If you’re asking this question, you’ve probably already tried the traditional medication route of one antidepressant, then another, or several and maybe even mood stabilizers and antipsychotics. Any you’re prbably skeptical of anything that sounds like too good to be true. So let me give you the answer the way I’d give it to a patient during a consultation: yes, TMS works, the evidence is strong, and it’s also not magic. Here’s the honest picture, in plain language.
Start with the FDA and why that means something here
Transcranial magnetic stimulation was first cleared by the FDA for major depressive disorder in 2008, after randomized controlled trials showed it outperformed sham treatment. Since then the clearances have kept coming, each backed by new data. Heres info on the machine we use at WIP, MagVenture: cleared for MDD in 2015, OCD in 2020, depression with comorbid anxiety symptoms, “anxious depression” in 2025, and accelerated TMS protocols in May 2026. Adolescents aged 15–21 were added in 2025 as well. Standalone anxiety disorders remain an off-label use, though treating the anxiety that travels with depression is now on-label, which matters, since more than half of depressed patients have it.
The FDA isn’t perfect. But it does not clear treatments without solid evidence, and here’s a detail worth pausing on: TMS got through without a pharmaceutical marketing machine behind it. There’s no drug reps, no television commercial, no billion-dollar ad budget pushing TMS into your feed. The fact that it earned FDA clearance, insurance coverage, and nearly two decades of growing clinical adoption by physicians is largely on the strength of its data not its marketing. That alone should raise your curiosity and confidence.
It’s not a medication and that matters
TMS is a purely biological approach to depression that doesn’t involve a drug at all. A magnetic stimulator, positioned against your scalp (placed in front and on the left), delivers focused pulses that stimulate brain cells in regions that are underactive in depression. Most importantly the ‘left dorsolateral prefrontal cortex’, an area for mood regulation that consistently shows reduced activity in depressed patients.
Nothing enters your blood. That’s why the side-effect profile looks nothing like an antidepressant’s: no weight gain, no sexual side effects, no sedation, no withdrawal when you stop. The most common side effect is a headache, typically in the first week or two although even that is rare now, and scalp discomfort at the stimulation site. The serious one to know about is seizure which is the reason TMS is a medical procedure and not a wellness service. The risk is rare, estimated at fewer than 1 in 10,000 treatment sessions, but it’s real, which is why screening for seizure risk factors is part of any responsible TMS evaluation.
For the biochemistry-inclined: each magnetic pulse induces a small electrical current in the cortex beneath the coil, depolarizing neurons and triggering action potentials. Repeated over a session, this drives synaptic plasticity, long-term potentiation-like changes, not just in the stimulated region but across the connected mood-regulation network, including deeper structures the coil can’t reach directly. It quite literally, retraining a circuit.
And because TMS is technology, not a molecule, it isn’t static. A drug is the same drug forever. TMS keeps improving in efficacy, speed, and tolerability. Sessions that took nearly 40 minutes in 2008 now take as little as 3 minutes with modern theta-burst protocols. That trajectory is still climbing.
So what do the numbers actually say?
Here’s where I’ll be more careful than most websites you’ll read.
One scoping note first: everything in this post is about depression. TMS for OCD is FDA-cleared too, but it’s a different protocol with its own evidence base and its own numbers so it deserves its own article.
In the large real-world studies of standard TMS, patients treated in actual clinics, not trials-roughly 50 to 70 percent of people respond (meaning their depression improves substantially), and roughly a third reach full remission. The landmark multisite study by Carpenter and colleagues found a 58% response rate and 37% remission rate, and subsequent studies have landed in the same range.
The newer accelerated protocols post even more striking numbers. Stanford’s SAINT protocol an fMRI-targeted, five-day accelerated form of TMS achieved remission in roughly 79% of patients in its randomized controlled trial, and 86–90% in open-label studies, in people who had failed an average of five medications. Those are the studies behind the “70–90%” figures you may have seen. Whats still unknown is how much of the fMRI targeted part led to improvement vs the speed and frequency at which it was given. Still leaves the expensive need for fMRI or simply doing a standard location called motor threshold or ‘mapping’ up in the air. fMRI is available around the country but makes it very inaccessible for most people due to its cost.
Now the honest caveat: clinical trial populations are much cleaner than real life. Studies typically enroll people with major depressive disorder and little else. Real patients often carry depression plus anxiety, trauma, ADHD, or substance use and comorbidity tends to pull real-world numbers below the headline trial figures. That’s not a knock on TMS; it’s true of every depression treatment. But a practice that quotes you only the best-case number isn’t preparing you properly.
