NIMH Multimodal Brain Stimulation Speaker Series: Colleen Hanlon, PhD
>> Now, I’d like to introduce Lorenzo Leggio, who was kind enough to provide an introduction for Dr. Hanlon.
>> Thank you Tom. So it is our pleasure to introduce our first speaker today, who is Dr. Colleen Hanlon from Charleston. Dr. Hanlon is an associate professor at the Department of Psychiatry joint with the Department of Neuroscience at the Medical University of South Carolina. She’s also the Scientific Co-Director of the Brain Stimulation Division at the Medical University of South Carolina.
>> Colleen received her PhD from Duke University. Then she moved to Wake Forest as a post-doc fellow. And after a brief junior faculty appointment as an instructor she moved to MUSC as an assistant professor where she was on faculty as an assistant professor for five years funded by NIDA through a K01 and then when the K01 was over she was promoted to an associate professor which is her current faculty position along with co-director of the Brain Stimulation Division.
>> And she received an R01 grant from NIDA and NIAAA recently and is now actively is a PI in a NIDA grant R01. She’s also a PI in another NIDA R21. She’s also a PI of a component of the MUSC Charleston Alcohol Research Center which is funded by NIAAA, and she’s also a project leader of a P20 which is investigating the neurobiological basis for loss of in cortical laterality in chronic stroke patients.
>> And Colleen has many great achievements, including she has already published 44 papers, 50% of them as either first or last author. She has received many awards – you will forgive me for not mentioning all two pages of awards from your CV, I will just mention two from HoBP and ACNP. She also served on the editorial board of the Experimental Neuropsychopharmacology journal. She’s an associate member of ACNP and in addition to her research she does a lot of teaching at MUSC where she has already established herself nationally and internationally for her expertise on neural imaging and brain stimulation. She has given 60 talks both nationally and internationally. To summarize, basically, Dr. Hanlon’s overall theme in her scientific research has been to map neural circuitry irregularities in alcohol and substance use disordered patients and then to modulate these circuitries using brain stimulation techniques or neural feedback.
>> So with that, I’m going to welcome Dr. Colleen Hanlon today. The title today will be “From Mapping to Modulation: using intrinsic and neural architecture to develop TMS as a new treatment for craving and control.”
>> Colleen, thank you so much for being here and we look forward to your talk.
>> Thank you all very much. Thank you for such a nice introduction, that was great. I’m extremely happy to be here. Let me see if I can get my talk started. And today, first actually I want to thank NIDA and NIH and NIAAA and whoever else, and NINDS, that have helped organize this.
>> I know the neural modulation unit here at NIMH has organized this great symposium and I heard that hundreds of people are potentially watching online which I’m sure doesn’t make the slightest bit nervous. But I will do my best.
>> But anyway, I really want to say that it's great to see all of the enthusiasm from everyone about really trying to develop TMS as a rigorous evidence-based treatment for various psychiatric and neurologic disorders. What I’m going talk to you about most today is about addition.
>>So addiction can be seen as a disease of both cravings and loss of control. And that will be a conceptual model and theme that kind of falls through the talk.
>> So the sort of obligate conflict of interest slide -- all the work that I have done has fortunately been funded through the NIH and I'm very thankful for that. And there's a ton of people who have contributed to this work and I just happened to be their investor.
>> To start off, what do we know about the brain? Well there’s many models of neural circuitry. Certainly in the last decade or two we have had significant advances in our understanding of frontal-striatal circuitry in the brain.
>> We know that there are many dopaminergic circuits in the brain, one of which is the mesolimbic dopamine circuit. This is supposed to be involved in motivation, vigilance, and sometimes this is called “hot cognition”
>> Additionally we have the mesocortical dopamine system. This is more involved in executive processing, response in addition. Sometimes this is referred to as “cold cognition.”
>> And then we have the nigrostriatal dopamine system. And although I spend a lot of my time thinking about the mesocortical and mesolimbic systems, as we do in the department of psychiatry -- if I were to walk across the street to the department of neurology or to NINDS, they, in fact, spend a lot of time thinking about the nigrostriatal dopamine system.
>> And of course we know that these dopaminergic circuits are sort of kind of go through pre-clinical and clinical models. Suzanne Haber and Brian Knutson have done a really elegant job on their review article in Neuropsychopharmacology about nonhuman primate track tracing work of these circuits. Another famous paper by Alexander, DeLong and Strick has kind of sort of told us that these are parallel and integrated circuits. And this set up a lot of work in the field.
>> And then a more recent paper, in Nature Reviews Neuroscience was also very well done and again it mimics this same idea of these parallel and integrated frontal striatal circuits and their relative control over behavior.
>> One of which is more associated with limbic tone and drive, another is executive control, response inhibition, and another one is, in fact, you could argue behavior. All of behavior requires action.
