From Our Neurons to Yours

Transcranial magnetic stimulation and the future of psychiatry | Corey Keller

• Wu Tsai Neurosciences Institute at Stanford University, Nicholas Weiler • Season 9 • Episode 14

Share episode

Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.

0:00 | 46:21

Transcranial magnetic stimulation (TMS) is one of the most promising new technologies in psychiatry. Essentially, it allows us to reach inside the brain, stimulate it, and alter the brain circuits that go awry in disorders like depression. 

As the late Nolan Williams once explained, we've known for a long time that mental health disorders are caused by changes in electrical circuits in our brains, until TMS came along, we didn't had a good way of nudging those circuits back toward healthier function.

But for all its promise, there is still much we don't understand about what exactly TMS is doing. We know it sends electromagnetic waves into the brain and changes the flow of electricity within. What we don't know is how it changes brain circuits and networks or why that helps treat depression, obsessive-compulsive disorder, or any of the other conditions where it's being used.

Today's guest, Corey Keller, is dedicated to addressing that gap. He's been developing methods for recording inside the brain to help us understand how treatments like TMS change our brain circuits and address psychiatric disease.

Learn more

Send us a text!

Thanks for listening! If you're enjoying our show, please take a moment to give us a review on your podcast app of choice and share this episode with your friends. That's how we grow as a show and bring the stories of the frontiers of neuroscience to a wider audience.

We want to hear from your neurons! Email us at at neuronspodcast@stanford.edu

Learn more about the Wu Tsai Neurosciences Institute at Stanford and follow us on Twitter, Facebook, and LinkedIn. 

Nicholas Weiler (00:10):

This is From Our Neurons to Yours, from the Wu Tsai Neurosciences Institute at Stanford University, bringing you to the frontiers of brain science. I'm your host, Nicholas Weiler. You've probably heard of transcranial magnetic stimulation. And if you haven't, it is one [00:00:30] of the most promising new technologies in psychiatry, essentially allowing us to reach inside the brain and stimulate and alter the brain circuits that are going awry in psychiatric disorders. We talked about TMS before on the show with Nolan Williams a couple of years back, who's one of the pioneers of this field.

(00:49):

The way he framed it for us was that while we've known for a long time that mental health disorders, things like depression, are caused by changes in the flow of electricity in particular circuits [00:01:00] within our brains, until now, we haven't had a really good way of getting in there and changing those circuits of bumping them back into a healthier function. Sadly, Nolan passed away last year after his own long battle with depression, which was a huge loss for our community and for the field as a whole, and a story that just goes to show that for all the progress we've made, we still have a long way to go in understanding these conditions enough to really help people get their lives back when struggling with depression.

(01:30):

[00:01:30] For all the promise of TMS, one of the biggest gaps in our understanding that is holding back this exciting field is that we still don't exactly understand what TMS is doing inside the brain. We know that it's a magnetic stimulation that goes through the skull, that creates electrical currents inside the brain and changes the flow of electricity within our brain networks. But what we don't know is how is it actually changing those brain networks [00:02:00] in a way that helps people with the symptoms of depression or with OCD or with PTSD or any of the other conditions where it's being used?

(02:08):

Today's guest is dedicated to understanding exactly this question. Corey Keller is an associate professor of Psychiatry and Behavioral Sciences at Stanford Medicine and a member of the Wu Tsai Neurosciences Institute who has been developing technologies for recording inside the brain to help us understand how are new treatments like TMS actually changing our brain circuits [00:02:30] and helping us get better from conditions like depression. I started my conversation with Corey by asking him to give us a little primer on where we are with TMS, what he sees as the promise of this new technology, and what are some of the things that are holding us back?

Corey Keller (02:50):

The exciting thing about TMS is it was used for many, many years in motor physiology and in neurology. And Mark George and Holly Lisanby brought it into psychiatry [00:03:00] and tested it in depression in a brain region, the left prefrontal cortex that we had known for a long time that was hypoactive or decreased activity in patients with depression. So could we sort of bump it up, like you said, can we activate it? Can we strengthen that part of the brain with downstream circuitry with something like TMS? And clinical trials time and time again against sham or placebo has been very effective. So it's in [00:03:30] the clinics, hundreds of thousands of patients with treatment-resistant depression are treated worldwide and it can be very effective.

(03:38):

And yet from a neural mechanism point of view, the exact circuits that it modifies and how long lasting the effect is, we don't know that much about. And part of that lies from the lack of non-invasive neuroscientific tools to explore this. We have functional [00:04:00] MRI and EEG and both of which have spatial or temporal limitations. And so we can take pictures of what happens in someone's brain before and after an entire treatment course of TMS, but we don't know at a granular level, both inside the brain at a millisecond level, what is happening? What is the neuroplastic effect that TMS is inducing? How does that accumulate over time throughout the treatment course and how does that eventually [00:04:30] translate to a clinical effect?

