Labyrinth Psychiatry Group · Published · Clinical review by Dr. Salvatore G. Savatta pending
The short answer
For depression that has not improved with medication, studies of everyday clinic patients found that 33% to 58% responded to TMS and 20% to 37% reached remission, depending on the study and the scale used. A 42-practice US study of everyday care (307 patients) found 58% response and 37% remission as rated by clinicians. Response means your symptom score drops by at least half; remission means it drops to a minimal level. In one large clinic, patients generally began reporting improvement within the first week (UCLA Health), and of those who respond, 62.5% were still responding a year later.
What “success” means: response and remission
TMS studies do not ask “did it work?” They measure depression on a rating scale before, during and after the course, and report two numbers.
- Response means your score fell by at least half from where it started. That is the usual definition in TMS research (UCLA, 2023; Toronto, 2018). You feel clearly better, but some symptoms may remain.
- Remission means your score fell below a set line, into the range where symptoms are minimal. It is the stricter goal, so remission rates are always lower than response rates.
Which scale is used changes the answer. In the 42-practice study, the same patients scored 58% response on the clinician’s rating, 56.4% on the PHQ-9 and 41.5% on the IDS-SR. A 708-patient UCLA study found that relying on a single scale carried a 14% to 36% risk of missing a response or remission that another scale detected. When you compare clinics’ numbers, ask which scale they come from.
| Scale | Who fills it in | Where remission starts | Used by |
|---|---|---|---|
| PHQ-9 | You. Nine questions, each scored 0 to 3, for a total of 0 to 27. | Scores of 5, 10, 15 and 20 mark mild, moderate, moderately severe and severe depression, so a score under 5 is below the mild range. (source) | Carpenter 2012, Leuchter 2023, Bastiaens 2024 |
| HAMD-17 (also written HRSD-17) | A clinician, in an interview. | 7 or lower was the remission line in the STAR*D study. (source) | THREE-D 2018, Leuchter 2023, O’Reardon 2007 |
| MADRS | A clinician, in an interview. | Under 11 was the remission line in the SAINT study. (source) | O’Reardon 2007, SAINT 2020 and 2022 |
| IDS-SR | You. A longer self-report questionnaire. | Set by each study. | Carpenter 2012, Dunner 2014 |
| CGI-S | A clinician’s overall rating of how ill you are. | Set by each study. | Carpenter 2012 (its main measure), Dunner 2014 |
TMS success rates in the major studies
These are the figures as each study reports them. Sham-controlled trials compare real TMS with a fake treatment that looks and sounds the same, which shows how much of the benefit comes from TMS itself. Practice studies show what happens to ordinary patients in ordinary clinics.
| Study | Who was treated | Response | Remission |
|---|---|---|---|
| Carpenter 2012, 42 US practices | 307 outpatients still depressed after an average of 2.5 adequate antidepressant trials; standard TMS as labeled | 58.0% (clinician-rated); 41.5% to 56.4% (patient-rated) | 37.1% (clinician-rated); 26.5% to 28.7% (patient-rated) |
| Leuchter 2023, UCLA clinic | 708 patients in routine rTMS care, rated weekly on four scales | 33% to 50%, depending on the scale | 20% to 24%, depending on the scale |
| Bastiaens 2024, multi-site group practice | 89 patients who had a full TMS course, compared with 583 in usual care, over six months | Not reported as a rate. PHQ-9 fell 5.7 points with TMS and 4.2 with usual care | Not reported as a rate |
| O’Reardon 2007, sham-controlled | 301 medication-free patients who had not benefited from earlier treatment; 4 to 6 weeks | Higher with TMS than sham on all three scales at weeks 4 and 6 | About twice the sham rate at week 6 |
| George 2010, sham-controlled | 190 medication-free outpatients at four university clinics; 3 to 6 weeks | Not the main measure | 14.1% with TMS vs. 5.1% with sham; almost 30% in the open-label phase that followed |
| Berlim 2014, 29 sham-controlled trials | 1,371 people, after about 13 sessions on average | 29.3% vs. 10.4% with sham | 18.6% vs. 5% with sham |
| Levkovitz 2015, deep TMS, sham-controlled | 212 outpatients who had failed one to four antidepressants; 20 sessions over 4 weeks | 38.4% vs. 21.4% with sham (week 5) | 32.6% vs. 14.6% with sham (week 5) |
| Blumberger 2018 (THREE-D) | 414 adults with treatment-resistant depression; theta burst vs. standard 10 Hz rTMS, 4 to 6 weeks | Depression scores (HAMD-17) fell from about 23.5 to 13.4 in both groups: the 3-minute theta burst session did as well as the 37.5-minute standard session | Not given in the abstract |
| Cole 2020, SAINT, open-label | 22 people with treatment-resistant depression; 10 sessions a day for 5 days, MRI-guided | Not the main measure | 19 of 21 who finished (90.5%); 19 of 22 counting everyone who started (86.4%) |
| Cole 2022, SAINT, sham-controlled | 29 people at an interim analysis | MADRS scores fell 52.5% on average vs. 11.1% with sham, 4 weeks after treatment | Not given in the abstract |
Why the numbers run from 14% to 90%
Every figure in the table is real. They differ because the studies asked different questions:
- Who was treated. The early sham-controlled trials took people off antidepressants first (O’Reardon 2007, George 2010). Practice studies treat people who stay on their medication.
