoc statistical correction for a baseline difference between groups. Active TMS was superior on the HAMD17 and HAMD24 at weeks 4 and 6, and remission was described as "approximately twofold higher" at week 6, significant on two of three scales. Only 4.5% dropped out for adverse events, mostly transient scalp discomfort.

Exact response and remission percentages from this trial are widely quoted online, but they could not be verified against the full text for this review, so they are not repeated here.

George et al., 2010 (OPT-TMS)

The NIMH-funded trial added independent funding and a sham designed to improve blinding. It enrolled 199 antidepressant-free outpatients at four university sites and used an active sham, with scalp electrodes to mimic the sensation of real stimulation. In the intent-to-treat sample, 14.1% of the active group remitted versus 5.1% with sham, an odds ratio of 4.2 and a number needed to treat of 12. Most remitters had low levels of prior treatment resistance. In the later open-label phase, about 30% of participants remitted.

Those numbers look modest, and they are. They are also honest: an ITT, sham-controlled, independently funded result. Many later patient-facing claims are several times higher because they come from different kinds of studies.

Deep TMS and the per-protocol question

Levkovitz 2015 tested the BrainsWay H1 coil as monotherapy in outpatients aged 22 to 68 who had failed one to four antidepressants. It enrolled 233 people; the ITT set was 212 and the per-protocol set 181. At week 5, ITT response was 37.0% vs 27.8% and remission 30.4% vs 15.8%; the per-protocol figures were higher, 38.4% vs 21.4% response and 32.6% vs 14.6% remission. The abstract's headline numbers came from the per-protocol population, and the ITT primary endpoint, a slope comparison, was not statistically significant (p=0.0578). The categorical ITT outcomes still favored active treatment, and BrainsWay's depression clearance followed in January 2013.

This trial is a good illustration of why the analysis population matters. The deep TMS vs standard TMS page covers how the H-coil differs.

Negative trials worth knowing

Yesavage et al., 2018 (Veterans Affairs)

This nine-site VA trial gave 10 Hz rTMS at 120% MT and 4,000 pulses per session to veterans with treatment-resistant depression. Remission (HRSD ≤10) was 40.7% with active treatment and 37.4% with sham, with no significant difference. Roughly 4 in 10 patients remitted in both groups, a reminder of how large the sham response can be. The VA/DoD guideline cites this trial and rates its own recommendation for rTMS as "weak for", with very low confidence in the evidence.

Croarkin et al., 2021 (adolescents)

The only large sham-controlled trial of 10 Hz TMS in adolescents enrolled ages 12 to 21 at 13 sites and found no difference: response 41.7% vs 36.4% and remission 29.2% vs 29.0%. There were no new safety signals. Later FDA clearances for ages 15 to 21 as an adjunct relied on open-label registry data, and this negative trial does not appear in the supporting tables of the first adolescent 510(k) summary. If you are thinking about suicide or are in crisis, call or text 988 (the 988 Suicide & Crisis Lifeline) or go to the nearest emergency room.

BRIGhTMIND, 2024

This five-center trial compared connectivity-guided iTBS (n=128) with MRI-navigated standard rTMS (n=127). There was no difference over 26 weeks, and two serious adverse events (mania, psychosis) were possibly TMS-related. It did not test TMS against sham, so it says nothing against TMS itself, but it tempered enthusiasm for fMRI targeting; see finding the TMS target.

Faster protocols: theta burst and accelerated TMS

THREE-D (Blumberger et al., 2018)

THREE-D randomized 414 adults with treatment-resistant depression at three Canadian hospitals to standard 10 Hz rTMS or intermittent theta-burst stimulation (iTBS), five days a week for 4 to 6 weeks, with patients on stable antidepressants. Depression scores fell almost identically, from 23.5 to 13.4 with 10 Hz and from 23.6 to 13.4 with iTBS, meeting the non-inferiority criterion. According to FDA's later SAINT summary, the iTBS arm had 49% response and 32% remission. iTBS sessions took about 3 minutes instead of 37.5, though iTBS was rated slightly more painful.