(04:33):

Without that mechanism, I think there's a lot that we're missing out in terms of the clinical efficacy. TMS, even for the newest treatments a month after the treatment, only has about a 50% clinical response. That is 50% of our patients show about a 50% reduction in their depressive symptoms. So there's still a long ways to go in terms of getting an effective and durable [00:05:00] effect with TMS.

Nicholas Weiler (05:02):

I want to compare that 50% that you cited. I mean, one, 50% is not nothing. And my understanding is that's kind of typical of a lot of different therapies, right? SSRIs, if I remember correctly, also have maybe roughly a 50% efficacy rate. And if I remember correctly, the sort of accelerated protocol that was developed by your late colleague, Nolan Williams, the SAINT protocol, we had him on the show a couple of years ago [00:05:30] talking about this, had something like an 80% remission rate for people with treatment resistant-depression, and that was just five days of stimulation. So I'm playing devil's advocate here a little bit because I'd love to hear you lay out... I'm never opposed to more research, but if we are already getting pretty positive results from TMS, what is the clinical need to understand the mechanism better?

Corey Keller (05:54):

Yeah, it's a great question. So why do we need to do better than these numbers? And you're right, 50% [00:06:00] is not nothing. And when it works, it works really well. And keep in mind, these are treatment-resistant patients. These are patients that have failed multiple rounds of antidepressants. So to have an effective treatment with minimal side effects at that stage is really valuable. And that's why it's FDA cleared. That's why I'm really excited about it. That's why I use it in the clinic a couple of times a week. SAINTS took it one step further, Nolan Williams' work where you could deliver the entire treatment course, which used to be [00:06:30] six weeks, once a day treatment. Now we can deliver treatment in one week.

(06:34):

That is 10 treatments a day for five days, which is a huge advance in terms of access to care, right? People can fly in for a week. People can drive from long distances and stay in a hotel room for a week and get their care. That 80% number is a little inflated, and it depends on the statistics. That number came from at any point in time in those five days, did someone remit from [00:07:00] the treatment? That wasn't at day five at the end of treatment course, what was the percentage of patients that were remitted?

Nicholas Weiler (07:07):

Meaning they said, "I feel better. I don't need this anymore." What do you mean? What does that mean?

Corey Keller (07:11):

Meaning when you assess every single day at one point in their treatment course, they did not meet criteria for depression. So it was very effective, but not necessarily at the end of treatment. Usually clinical trials assess only at the end point what was that clinical response rate. [00:07:30] So it's a little bit higher than it would have been at day five, but more importantly, one month after that treatment, the clinical response rate, that is the percentage of people who get 50% better or more was still at 50%. So it's actually the same number as your more traditional treatments. So it was very effective in those five days, but did not last. And that was one month afterwards. So a lot of people [00:08:00] relapse. So I think we can still do a lot better in terms of creating a durable treatment that changes one's circuits for longer than one month.

Nicholas Weiler (08:11):

Yeah. And one of the things I think we'll get into as we dive deeper into this is also, do we need to think about how do we personalize the treatment, right? Do we need to do something different for some people than for others? It's still sort of a one size fits all approach, right?

Corey Keller (08:30):

[00:08:30] Yeah. And it depends on how you look at it. So there are certain components of personalization right now. We personalize the intensity of the stimulation based on how much energy it takes for you to deliver TMS and get your finger to twitch. And what Nolan did with the SAINT protocol is to use image guidance to figure out where within that prefrontal cortex we should be stimulating. So I think we have a couple of the pieces. What we're missing is once we [00:09:00] have the target and the intensity, everything else is exactly the same. So the pattern of stimulation, the duration of stimulation, the time between treatments, the number of total treatments, all of those are exactly the same for every individual. And that's the piece that I think can be personalized and optimized significantly.

Nicholas Weiler (09:30):

[00:09:30] Let's take a quick step back. I want to make sure that everyone listening has sort of what we need to know about this technology to dive into your research and some of the progress that we're making in understanding what exactly is it doing in the brain? [00:10:00] How could we make it more effective? How could we personalize it? Could you give us a pretty concise primer on the technology? Just sort of what should listeners know about what exactly is TMS and how is it currently being used?

Corey Keller (10:14):

Yeah, that's a great question. And maybe we start with what do we know about depression and why are we using this tool in the first place?

Nicholas Weiler (10:20):

Sure.

