TMS for Depression: Response Rates, Protocols, and the SAINT Question

So the questions I get most often about TMS are, will it work or who will it work for? How does it work? And hey, what about this SAINT protocol? So in this video, I want to talk about transcranial magnetic stimulation for depression, major depressive disorder specifically. So will it work for you or for your patient? Well, what we can say is quite reliably, two-thirds of people, at least 60 to 70%, with

depression will respond to TMS. Now, what does "respond to TMS" mean? Well, that means if you look at assessments of depression symptom severity, whether that's a self-report, something like the PHQ-9 a patient fills out, or clinician-rated, something like the MADRS, which is a structured interview about depression symptoms, there is a reduction in symptom severity by at least half. So the score might go from a 28 to a 14 or less. And that 60% of people,

uh, that number is borne out by real data, big and small. So the NeuroStar Outcomes Registry, NeuroStar being one of the big TMS device manufacturers, now includes 17,000 patients across their national network, and the self-report scales with them show 69% of people showing that significant response. And about half of those reach full-on remission, meaning no major depression symptoms after a TMS course. The

clinician-rated numbers are actually higher than those, but I want to go with the self-report here because I think how the patient feels is probably more important to the patient than how the clinician says they feel. But those numbers too, I will say, are borne out just in the sampling of my own clinic, so that's certainly reassuring to me, you know, that the large-scale NeuroStar registry data does not seem to be biased, uh, by the manufacturer's collection methods, because essentially I see those same proportions, two-thirds

responding, a third remission. Those are the same proportions I see with my own eyes. So nothing like seeing something with your eyes to, uh, to believe, right? So at least with well-trained TMS sites, I'm pretty confident in the likelihood of response. Odds are you'll have a response. But I think what is still an open question is who it is that responds among people with depression. And I think the field is still awaiting reproducible data in terms of predicting

which patients are in that two-thirds that responds or even more so in the one-third that fully remits. Now, there may be a future where knowing what meds did or did not work in the past, or looking at inflammatory markers or particular features of the depression, you know, an agitated versus an anhedonic depression, maybe those things can tell us who's a good candidate somewhere down the road. Now, if you'll indulge me, to be a bit anecdotal, I will say

that in my clinical experience, it's that moderate severe depression where a primary issue is sort of a lack of activation energy. You know, trouble getting started at things, especially at more complex tasks, but maybe not the full sort of paralytic or catatonic features. But in that kind of insidious depression where if you're a therapist, this is someone that you're thinking, uh, we need behavioral activation. If that's your goal, then I think TMS

is a biological intervention that seems to dovetail well with that, at least in my experience. Now, of course, there are limitations, someone needs to have the sort of psychosocial stability that allows for coming into a TMS clinic Monday through Friday for about six weeks, and there are some contraindications, the main one being metallic hardware above the neck since we are generating an electromagnetic field right at the head, so no ferromagnetic implants

near that coil, that powerful magnet. Beyond that, real caution around anything that lowers seizure threshold, so if someone is on high-dose bupropion, especially stacked with something else, tramadol, clozapine, that combination is a, a conversation and a risk-benefit analysis before referral. Practically speaking, most patients will drive themselves to and from TMS sessions. The cognitive effects are essentially nil, so that's a real

advantage over ECT, say. The most common side effect of TMS is scalp discomfort at the stimulation site. Tends to be worse in the first few sessions, and then most people adapt sort of like, uh, a sore muscle from a new use. On the insurance side, most commercial payers and Medicare will cover TMS for treatment-resistant depression, maybe with prior authorization, so documenting that there have been, let's say, two medicine or two treatment failures in the current episode is

generally gonna get this covered by most insurance companies. But if you're a clinician, I would lay out those medication trials or other treatment trials when you're putting in the referral. Now again, there's a lot we don't know in terms of who is empirically a good candidate, and beyond that, there's things we don't know about which protocol of TMS we might recommend for a patient. So first off, regardless of TMS protocol, the overall mechanism

is about inducing neural stimulation using an electromagnetic field. So if we can remember our physics, an electric current produces a magnetic field perpendicular to the current, right? So if the coil has a pulsing electric current running like this parallel to the brain, a magnetic field is going to be directed perpendicular to that towards the head, and by Faraday's induction, a rapidly changing magnetic field that is entering the

brain will actually induce an electric field upon reaching the brain, it being a conductive substrate, right? An electrophysiologic substrate. That electric field induced by the electromagnetic, uh, induction will depolarize neurons. So in terms of protocols, the 10 hertz repetitive TMS protocol, which was the first FDA approved to the dorsolateral PFC, has a stimulatory effect on the

neurons in that region, right? So dorsolateral PFC, if we think about there being evidence that depression seems to represent a sort of retreat of brain activity from these lateral task positive networks towards the midline, the default mode, and depression ruminations seem to be sort of fixed along that overactive default mode network. We can think about antidepressant treatments as having the goal of moving the activity back laterally,

waking the brain out of this sort of ruminative retreat along the midline and into the more task-positive, externally-focused lateral networks. So we're stimulating a lateral prefrontal cortex network to bring that activity back out as it were. That's the general logic of why stimulating the lateral prefrontal cortex would help depression. So the other now standard protocol in addition to the 10 hertz repetitive TMS is theta burst, also called