Because there was no sham arm, THREE-D shows that iTBS is about as good as 10 Hz, not how much better either is than placebo. It was the basis for iTBS clearance in 2018. See the theta-burst explainer.

SAINT and the accelerated trials that followed

The SAINT story has two parts. Cole 2020 was an open-label study of 22 people in which 19 of 21 completers (90.5%) remitted. Cole 2022 was a double-blind sham-controlled trial stopped early for superiority at its interim analysis: 29 people were treated, and remission was 11 of 14 (78.6%) with active treatment versus 2 of 15 (13.3%) with sham. FDA's summary noted that only 19 of 29 completed the blinding questionnaire, leaving "uncertainty" about blinding, and that fatigue and back or neck pain were common. No independent sham-controlled replication was found in the literature search for this guide, so the large effect needs confirmation; details are in the SAINT guide.

Larger accelerated trials have since appeared. In Ramos 2025, 100 patients received 45 accelerated iTBS sessions over 15 weekdays without neuronavigation; response was 52% vs 22% and remission 34% vs 16% versus sham. In Hanlon 2026, a BrainsWay-sponsored non-inferiority trial of 104 adults, accelerated H1-coil iTBS was non-inferior to standard deep TMS on HDRS-21 change at week 6 and reached remission sooner (21 vs 28 days), but its remission rate was numerically lower (78.0% vs 87.5%) and adverse events were more common. The accelerated TMS overview puts these protocols side by side.

Beyond depression: OCD and smoking

Carmi 2019 randomized 99 people with OCD at 11 centers to deep TMS or sham over the medial prefrontal and anterior cingulate cortex, delivered after individualized symptom provocation, daily for 6 weeks. Response, defined as at least a 30% drop in Y-BOCS score, was 38.1% vs 11.1% after treatment and 45.2% vs 17.8% one month later. FDA granted BrainsWay's De Novo request for OCD, as an adjunct treatment for adults, in August 2018. The site's OCD treatment page covers the wider approach to OCD care.

Zangen 2021 randomized 262 chronic smokers to deep TMS or sham targeting the lateral prefrontal and insular cortices. Four-week continuous abstinence through week 18, verified by urine cotinine, was 19.4% vs 8.7% in the ITT analysis, and the authors state it led to FDA clearance as an aid to smoking cessation. A patient guide is at TMS to quit smoking.

Real-world studies: useful, but different

Open-label studies show how TMS performs in routine practice, but they cannot separate treatment effect from placebo response or selective reporting. Carpenter 2012 followed 307 patients at 42 U.S. practices who had failed a mean of 2.5 antidepressant trials; clinician-rated response was 58.0% and remission 37.1%, while patient-rated remission ranged from 26.5% to 28.7%. Sackeim 2020, a NeuroStar registry of 7,759 participants at 103 sites, reported response from 58% to 83% and remission from 28% to 62%, depending on the sample and scale. Those registry figures are open-label, mostly self-report, from voluntarily reporting sites and partly completers only, and should not be set beside RCT rates as if they were comparable.

The same caution applies to the BrainsWay post-marketing analysis often quoted as "81.6% response, 65.3% remission": it was open-label, used varying scales, and counted only patients with data on at least one scale after 30 sessions. How long benefits last in these cohorts is covered in how long TMS results last.

What the meta-analyses say, and what it adds up to

• Brunoni 2017, a network meta-analysis of 81 RCTs and 4,233 patients, found high-frequency rTMS more effective than sham for response (odds ratio 3.07), along with bilateral, low-frequency, priming and theta-burst protocols. It did not find accelerated, synchronized or deep rTMS better than sham, but that evidence base was small at the time.
• Zhang 2026 pooled 141 RCTs and 10,587 participants; every major protocol beat sham, but only 3 trials were at low risk of bias and confidence in most comparisons was low or very low.
• Eren and Takım 2026 pooled 30 sham-controlled comparisons (1,850 patients): response odds ratio 4.78 and remission odds ratio 3.55, rated high certainty, with larger effects in single-blind than double-blind trials.
• Santos, Brunoni and colleagues 2026 pooled 15 sham-controlled accelerated iTBS trials (704 patients): response risk ratio 1.99 (NNT 5) and remission risk ratio 1.77 (NNT 10).