Corey Keller (10:21):

So there's a long line of evidence, both with imaging and PET work and deep brain stimulation to suggest that the subgenual cingulate [00:10:30] cortex buried pretty deep in the prefrontal cortex is hyperactive in depression. And if we can turn that region down, we can help one's depression. Now in order to access that brain region, it's pretty far into the brain. And TMS really only hits superficial regions. And so what Mark George and Holly Lisanby and others thought is, can we target a region that would turn down the subgenual [00:11:00] cingulate cortex? And that's the dorsolateral prefrontal cortex. So it's right around here. If we apply a TMS coil to that brain region and activate that brain region, it will turn down the subgenual cortex based on the connection with it.

Nicholas Weiler (11:13):

You're pointing to like your left temple here, so somewhere right here.

Corey Keller (11:18):

It's right around there.

Nicholas Weiler (11:18):

So we're saying, so there's this deeper region, the subgenual cingulate, which is sort of buried underneath the frontal lobe of the brain. And [00:11:30] there are executive control regions up in the top of the prefrontal lobe that ideally should be able... They're supposed to be controlling activity, giving us the ability to regulate our moods, regulate our emotions, things like that. And so the idea is if there's too much activity in this emotion regulation area, maybe we can strengthen our own brain's cognitive control over that. Am I interpreting this correctly?

Corey Keller (11:59):

Yep. [00:12:00] Exactly. And so now the question is how do you actually do that from outside the brain? And that's where TMS comes in. So this is an electromagnet that when placed over a certain part of the brain and when discharged will induce a time-varying magnetic field that will go through your skull, will turn into an electrical current and will activate your neurons, will cause an action potential and activate the dorsolateral prefrontal cortex.

Nicholas Weiler (12:24):

This thing looks pretty cool. We should have brought a prop in [00:12:30] now that we're on video, but it's sort of like two big thick plates that almost look like a figure eight that are just like this magnetic coil and it can create a brief magnetic pulse. And you can just sort of bring that up in front of your head and create a magnetic pulse. And so the idea is because there's a magnetic pulse, it does something to neurons because they're electromagnetically active?

Corey Keller (12:56):

Through Faraday's law, and it's been a while, but the time-varying magnetic [00:13:00] field will pass through through the right hand row, will pass through the skull and induce a time-varying electrical field. And that electrical field will change the membrane potential of the neurons and cause that neuron to fire, those neurons to fire.

Nicholas Weiler (13:16):

And this is different from things people might've heard about with direct current stimulation, right? I mean, there are certain kinds of electrical stimulation that have been used in the past and some are still used in psychiatry, but this is something... You don't really feel it, right? [00:13:30] Or it's pretty mild.

Corey Keller (13:31):

You can feel it. It does activate the scalp neurons that are passing over the skull. And so it hurts a little bit, but it is different than the direct electrical current. When you apply electrical current to the skull, it's very hard to apply a strong enough current to actually get through the skull. So the way that you can get through the skull is through the magnetic field that will pass through. And so that's just a single pulse of TMS. And so what we deliver for TMS treatment [00:14:00] is a pattern of stimulation that has been known for many, many years. What we use now is theta burst stimulation, which is a tool in slice physiology and preclinical work that is thought to induce neuroplasticity.

(14:15):

So not just a single pulse, but it's actually 600 pulses in three minutes in a very specific timing sequence that is thought to not only activate that part of the brain, but induce neuroplasticity in that part of the brain and the connected regions, [00:14:30] that being the subgenual cingulate cortex. So the idea is after a dose of TMS, you are down-regulating or inducing long-term depression in the subgenual cingulate cortex. So it's not just for that moment it's turned off, but it's actually a weaker activation when the DLPFC gets activated.

Nicholas Weiler (14:51):

And so we should clarify the term long-term depression is not related to clinical depression. It's a [00:15:00] brain plasticity term, meaning a weakening of connections essentially, right? You're trying to weaken the connections that are keeping this hyperactivity, this excessive activity in the subgenual cingulate. And maybe you're also trying to strengthen connections from the more executive regions. I mean, this is giving me an image of almost like physical therapy for the brain, like you're trying to work out a particular circuit and weaken something that's, I don't know, you could think of it...

(15:28):

I'm going way far afield [00:15:30] on the metaphor here, but maybe you can think of it almost as like your muscles tied up in a knot and there's this other thing that's not balancing it correctly. You're trying to loosen one thing that is overly activated and strengthen something else to help balance it out. That's a pretty weird metaphor for the brain, but-

Corey Keller (15:48):

It's actually the exact metaphor that I use with my patients.

Nicholas Weiler (15:50):

Oh good.