For context: in people who have already failed two or more medications, the chance that the next medication produces remission drops below 15%. Against that baseline, TMS’s numbers are not incremental, they’re a whole different tier.
It works by the numbers but will it work for you?
If you’ve been through a few medication trials already, the numbers above suggest yes, plausibly. But the more honest framing isn’t a promise, it’s this: at that point in your treatment history, TMS is a reasonable next strategy with higher efficacy than trying yet another antidepressant. That’s the actual comparison that matters. You’re not choosing between TMS and a guarantee; you’re choosing between TMS and medication trial number four, whose odds you’ve already seen.
What about using TMS first- before medications at all? A lot of us in this field are hoping that question gets answered, but right now there simply isn’t data on TMS as a first-line treatment. One plausible reason: the people who respond well to SSRIs may have a different underlying biology than the people TMS helps so a first-line trial would be treating a population the studies haven’t characterized. Best case today: try an SSRI, then an SNRI and/or an atypical antidepressant like Wellbutrin (bupropion), if anxiety isn’t a major feature of your depression and then have the TMS conversation with your psychiatrist. If your psychiatrist isn’t familiar with it, my suggestion would be to have an evaluation with a psychiatrist who does TMS and who collaborates with you psychiatrist.
That said, the ground is shifting. Some insurers no longer require two failed medication trials before covering TMS- a quiet signal that insurers are moving toward or at least asking where TMS belongs on the treatment algorithm. Its lack of systemic side effects and its long-term durability are exactly why that question is really active across the interventional field. But we’ll be straight with you: we still wouldn’t recommend TMS as first-line therapy for depression. When that changes, it’ll be because the data changed and your psychiatrist should say so.
Who should think twice- honest caveats
A few situations deserve straight answers, because they change the calculus without necessarily closing the door:
Bipolar depression. The strong evidence for TMS is in unipolar major depression. For bipolar depression, the evidence base is much thinner with some encouraging signals, but not the robust trial data that backs TMS for MDD. If your depression is bipolar, that doesn’t automatically rule TMS out, but it means the conversation with your psychiatrist needs to be more careful, and getting the diagnosis right beforehand matters enormously.
Alcohol use. Regular or heavy alcohol use raises seizure risk- both through alcohol itself and especially through withdrawal which stacks on top of TMS’s own small seizure risk. It’s not an absolute contraindication, but it’s something your evaluating psychiatrist needs to know about honestly, and it may need to be addressed before or alongside treatment. Also, alcohol is a depressant so identifying changes while doing TMS and using alcohol is extremely difficult.
Medications that lower the seizure threshold. Certain medications including some antidepressants and stimulants among them can increase seizure susceptibility. That includes, notably, bupropion- the same medication we just mentioned as a reasonable trial step. That’s not a contradiction; it’s a sequencing problem, and sequencing problems are exactly what psychiatrists are for. A psychiatrist looking at your whole picture knows when bupropion is the right trial, when it needs adjusting before TMS, and how to weigh the combined risk- a technician running a machine, or a prescriber who isn’t talking to your TMS clinic, doesn’t. Usually not a hard stop, but it belongs in a risk assessment.
Metal (but the details matter). The formal rule, per FDA guidance, is that TMS is contraindicated with conductive, ferromagnetic, or magnetic-sensitive metal implanted in the head or within about 30 centimeters of the coil. The concern is twofold: ferromagnetic material (metal a magnet pulls on) can be heated or displaced by the pulse, and even non-magnetic conductive metal very close to the coil can heat through induced currents. That’s why certain older aneurysm clips, metallic fragments in or near the head, and magnet-containing devices like cochlear implants are the classic hard stops. But the geometry and the material both matter. Titanium is non-ferromagnetic and a poor conductor- titanium plates and hardware are generally compatible with TMS, evaluated case by case when they sit near the coil. And cardiac stents are generally fine: your chest sits well outside that 30-centimeter radius, and modern stents are non-ferromagnetic alloys anyway. (A stent in the neck, carotid, is a different conversation, because of that radius.) Implanted electronics like pacemakers, stimulators are their own category: the safety consensus considers TMS workable with pacemakers, assessed individually and mostly around where the leads run, not the metal itself. The short version: “I have metal in my body” rules out far fewer people than the internet suggests, but it’s the kind of detail your screening evaluation exists to sort out.
None of these are reasons to write TMS off. They’re reasons the screening evaluation is a real medical assessment and another argument for having it done by a psychiatrist rather than an intake form.
The real limitation isn’t efficacy, it’s the calendar
Standard TMS runs five days a week for four to eight weeks. Each visit is short — with modern protocols, about three minutes of stimulation but you have to show up, daily, for weeks. Adherence is the single most important factor in whether TMS works, which is why we tell people something you’ll rarely hear a practice say and it’s the same thing we tell patients in our own consults: if everything else is equal, choose your TMS clinic primarily by proximity. The best protocol in the world doesn’t work if the commute makes you quit in week three.