Guideline bodies read this evidence consistently. CANMAT 2023 rates rTMS as first-line for treatment-resistant depression with Level 1 evidence for acute efficacy. The international IFCN guideline rates high-frequency left DLPFC rTMS as Level A, definite efficacy.

Taken together, TMS reliably outperforms sham for depression across many trials and several protocols. The advantage is real but moderate, sham response is large, and the most impressive single numbers come from small, early-stopped or uncontrolled studies. For a patient, that means a reasonable chance of meaningful benefit, not a guarantee. Comparison with other options, including ECT, is covered in TMS vs ECT; the full timeline is in a short history of TMS.

For the complete set of evidence pages, start at the TMS guide. To discuss whether TMS fits your history, see the TMS therapy page.

Frequently asked questions

What is the single most important TMS trial?

There is no single one. O'Reardon 2007 supported the first clearance, but the NIMH-funded George 2010 trial is often seen as the cleanest test because it was independent, used an active sham and reported intent-to-treat remission.

Why do TMS success rates in trials look lower than clinic websites?

Trials compare against sham and count everyone randomized. Clinic and registry figures are usually open-label, may count only completers and often use self-report scales, all of which push numbers up.

Has any major TMS trial failed?

Yes. The VA trial in veterans and the adolescent NeuroStar trial found no difference from sham, and BRIGhTMIND found connectivity targeting no better than standard targeting. Negative trials help define where TMS works best.

Is there a Cochrane review of TMS for depression?

A search of PubMed for this review did not find a Cochrane review of rTMS for adult depression, so none is cited here. Large network meta-analyses from 2017 and 2026 are the main pooled sources.

Are SAINT's results proven?

SAINT's sham-controlled trial was small (29 treated) and stopped early, and FDA noted uncertainty about blinding. The results are promising, but no independent sham-controlled replication was found in the literature search for this guide.