Corey Keller (15:54):

So not so far off. And yes, I think you were saying that the DLPFC or the dorsolateral prefrontal cortex, it's not [00:16:00] only connected to the subgenual cingulate cortex. It's actually connected to many, many different brain networks. And so we're not only changing that connection with the subgenual cingulate, we're also changing the connections with the frontal parietal network and the default mode network. So it's complicated in terms of what we're actually doing and which of those network modulations actually induce the symptom change that we see when people get better with depression.

Nicholas Weiler (16:27):

And I love what you mentioned about the theta [00:16:30] burst stimulation that people are using now to try to get the sort of the timing of the stimulation right. Because it comes from these very longstanding basic neuroscience research, as you said, like slices of mouse brains in a dish. Scientists figured out, well, the brain seems to like to communicate at this particular frequency band. That can induce plasticity, that can make circuits change. If we do that with TMS, we get stronger results because [00:17:00] hopefully we're talking the brain's language and inducing plasticity. I think that's so cool.

Corey Keller (17:06):

Yeah. No, it's amazing. I think there's always complexity with any of these things, right? Taking a mouse hippocampus neural activity and trying to translate that idea to something in the human prefrontal cortex is a lot more complicated than it may seem. And so it should work, right? And in a lot of ways when we apply TMS, it does work for depression, but [00:17:30] how it works and why it works, we have not figured that out yet.

Nicholas Weiler (17:35):

Well, let's get into that then. Let's start talking about some of the work that you've been doing to try to figure out, well, what is happening when we do this stimulation? How are neurons actually responding? And one of the things that you've said you're really focused on, that your team is really focused on, is trying to come up with some clinically useful biomarker for is this TMS stimulation effective? [00:18:00] Are we doing the thing in the brain that we are theoretically trying to do? And it's interesting, you use the term biomarker for this. And I think most of us think of biomarkers as a molecule in the blood that reports on, is there potentially a tumor somewhere?

(18:18):

And I give this medication and you see less of that molecule in the blood, so hopefully we're reducing the tumor. But you're talking about something very different. You're talking about a biomarker in brain activity. [00:18:30] So before we dive into all the work that you've done to try to define what this should look like, how do you define a brain activity biomarker? What does that look like to you?

Corey Keller (18:42):

Yeah. It's a very general term that is thrown around quite often. My definition is, as long as you define what you mean by biomarker, you can use it in a lot of different ways. You can have a biomarker for, let's say, neuroplasticity, right? And that might be a brain [00:19:00] signal that you see somewhere in the brain that tracks neuroplasticity. It could also be a biomarker for a clinical treatment or a clinical severity. So I could have a brain signal that I'm watching for that if moving in the right direction, that is going to track symptom changes, that is going to predict someone's symptom change at the end of treatment.

Nicholas Weiler (19:23):

Well, let's get into the science. So the two things I really want to talk about, and these are two big areas of your research, [00:19:30] are first, you're doing research with people who have epilepsy, who already have electrodes implanted in their brains as part of sort of preparation for neurosurgery, planning for neurosurgery. And this lets you take a look at sort of the ground truth of when you do these TMS pulses, how are neurons in different parts of the brain actually reacting? Are they doing what you would predict them to do? Are they doing something different? How is the effect actually occurring in the brain? [00:20:00] And the second piece is developing noninvasive tools, mostly based on EEG sort of scalp electrodes as a tool that could be used more regularly in the clinic to guide and personalize treatment. Obviously we don't want to be implanting electrodes in people's brains if we can help it.

(20:19):

And I'd love to start by talking about the single neuron recordings, what you're learning there, and then we can shift into talking about developing these noninvasive biomarkers. Maybe we can start by talking about [00:20:30] a pre-print that you posted earlier this year on bioRxiv where you recorded from neurons in different parts of the brain while you gave TMS stimulation in this DLPFC dorsolateral prefrontal cortex sort of executive control region. Can you give us a top line of some of the things that you've learned from that study? Were the neurons responding in the way that you expected or not so much?

Corey Keller (20:59):

Yeah. I'd love [00:21:00] to step back a little bit. The broad strokes goals of the lab are twofold as you defined, and they're connected, right? I think it's this thing that has driven me is we don't know what TMS is doing in the brain and I want to understand it. So how do we do that, right? We can take advantage of this access to the brain that is involved in these epilepsy patients that have had seizures uncontrolled by medications that get electrodes placed deep into the brain. And then sometimes [00:21:30] have micro wires that can actually pick up single neurons. And we can apply TMS during that time to try to understand this. Let's understand at the most ground truth level as we can in humans what TMS is doing to the brain.