intermittent theta burst stimulation, ITBS. So you can think about the classic rTMS at 10 hertz versus theta burst, ITBS. And functionally, the main difference will be that each time you come in for treatment, theta burst is a much shorter session. So theta burst you could say packs more of a punch into a shorter window in that each treatment is only a few minutes as opposed to at least 20 minutes for the 10 hertz standard protocol. And that's because with theta burst you're actually

getting these nested pulses. Instead of a 10 hertz repetitive pulse, you're getting 50 hertz stimulation within five hertz pulses. So that can be a bit more intense in terms of the sensation, the sort of, woodpecker on the scalp. But it's also much shorter and when this has been compared to 10 hertz what we can say is that it's non-inferior. That is when you compare 10 hertz TMS to theta burst, theta burst is neither better or worse. It's an

equivocal outcome. So again, the question is about treatment session logistics. The fact that theta burst is substantially shorter for some people makes it more attractive in and of itself. But either way the overall course is six weeks or 30 to 35 treatments. So it is a time commitment in the long term regardless. So that then brings us to the SAINT protocol. SAINT being the Stanford Accelerated Intelligent Neuromodulation Treatment. This is developed by Dr. Nolan

Williams, may he rest in peace, a great inspiration to myself, but more so a great inspiration and influence on the entire field or fields of neuroscience and psychiatry. So Nolan Williams wanted to harness the natural mechanisms of the mind, and in particular of learning and memory, something called spaced learning. Uh, he would use the example of studying with flashcards, something that I think certainly,

med students have always related to, though nowadays maybe it's not the paper flashcards, right? It's Anki, at least at last I checked, Anki digital decks on the computer. But the point is the same, learning with flashcards, learning with repetition. Nolan would ask, how do you study? When you study with flashcards, do you just take the first one and just look at that for a while? Look at it until you remember it? No, of course not. You look at it and then you go through the whole stack. But nor once you're

finished with the stack do you just put that aside and let it rest for days or weeks, because you would forget everything, right? So what you want is to go through the stack such that you're priming each of these little memory traces that you're trying to learn, and then when there's still a memory trace, you know, after a refractory period but before the trace disappears, you reinforce it. So you reinforce it while it's in that sort of soft encoding phase.

There's a refractory period after each flashcard, but in about 50 minutes or once you're through your stack, you can reinforce it, and that's what you do. That's what works. So if you think about TMS as a kind of learning process, but where some of the learning is with exogenous stimulation rather than the pure internal stimulation that students might rely on, SAINT would be a version of TMS that is really intensive in this sort of spaced learning version. So each day,

rather than doing a single session, you actually do 10 sessions with 50 minutes in between. So it's a theta burst session, 50-minute refractory period, another theta burst session, essentially all day, eight to 10 hours of treatment, but that whole course is just five days. And so who this really seems to benefit is the more severe end of the spectrum. Originally it was done on inpatient units. Uh, it's people that need a robust and rapid response because their

symptoms are acute, and it is then this more intensive procedure. So there are really two things that distinguish SAINT from the standard TMS protocols we talked about before. One is that 10 sessions a day for five days with SAINT as opposed to one session a day for 30 days with standard TMS. But the other thing about it is that intelligent targeting part, right? So it's, it's precision targeting. With standard TMS, we localize the prefrontal cortex,

which is our target, by first finding the motor cortex and then the motor cortex, that being the site where each stimulation produces a maximal hand twitch. Once we know where that is, we can use skull landmarks and calculations to move the coil to the dorsolateral PFC, uh, based on its relative location with that motor cortex. Nolan Williams and Stanford decided that they wanted to find an even more specific node in the dorsolateral PFC.

Nolan wanted to find the point that was the most anti-correlated with the midline anterior cingulate. So if you remember that, that evidence about depression being sort of overly active and fixed along midline in terms of brain activity, especially in that anterior cingulate. And if the goal is a shift towards lateral PFC activity, the hypothesis with SAINT is that there's one sort of best node per individual that maximally increases

lateral activity while simultaneously decreasing that midline cingulate activity. So fMRI and a connectivity algorithm can do that. Uh, that's what SAINT uses. But of course, getting every patient an fMRI, which is not even just a single sort of MRI snapshot, it's time in the scanner where the activity is getting recorded. For now, this is just prohibitive for most sites. There just aren't that many fMRI scanners, and

both the time and the labor to get that data is very expensive, very involved. So SAINT is restricted, as it stands, to a handful of academic medical centers and specialty sites operated through Magnus Medical, who owns the patent. But I think one thing that remains to be determined is how necessary is that degree of target localization actually? Could there be something like a SAINT protocol in that it uses the

spaced learning, something like 10 treatments a day for five days, but uses a more traditional motor cortex-based targeting, right? So what would be the number needed to treat that would actually make getting the fMRI, uh, worth it in terms of finding that target? I have not seen this done, but of course, it's always possible that I'm not 100% up to date or I'll admit, maybe I'm not even 100% correct about the claims I'm making. This is my best

understanding of the state of the field, and I certainly welcome any feedback or corrections about, uh, these issues I've laid out. For now, I hope this was a helpful overview of TMS for depression. Thank you.