- How many sessions. The 29 trials in the 2014 meta-analysis averaged about 13 sessions, far fewer than the 36-session course Labyrinth uses for depression.
- Which scale. The same group of patients can show 33% or 50% response depending on the questionnaire (Leuchter 2023).
- How hard the depression has been to treat. The number of earlier treatments that failed was the strongest predictor of outcome in a 301-patient analysis (Lisanby 2009).
- Which protocol. SAINT’s 90.5% remission came from 21 people given 50 sessions, 10 a day for five days, with MRI-guided targeting (Cole 2020). That is a different treatment from a standard daily course.
Kinds of TMS, and when the FDA cleared them for depression
“TMS” covers several protocols. The FDA cleared the first rTMS device in 2008 for people whose depression did not respond to at least one antidepressant in the current episode. Since then:
- Deep TMS can use two coils to reach larger structures deeper in the brain (NIMH). A deep TMS system was cleared on January 7, 2013 (FDA K122288).
- Theta burst gives a session in about 3 minutes instead of 37.5. A system with theta burst was cleared on August 14, 2018 (FDA K173620), the same year the THREE-D trial found it did as well as standard TMS.
- Accelerated, MRI-guided TMS (SAINT) gives 10 sessions a day for five days (Cole 2020). Its system was cleared on September 1, 2022 (FDA K220177).
Sessions at Labyrinth run 19 to 45 minutes, depending on the diagnosis. Which protocol the practice would use for you is decided at the evaluation; ask about it there.
How long until TMS starts to work
TMS is not a one-day treatment, and the studies on this page describe a change that builds over the first few weeks. Here is what the studies show, week by week:
| When | What the research shows |
|---|---|
| Week 1 (sessions 1 to 5) | The first session includes the measurements that set the machine and place the coil. In UCLA’s clinic, patients generally started reporting improvement within a week of starting treatment. |
| Week 2 (sessions 6 to 10) | The first useful checkpoint. More improvement by sessions 5 and 10 predicted response on all four scales in a 708-patient study, and in one Toronto clinic, people with under 20% improvement at session 10 went on not to respond 88.2% of the time. |
| Weeks 3 to 4 | When TMS is added to an antidepressant, response at about 2.7 weeks was 43.3% vs. 26.8% with sham. A 301-patient sham-controlled trial from 2007 measured its main result at week 4. |
| Weeks 5 to 6 | Response and remission were still higher than sham at week 6. NIMH describes a typical course as daily sessions, five days a week, for 4 to 6 weeks. |
| Weeks 6 to 8 | Labyrinth’s standard course for depression is 36 sessions, five days a week, which takes about six to eight weeks. |
Sources: Leuchter 2023 and the UCLA release, Feffer 2018, Berlim 2013, O’Reardon 2007, NIMH.
A slow start is not the same as no response. The early-improvement findings are probabilities for groups of patients, not a rule for you. Labyrinth’s care team tracks your symptoms, side effects and daily functioning through the course and explains how you will decide together whether to continue. Ask at your evaluation which scale the team uses and how often it checks it.
How long TMS results last
For people who respond, the benefit often holds for months. The evidence:
- One year in everyday practice. Of 120 people who responded or remitted after TMS, 75 (62.5%) still met response criteria through 12 months of follow-up. They continued antidepressant medication, and 93 of the 257 followed (36.2%) had TMS again at some point, for an average of about 16 treatment days (Dunner 2014).
- Across 19 studies. Among responders, 66.5% were still responding at 3 months, 52.9% at 6 months and 46.3% at 12 months. Studies that gave maintenance treatment, and studies with more women, had higher rates (Senova 2019).
- Deep TMS. In the 212-patient trial, the advantage over sham held through a 12-week maintenance phase (Levkovitz 2015).
Follow-up care at Labyrinth continues after the last session. Whether maintenance or repeat sessions would be part of your plan is a question for your evaluation.
What makes success more likely
No test predicts your result. These are the patterns the trials found:
- Fewer failed treatments so far. The strongest predictor of response in the 301-patient trial analysis (Lisanby 2009). In a 190-patient sham-controlled trial, most people who reached remission had low treatment resistance (George 2010).
- A shorter current episode. Linked with better outcomes in the controlled trial (Lisanby 2009).