(21:46):

And then let's translate that up to the scalp in a space where we can noninvasively bring something like EEG into the clinic. So if we had EEG biomarkers of neuroplasticity [00:22:00] as defined by single neurons in the brain, we can actually track that in real time. We can actually use that tool to actually close the loop and optimize treatments, right? So I think tackling both of those are really important. Understanding the mechanism and understanding what the noninvasive correlates are of those measures. So how do I think about it? In the end, I made a bet a long time ago, I probably should have went into [00:22:30] neurology and epilepsy surgery. My hope was that 20 years from now, we don't need electrodes placed in the brain in some cases, right?

(22:38):

And over the next 20 years, we will learn a large amount about basic neuroscience with these tools. But if we can get noninvasive brain signals to the point where we can really track these measures in real time, we will have something that will be very powerful. And so that's where I'm coming at it from, [00:23:00] that these patients with depression, with PTSD, we don't want to have to have them undergo surgery in order to understand what TMS is doing to their brain and actually help improve it. I want to bring these, have really clinically usable tools that could actually be valuable.

Nicholas Weiler (23:18):

It is amazing how much we've learned in human neuroscience from patients with epilepsy who need to go through neurosurgery for one reason or another, who generously volunteer their time while they're [00:23:30] undergoing treatment of their own to advance the field. So let's talk a little bit more about that science. I'd love to hear about, you have this new study. This is a pre-print on bioRxiv right now, I think. So it's presumably going through the peer review process as we speak. So basically you're recording from neurons in various parts of the brain. You give TMS stimulation in this executive control region, the DLPFC, dorsolateral [00:24:00] prefrontal cortex, which is the normal target for TMS for depression. What are some of the top line findings here? What are you learning about how that region and other parts of the brain actually respond when you give TMS? This is the thing that you've wanted to understand for a long time.

Corey Keller (24:15):

Yeah. So we're really excited about this research direction. This is the first of many studies to come out of some of these experiments that we've been doing. And it's worth noting here we asked what types of neurons respond and where to [00:24:30] a single pulse of TMS. So again, this is just 1 of 600 pulses that we normally deliver in a three minute treatment. So here we're just looking at what is the brain's acute immediate response a few hundred milliseconds after a single pulse of TMS. And so the top line findings were almost 50% of the neurons that we sampled from, and we were looking pretty deep in the brain and pretty broadly, responded to a single pulse of TMS. [00:25:00] That was actually pretty surprising because we were sampling really deep in the brain.

(25:05):

And this idea that TMS is only superficial. Yes, TMS only gets to that superficial region, but then it propagates throughout the brain very quickly. And even within 10 milliseconds, the thalamus, a very deep region becomes activated. Single neurons fire as opposed to when you applied single pulses of TMS to the prefrontal cortex. Are there interesting findings? So this idea of the DLPFC and [00:25:30] subgenual cingulate relationship we talked about earlier, subgenual cingulate is part of the limbic circuitry. And so the longstanding question is when you activate the DLPFC, do you have direct inhibition of the subgenual cingulate and other limbic structures? Or does it get propagated through another region?

Nicholas Weiler (25:53):

So the limbic system, generally thought of, the brain's always more complex than our simple definitions, but generally [00:26:00] thought of as being involved in emotion, memory. It's got things like the amygdala, people have probably heard of that's involved in sort of threat assessment. It's got the hippocampus, which is involved in memory. It's got this subgenual cingulate and other areas. But when we say limbic, we're thinking sort of emotion, memory, things like that.

Corey Keller (26:18):

Yes. Thank you. So what we found was the first types of neurons that come online were actually those very deep thalamic putamen striato-thalamic [00:26:30] circuits and not the limbic structures. The limbic structures turned on about 300 milliseconds after that single pulse of TMS. The types of neurons that were activated in the striato-thalamic circuitry were inhibitory neurons. So the inhibitory neurons were activated early. And then later, the limbic excitatory neurons became suppressed. So none of this is causal. [00:27:00] None of this pathway is causal right now and we still need to work it out. But it may be that yes, those limbic circuits do get inhibited when you apply a single pulse of TMS to the DLPFC as we sort of long thought, but it might be through this indirect pathway through the thalamus. And that is a relatively new finding that, again, needs some causal validation and future work, but I thought was pretty interesting.

Nicholas Weiler (27:28):

We don't hear enough about the thalamus. [00:27:30] This is me because I was really interested in the thalamus when I was in graduate school, but it's this amazing node that brings in all the sensory information passes through the thalamus. It's like the grand central station of communication through many different parts of the brain. So it's so interesting that you see, okay, we stimulate in right behind on the forehead, right behind the forehead prefrontal cortex. We see these inhibitory neurons in the thalamus get activated [00:28:00] and that seems to tamps down activity in our target area in the limbic system. So it's sort of like you are activating this sort of control center to tone down activity in the area you're trying to hit.