But use judgment with that advice, because TMS clinics are not interchangeable. As the field has grown, many clinics have opened that are not run by psychiatrists. Close-to-home only wins if the clinic clears a quality bar. What to check for:
Who’s actually overseeing your care? A machine can be operated by a technician- that’s normal- the cost would be too high if every treatment was run by a doctor but your depression shouldn’t be managed by that technician. People genuinely improve at weeks 4, 6, and 8 and those changes are often subtle in ways only a trained eye catches. Someone qualified, ideally a psychiatrist should be formally checking your progress at those milestones and interpreting what’s changing, not just running the machine for 8–12 weeks and measuring at the end. And the bar for continuing matters: improvement should be significant, not marginal, to justify ongoing treatment. If the clinic doesn’t have a psychiatrist doing this, make sure your own psychiatrist is on board and tracking your response in parallel. Also, discussions on sleep need to be addressed- all psychiatric treatments work better when sleep is managed.
What machine do they use, and for what? Different TMS systems differ meaningfully, and different conditions- depression, OCD, anxiety (an off-label use on its own) involve different coil placements and can require different equipment. A clinic should be able to tell you exactly what device they use and why it fits your diagnosis.
What happens after it works? This is the most neglected question. If TMS succeeds, you should still be checking in with a psychiatrist at that clinic because in our experience, many patients are never properly educated on maintenance. Follow-up then gets dictated by insurance approvals, or by waiting until things fall apart again. Best practice is the opposite: re-treat when symptoms begin to worsen, slightly not when they’re terrible. That means knowing your own subtle early signs: a change in sleep patterns, a little more difficulty getting out of bed, feeling more tired than usual during the day. A good clinic teaches you to recognize these and has a plan ready for when they appear.
Accelerated TMS (aTMS) compresses the course into about one week- multiple ~10-minute sessions per day. Most people take the week off, though at our office patients typically get a private room with a desk and couch, so many keep working virtually between sessions. Expect to be genuinely fatigued by the end of each day and by the end of the week- that’s the concentrated stimulation, and it’s temporary. Daily standard TMS, by contrast, causes little to no fatigue; three minutes of stimulation a day is just much more tolerable.
“Do I need the fMRI-guided version?” — a question we get a lot
If you’ve read about accelerated TMS, you’ve probably seen that the Stanford protocol used fMRI to individually map each patient’s stimulation target. So people reasonably ask: do I need brain imaging for TMS to work?
Honest answer: fMRI-guided targeting is probably better and it’s far more expensive, with no insurance coverage, which makes it financially inaccessible to most people. Here’s the context that headline misses:
Motor-threshold-based targeting has been the clinical standard for decades, and it’s effective. All of the outcome data for standard TMS cited above the 50–70% real-world response rates was generated with motor-threshold targeting, not imaging. The method works like this: we find the precise spot on your motor cortex that makes your hand twitch, calibrate the dose to your brain’s threshold, and locate the treatment target from that individualized anatomical anchor. It’s worth understanding that this is a clinical skill applied to your individual brain, not a one-size-fits-all template. Done well, it’s personalization by examination rather than by imaging.
And here’s the part nobody selling fMRI-guided TMS will volunteer: no study has directly compared the two approaches. The Stanford results are impressive, but that protocol bundled fMRI targeting with an intensive accelerated schedule and other innovations. There’s (as of writing this) no trial isolating how much the imaging itself adds over skilled motor-threshold targeting. “Probably somewhat better” is a fair read of the indirect evidence; “necessary” is not supported by anything.
So our recommendation, for the cost-benefit to patients: motor-threshold-based treatment. If fMRI-guided targeting becomes insurance-covered and its added value gets demonstrated head-to-head, that math changes and, as with everything else in this article, we’ll say so.
The bottom line
Does TMS really work? The evidence says yes and has been FDA-cleared since 2008, response rates of 50–70% in real-world practice, remission rates approaching 80% in the newest accelerated trials, and a side-effect profile most antidepressants can’t compete with. What it asks of you isn’t tolerance of side effects; it’s showing up.
And if TMS turns out not to be your answer? This is where being treated at a comprehensive practice matters. TMS is one tool of many- ketamine, Spravato, SGB, medication optimization, and more, under one psychiatrist-led roof, so a non-response isn’t a dead end, it’s information that directs the next step.
If you’re reading this and you’re in the DMV, we’re in Washington, DC and Rockville, MD now.