- No anxiety disorder alongside the depression. Linked with better outcomes in the open-label extension (Lisanby 2009).
- Less severe depression at the start. Higher starting severity meant a lower chance of remission (Leuchter 2023).
- Early change. Improvement by sessions 5 and 10 predicted response (Leuchter 2023).
In these trials, TMS tended to work better for people with fewer failed treatments and a shorter current episode. That is a reason to ask about it sooner rather than later, not a promise.
How TMS compares with trying another medication
The STAR*D study of medication steps, which started with 3,671 outpatients, found remission in 36.8% of people with their first treatment, 30.6% with the second, 13.7% with the third and 13.0% with the fourth. The TMS figures above come from people who had usually already been through two or more medications. These are different studies with different scales, so they are not a head-to-head comparison. They do show that each further medication step helped fewer people.
TMS is not the only option. NIMH notes that rTMS does not match the effect of ECT, which needs anesthesia. SPRAVATO (esketamine) is a nasal spray taken in the office. Our TMS vs. SPRAVATO page compares the two, and treatment-resistant depression explains when each is considered.
When TMS is not the right choice
TMS does not work for everyone, and it is not safe for everyone. It is generally not used for people with a seizure disorder, metal or electronic devices in or near the head, implanted heart devices close to the coil, or head tattoos made with metallic ink. Our article Is TMS safe? covers side effects and screening.
Insurance, in brief
Medicare and many commercial plans cover TMS for depression when their criteria are met, and most require approval before the course starts. Our TMS cost in New Jersey article lists 2026 Medicare amounts and what each insurer requires.
Getting TMS in Cranford
Labyrinth Psychiatry Group and TMS Center gives TMS at 11 Commerce Drive, Lobby Floor, in Cranford, a short drive from Westfield, Garwood, Clark, Springfield, Union and Summit. The standard course is 36 sessions for depression and 30 for OCD, five days a week. There is no anesthesia, so you can drive yourself to and from each visit.
Questions worth asking at your evaluation:
- Which TMS protocol would you use for me, and why?
- Which depression scale do you track, and how often?
- What happens if I have not improved by the second or third week?
- Would maintenance or repeat sessions be part of my plan?
Call 908-336-1187 or use the evaluation request form. The TMS therapy page walks through the course from the first session to the last.
Questions about TMS success rates
What is the success rate of TMS for depression?
It depends on how success is measured. In 42 US practices treating 307 people whose depression had not improved with medication, clinicians rated 58% as responding (symptoms at least halved) and 37% as in remission (symptoms down to a minimal level) at the end of TMS. Patient questionnaires in the same study put response at 41.5% to 56.4%. Early sham-controlled trials, which took patients off antidepressants and used shorter courses, report lower rates.
How long does it take for TMS to work?
Change often starts early, but the full course still matters. In UCLA’s clinic, patients generally began reporting improvement within a week, and in that 708-patient study more improvement by session 5 and session 10 predicted who would respond. A standard course is five sessions a week for several weeks; at Labyrinth it is 36 sessions for depression, which takes about six to eight weeks.
How long do TMS results last?
In a one-year follow-up of 120 people who responded to TMS, 62.5% still met response criteria at the end of the year while staying on antidepressant medication, and about a third of everyone followed had some TMS again during that year. A review of 19 studies found that about half of responders were still responding 6 to 12 months later.
What happens if TMS does not work for me?
TMS does not work for everyone; in the 42-practice study about 4 in 10 people did not respond. Your psychiatrist reviews what to do next, which can include SPRAVATO (esketamine), a medication change, or another treatment. If there has been little change by about the second week, ask your care team how they read your progress.
Is theta burst or deep TMS more successful than standard TMS?
In the THREE-D trial of 414 adults, a 3-minute theta burst session worked as well as a 37.5-minute standard session. A deep TMS trial of 212 people found 38.4% response and 32.6% remission at week 5, compared with 21.4% and 14.6% on sham. The studies on this page did not compare deep TMS directly with standard TMS. Ask at your evaluation which protocol the practice would use for you.
What is the success rate of TMS for OCD?
In an 11-center trial of deep TMS for OCD, 38.1% of people responded after six weeks of daily treatment, compared with 11.1% on sham. One month later the figures were 45.2% and 17.8%. Response in that trial meant at least a 30% drop on the Y-BOCS OCD scale. Labyrinth’s standard TMS course for OCD is 30 sessions.
Why do some clinics quote success rates of 80% or 90%?
Those figures usually come from SAINT, an accelerated, MRI-guided protocol given as 10 sessions a day for five days. In its open-label study, 19 of 21 people who finished met the remission line. The study had 22 participants and no sham group, so its authors called for controlled trials. It is a different protocol from a standard daily course, and its numbers do not carry over to other kinds of TMS.