Corey Keller (28:16):

Yeah. And we'll see. It's one study. It's a small number of patients. These are epilepsy patients. One of them happened to be depressed, but there's a lot more to do here. And I do want to clarify that this is just a single [00:28:30] pulse of TMS. So we have a lot more to do in terms of understanding what does a single dose of TMS do. There's 30 doses in a treatment course. How do those doses accumulate over time or induce what we call metaplasticity? And eventually, how does any of this relate to the clinical response?

Nicholas Weiler (29:00):

[00:29:00] Okay. Well, let's talk about what you were just saying about sort of how do we get this to a state that's more clinically useful. We're learning a lot about some of the mechanisms, how are the circuits impacted. A lot of this is actually being done in the Koret Human Neurosciences Lab at the Wu Tsai Neurosciences Institute, just to give a little institutional shout out, where you can do both TMS and EEG at the same time in research participants and see what can we measure without [00:29:30] going through the skull, without going into the brain, what can we measure about what's going on when we give TMS? And of course the trade off is with EEG, it doesn't give you nearly as much information as electrodes deep in the brain.

(29:46):

I often like to compare this to imagining you're at a sports stadium and the deep electrodes are like having a cameraman going around and interviewing individual people in the crowd. And EEG is like having a bunch [00:30:00] of microphones hung up over the stadium. You can probably figure out which sections are cheering and maybe you could get a sense of what's going on in the game, but you don't know what the individuals are doing and you certainly don't know who's having a hot dog and who's having a Coke or whatever.

Corey Keller (30:15):

That's a great analogy.

Nicholas Weiler (30:16):

It helps me to think through all this. So what are some of the things that you are able to see using EEG that you think is going to be clinically useful for guiding TMS treatment?

Corey Keller (30:28):

Yeah. So I'll say a few things. So [00:30:30] back in the epilepsy patients with those single neuron recordings, we have simultaneous scalp EEG recordings. So at the same time as we're seeing those neurons fire deep in the brain, we can see what's happening with those microphones. So we have that sort of transfer function that we'll be making that when this brain region comes online or goes offline or plasticity gets induced, we can ask what's happening at the scalp. So the goal here is preclinical [00:31:00] neuroscience, even motor physiology has been revolutionized based on the idea that we have a tool to measure cortical excitability and cortical plasticity. In the prefrontal cortex in humans from outside the brain, we don't have that measure.

(31:20):

We can't get it from scalp EEG. We can't get it from fMRI right now. That really directly relates to some measure of prefrontal excitability. [00:31:30] We have some sense using TMS paired with EEG that we might be able to get it. So this idea of perturbing a brain region and then measuring the brain's response is something that is very basic and fundamental if you talk to an animal neuroscientist. And that's what we're doing outside the brain. So we're applying that same single pulse of TMS to the prefrontal cortex and measuring with six electrodes right under the stimulation coil within 20 to 50 milliseconds. So think of it as a sonar. [00:32:00] You're pinging the brain and you can measure that brain's response and you see this sort of biphasic deflection within about 50 milliseconds.

Nicholas Weiler (32:08):

I was going to say, when you talk about excitability, you're saying like how sensitive is this brain region? What are you thinking about when you're talking about excitability?

Corey Keller (32:18):

You can think of it in terms of responsiveness. How responsive is the brain to a single pulse of TMS at that site? We're probably like [00:32:30] we are exploring and understanding, we're probing the corticothalamic-cortical circuitry and we're bouncing the signals down deep in the brain and measuring how responsive those brain circuits are. It's still understanding what that truly represents because again, we're outside the brain. And so this is where you can only get to some level of mechanism, right? [00:33:00] If you really want to see like this type of neuron is online or offline, you have to have invasive brain recordings.

Nicholas Weiler (33:07):

Right. Okay. But you have this sonar system. You can sort of ping part of the brain and record with the EEG electrodes to see how excitable is this particular area. What are some of the things you're learning that you think are going to have a clinical implication?

Corey Keller (33:23):

Yeah. So a lot of the work in the lab is we've seen this signal, but it is a really small signal compared to all of the other signals [00:33:30] that we see. And TMS is a loud auditory click associated with it when the capacitor discharges. It's relatively painful and you feel it on the scalp. And each of those sensory stimuli have an input into the brain and you have a brain response as a result of that. Each of those have different time windows and sort of signatures associated with it. But needless to say, there's a lot of artifacts, both neural and non-neural, [00:34:00] that come along with that more direct measure of cortical excitability or cortical responsiveness.