Sources

1. Efficacy and safety of transcranial magnetic stimulation in the acute treatment of major depression: a multisite randomized controlled trial. Biological Psychiatry, 2007. pubmed.ncbi.nlm.nih.gov
2. Daily left prefrontal transcranial magnetic stimulation therapy for major depressive disorder: a sham-controlled randomized trial. Archives of General Psychiatry, 2010. pubmed.ncbi.nlm.nih.gov
3. Transcranial magnetic stimulation (TMS) for major depression: a multisite, naturalistic, observational study of acute treatment outcomes in clinical practice. Depression and Anxiety, 2012. pubmed.ncbi.nlm.nih.gov
4. Efficacy and safety of deep transcranial magnetic stimulation for major depression: a prospective multicenter randomized controlled trial. World Psychiatry, 2015. pmc.ncbi.nlm.nih.gov
5. Effect of Repetitive Transcranial Magnetic Stimulation on Treatment-Resistant Major Depression in US Veterans: A Randomized Clinical Trial. JAMA Psychiatry, 2018. pubmed.ncbi.nlm.nih.gov
6. VA/DoD Clinical Practice Guideline for the Management of Major Depressive Disorder. U.S. Department of Veterans Affairs / Department of Defense, 2022. healthquality.va.gov
7. Effectiveness of theta burst versus high-frequency repetitive transcranial magnetic stimulation in patients with depression (THREE-D): a randomised non-inferiority trial. The Lancet, 2018. pubmed.ncbi.nlm.nih.gov
8. 510(k) Summary K220177 – Magnus Neuromodulation System with SAINT Technology. U.S. FDA, 2022. accessdata.fda.gov
9. 510(k) Summary K173620 – MagVita TMS Therapy System w/Theta Burst Stimulation. U.S. FDA, 2018. accessdata.fda.gov
10. Efficacy and Safety of Deep Transcranial Magnetic Stimulation for Obsessive-Compulsive Disorder: A Prospective Multicenter Randomized Double-Blind Placebo-Controlled Trial. American Journal of Psychiatry, 2019. pubmed.ncbi.nlm.nih.gov
11. De Novo Classification Request DEN170078 (BrainsWay Deep TMS for OCD). U.S. FDA, 2018. accessdata.fda.gov
12. 510(k) K122288 – BrainsWay Deep TMS System. U.S. FDA, 2013. accessdata.fda.gov
13. Clinical outcomes in a large registry of patients with major depressive disorder treated with Transcranial Magnetic Stimulation. Journal of Affective Disorders, 2020. pubmed.ncbi.nlm.nih.gov
14. Left prefrontal transcranial magnetic stimulation for treatment-resistant depression in adolescents: a double-blind, randomized, sham-controlled trial. Neuropsychopharmacology, 2021. pubmed.ncbi.nlm.nih.gov
15. 510(k) Summary K231926 – NeuroStar Advanced Therapy System. U.S. FDA, 2024. accessdata.fda.gov
16. Repetitive transcranial magnetic stimulation for smoking cessation: a pivotal multicenter double-blind randomized controlled trial. World Psychiatry, 2021. pubmed.ncbi.nlm.nih.gov
17. Stanford Accelerated Intelligent Neuromodulation Therapy for Treatment-Resistant Depression. American Journal of Psychiatry, 2020. pubmed.ncbi.nlm.nih.gov
18. Stanford Neuromodulation Therapy (SNT): A Double-Blind Randomized Controlled Trial. American Journal of Psychiatry, 2022. pubmed.ncbi.nlm.nih.gov
19. Connectivity-guided intermittent theta burst versus repetitive transcranial magnetic stimulation for treatment-resistant depression: a randomized controlled trial. Nature Medicine, 2024. pubmed.ncbi.nlm.nih.gov
20. Accelerated Theta-Burst Stimulation for Treatment-Resistant Depression: A Randomized Clinical Trial. JAMA Psychiatry, 2025. pmc.ncbi.nlm.nih.gov
21. Accelerated TMS with the H1-coil for depression: A multisite, randomized non-inferiority trial. Brain Stimulation, 2026. pubmed.ncbi.nlm.nih.gov
22. Deep TMS H1 Coil treatment for depression: Results from a large post marketing data analysis. Psychiatry Research, 2023. pubmed.ncbi.nlm.nih.gov
23. Repetitive Transcranial Magnetic Stimulation for the Acute Treatment of Major Depressive Episodes: A Systematic Review With Network Meta-analysis. JAMA Psychiatry, 2017. pubmed.ncbi.nlm.nih.gov
24. Repetitive Transcranial Magnetic Stimulation for Major Depressive Disorder: A Systematic Review and Network Meta-Analysis. Journal of Evidence-Based Medicine, 2026. pubmed.ncbi.nlm.nih.gov
25. Protocol-level moderators of antidepressant efficacy in prefrontal rTMS: A sham-controlled meta-analysis of randomized trials. Journal of Affective Disorders, 2026. pubmed.ncbi.nlm.nih.gov
26. Accelerated intermittent theta-burst stimulation for major depressive episodes: systematic review and dose-response meta-analysis. Brain Stimulation, 2026. pubmed.ncbi.nlm.nih.gov
27. A systematic review and meta-analysis on placebo response to repetitive transcranial magnetic stimulation for depression trials. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 2018. pubmed.ncbi.nlm.nih.gov
28. Canadian Network for Mood and Anxiety Treatments (CANMAT) 2023 Update on Clinical Guidelines for Management of Major Depressive Disorder in Adults. Canadian Journal of Psychiatry, 2024. pmc.ncbi.nlm.nih.gov
29. Evidence-based guidelines on the therapeutic use of repetitive transcranial magnetic stimulation (rTMS): An update (2014-2018). Clinical Neurophysiology, 2020. pubmed.ncbi.nlm.nih.gov

Educational information, not medical advice. Whether TMS or any treatment fits you is a decision to make with a psychiatrist who knows your history.