(34:08):

And so what we've had to do is build a whole system around it where we, in real time, can change coil angle, location and intensity to try to minimize those neural and non-neural artifacts and maximize the signal of interest to the point where an individual now comes into our office, into our [00:34:30] lab, and we can get that measure of cortical excitability on a single subject level. And that's something that, again, in noninvasive neuroscience isn't that typical. Usually you collect 30 patients or 30 people, average the signal and you get some sort of effect. We really want to do this on a single subject basis if we're going to think about using this signal to personalize treatment.

(34:54):

So there's a lot of closed loop optimization. There's a robotic arm that holds the TMS coil. We have real time EEG [00:35:00] processing and then a feedback loop, that within five minutes or so, we can optimize that signal. So we have a starting point, right? We have that measure of cortical excitability and plasticity is typically measured by the change in that cortical excitability after the intervention. And the intervention here is that 600 pulse three minute TMS treatment, if you will. So we're measuring that piece of cortical excitability before and after and seeing [00:35:30] how that changes.

Nicholas Weiler (35:31):

And are you ultimately able to get a signature of a change in cortical excitability? Have you connected this to the effectiveness of the treatment? I mean, because that seems like that would be the holy grail to use the standard metaphor that you're looking for here, which is here's what we can... We've done all this engineering. We can now get before and after, here's the excitability in this region. And if it looks like X, that [00:36:00] means we've effectively treated this and this person's going to experience a lot of improvement in their symptoms.

Corey Keller (36:06):

We're getting closer. So I think we're now at the point where we get really nice signals and we are now testing what is a single dose of TMS due to the brain. And we're starting to analyze that data to see if there are consistent, reliable effects using that measure of cortical excitability. Linking that to the clinical symptom change is the study that is undergoing right now. And [00:36:30] so it's just a different set of resources that are needed to deliver a full treatment course, which can take a long time compared to a single dose of TMS and measuring the more acute changes in excitability that we see.

Nicholas Weiler (36:47):

I mean, so yeah, you've got this clinical trial underway. So we definitely want to have you back once that comes out and you can tell us exactly what aspects of how you do the stimulation [00:37:00] is really predicting the patient's response. But there have been some interesting findings. One of your colleagues, Jessica Ross in your group was working on where they actually played music during stimulation, but she played a bunch of funk and R&B music and saw that the effectiveness of the stimulation or we should say the excitability in the brain changed depending on whether the TMS was [00:37:30] timed to the music. Is that right?

Corey Keller (37:32):

That's right. And Jess is actually starting her own lab at the Palo Alto VA if anyone's interested. She came in with a background in sensory motor neuroscience and brought that into the lab and wondered, what if we play music during TMS? Taking a step back, when people get treatment for depression, they're staring at a wall. Some of them are watching Netflix, some of them are listening to music, some of them are falling asleep. And we know [00:38:00] from our preclinical work that you have to activate the brain circuits when you stimulate them to actually induce neuroplasticity maximally. So I think we're very much under utilizing what TMS does because we're not putting those patients in that optimal brain state.

(38:17):

So that's been the last couple of years in our lab we've been focused on three different paths. One is playing different types of music and I'll talk about that in a second. The second is Umair Hassan is a sleep physiologist. So we've actually had people [00:38:30] nap and applying TMS during certain sleep stages. And sleep, as we know, is a period of maximal plasticity and consolidation. And so could we actually get a larger bang for our buck if we applied TMS during sleep? And more typical cognitive tasks that activate the prefrontal cortex. And if we apply TMS at those specific timings where we know that brain is active, can we get a larger effect? So in terms of the musical beats, [00:39:00] we've sort of methodically stepped through these different findings.

(39:05):

The first was in the motor cortex because there's actually a lot known about how music shows up in the sensory motor system. And so we applied single pulses of TMS at different timings with respect to the latency of the beats. And we saw that when you apply TMS just before each beat, which is a period of known maximal excitability of the brain, you [00:39:30] could actually get a larger evoked response. We've now translated that word to the prefrontal cortex using our measure of cortical excitability. And we see a larger cortical excitability signal at the specific timing right before each beat, if we apply TMS right before each beat compared to on the beat or after the beat. That's looking at cortical excitability.

(39:51):

And then there's the plasticity question, right? When you apply that three minute treatment, can you get a larger change in cortical excitability [00:40:00] if you were to play music, if you were to be asleep, if you were to be doing cognitive tasks? And in Jess's case, in the motor cortex, when we applied music during the TMS, we got a much larger plasticity effect and saw it in every single subject. So we're pretty excited about what could be done. And again, I think we are very much under utilizing what TMS can do from a treatment perspective partially because of the brain state, [00:40:30] partially because of the pattern of stimulation can also be personalized.

(40:34):

There's just a lot of parameters that we really haven't dove into, mostly because we haven't had the tools to measure it in such an acute way. Most people do the full TMS treatment and take a picture before and after using fMRI. And it's hard to understand that black box of each single pulse, each single dose and how they all interact. Once we get down to that level, we can do some pretty interesting neuroscience.

Nicholas Weiler (41:00):

[00:41:00] I love the idea of how do you get the brain in a receptive state for a treatment like this? And then also, how do you know when you're done, right? How do you know when you've done the thing that you set out to do, that you've got the plasticity that you were looking for? And I wanted to take a step back here just as we sort of come towards our close. One of the things that we talked about a little bit early on is the technologies [00:41:30] you're developing, these ways of measuring brain activity, measuring excitability, measuring plasticity in the human brain, particularly in these sort of more high level prefrontal areas, have a lot of implications beyond just TMS, right?

(41:44):

You could imagine looking at this in a different way with medications, with SSRIs, with psychedelics, with deep brain stimulation, maybe even with therapy. I don't know. I mean, it seems like the possibilities here are pretty broad [00:42:00] once we have a tool to say, when we do this intervention, how is that affecting the brain? And how does that link up to actually helping the person with their symptoms? I'd love to hear you just sort of imagine how this field looks in, I don't know, 5, 10, 15 years using this more sort of neural data, using these biomarkers. What does a more neuroscience oriented psychiatry [00:42:30] look like to you?

Corey Keller (42:31):

Yeah. I really do think this has pretty broad implications, not just in depression and not just with TMS. The idea of measuring neuroplasticity and using that to optimize treatment is something for any brain disorder that TMS is either FDA cleared for or being investigated, which is almost all of them. For any brain region, we've applied this to parietal cortex, to prefrontal cortex, and then we're now looking at the visual system. Can we optimize those signals, get measures [00:43:00] of cortical excitability, and then measure neuroplasticity with that change in cortical excitability? But really any disorder, any brain region, and the questions are the same, whether it's ultrasound, temporal interference, TDCS, TACS, ECT, SSRIs, you name it.

(43:20):

The question is, do you apply an intervention and how does that change the brain? This is just a different way to measure those brain changes that I think will be very valuable in the future. [00:43:30] So what does this look like five years from now? My hope is if someone's going to get TMS treatment, they might get a functional MRI as they do for the SAINT protocol to identify where in the brain we should be stimulating, where in the prefrontal cortex we should be stimulating. And then there's a day of a stimulation screening day where you run through with very, let's say, one second or two second chunks of [00:44:00] different stimulation parameters and different brain states, music, cognitive tasks, theta births, beta burst, different continuous forms of stimulation, different intensities.

(44:11):

And after each one of those, you're measuring how did that cortical excitability change? And is that moving in a direction that if you applied that over and over and over again, would it move in the direction of a less depressed state? And then you would choose that treatment paradigm for that [00:44:30] individual in the beginning and maybe check a few times during treatment. And you might need to adjust those stimulation parameters as that person's brain actually changes. So I'd like to see pieces of that incorporated in over the next 5 to 10 years. I think that would be pretty amazing.

Nicholas Weiler (44:49):

All right. Well, thank you so much, Corey, for joining us on From Our Neurons to Yours. It's been a pleasure.

Corey Keller (44:54):

Thanks, likewise.

Nicholas Weiler (44:56):

Thanks again so much to our guest, Corey Keller. He's an associate [00:45:00] professor of psychiatry at Stanford Medicine and at the VA Palo Alto Health Center. To read more about his work, check out the links in the show notes. And if you enjoyed the episode, please be sure to subscribe for more conversations from the frontiers of brain science. We also love hearing from listeners. If you have thoughts about the show or questions about the brain you'd like to hear us discuss in a future episode, send us an email. We're at neuronspodcast@stanford.edu, or leave us a comment on your favorite podcast platform. While you're at it, we'd [00:45:30] love it if you could give us a rating and share the show with your friends. I know everyone asks this, but it's tremendously valuable for us to be able to bring more listeners to the frontiers of neuroscience. Coming up on From Our Neurons to Yours.

Jonathan Long (45:45):

I think that the way to think about peptides is these are like Lego blocks. It's a very modular unit. The amino acid as a basic unit is very modular and you can stitch them together in all sorts of different ways. And it's very easy to stitch them together. The chemistry is very easy. And so it means that testing combinations in a pseudo rational [00:46:00] way is now much easier.

Nicholas Weiler (46:03):

From Our Neurons to Yours is produced by Michael Osborne at 14th Street Studios with sound design by Mark Bell. Our social media strategy is by Julia Diaz, additional editing by Nathan Collins. Our logo was designed by Amy Garza. I'm your host, Nicholas Weiler. Until next time.