>> So we probably should not neglect using psychiatry for the role of the nigrostriatal dopamine system.
>> What we know about cocaine? Well we know that cocaine directly affects dopamine in the striatum. Oand f all the substances of abuse, cocaine is one of the few that directly affects dopamine causes a large change in dopamine binding.
>> Of course methamphetamine is potentially even more efficient in this way.
>> Well my post-doc advisor Linda Perino demonstrated in a non-human primate model that over time, when you use cocaine, the effects of cocaine initially begin in the ventral aspects of the striatum, which is a more limbic domain, and then over time the effects grow out to sort of more dorsolateral aspects striatum. This is true not only in changes in dopamine but also changes in glucose consumption and other measures of excitation.
>> We then show that in humans, individuals who are chronically using cocaine also have these changes in the dorsal striatum.
>> And in fact, in individuals that have been using cocaine for 15 or more years, you lose the typical functional segregation between the ventral and the dorsal striatum.
>> Then, as neural imaging techniques have been grown through a lot of work that has been done here, at NIDA and NIH Intramural, we now know that there is all this sort of functional connectivity networks that exist in the brain. And through the addiction process, this is a figure that I pulled from a recent very good review by Powers et al, it shows that the engagement of various baseline functional connectivity circuitry in the brain changes over the course of the addiction process.
>> So to those of you who are not addiction researchers in the audience, addiction is seen as a progressive disease, one that behaviorally starts with a vulnerable individual who is given sort of opportunities to enjoy the reinforcing qualities of the substance, like cocaine. And initially drug-taking behavior is mediated by reward-based mechanisms.
>> But then over time, there is an evolution, and drug taking behavior becomes much more habitual. Much more behaviorally and ritualistically driven.
>> And what this figure shows here is this cascade from incentive cravings and goal formation which typically occurs in the early stages of the addiction process, eventually to actions and habitual control, which involves different stages of neural architecture.
>> So as you can see, there's now a lot of research on functional imaging, functional connectivity, and the role of these different cortical, subcortical networks in different stages of addiction.
>> What one thing is, is to spend a lot of time, like I did for most of my career and still do, just mapping the brain; just looking at these individuals and saying what’s different with their brain? And it's even more exciting to try to change that behavior.
>> And now in the last two years, we have a number of studies that you can see here -- that have used functional imaging in individuals that are undergoing rehab for various substances.
>> So this is a summary of all the studies we could find as of last year. That have looked at baseline brain activity to either executive control task, those are the ones in blue here, or to a limbic arousal task and looked at the ability of their brain activity at baseline to predict their treatment outcome.
>> What you can see here in this summary figure -- is that when engaged in an executive control task, the areas in blue -- they have decreased connectivity, in those areas, individuals are more likely to relapse. Individuals that present with relatively high connectivity in the red areas, also have an increased ability to relapse.
>> So that suggests that potentially there is a predictive role of baseline brain activity to eventual treatment outcomes.
>> That sets stuff up really nicely.
>> The preclinical groups have made tremendous strides in the last decade or two with really elegant molecular tools where they can casually manipulate these circuits.
>> What many have shown is, here's a sample of some of the papers that have been particularly influential, is that through optogenetic stimulation, of the pre-limbic versus infralimbic cortex, you can in a causal manner, increase or decrease cocaine self-administration.
>> So what that means is these same brain regions, the ventromedial prefrontal cortex, orbital frontal cortex, nucleus accumbens, all these limbic cortical/subcortical connections have a causal role in drug self-administration.
>> So we as human researchers who are looking to really translate these data into something we can actually deliver to the patients, would really like to see if we can do the same thing in humans. Can we take TMS and in the same way causally change activity in these networks and then change eventual treatment outcomes.
>> So that’s pretty much the goal of my lab. For the last few years and hopefully for the next few years to come, is to develop TMS as an evidence-based neural circuit specific treatment for cocaine addiction.
>> And we're getting close.
>> And I will tell you for the rest of the talk about logical progression, we’ve done at least six studies that logically have followed one another and have led us to a point where we are now running a double blinded placebo-controlled clinical trial of TMS as a treatment for cocaine users.
>> So the strategy that we have chosen is to stimulate the ventral medial prefrontal cortex. By quieting down this brain region -- that's kind of the conclusion. To just kind up whet your appetite and I’m going to tell you how we got to that.
>> Okay so the general outline for this presentation -- is to first present you with the conceptual framework for developing TMS as a treatment tool for addiction. It's already an FDA approved treatment tool for depression and there's a lot of people out there who are trying to optimize various parameters, etc. But in addiction, we’re way behind the depression crowd. We are probably way behind the pain crowd. But we have a lot of really good pre-clinical evidence that actually can help lead us in the right direction.
>> So first, I’m going to talk about a framework, and then I want to talk to you guys a little bit about some patient inspired method developments that we have done. Because the audience probably has some baseline knowledge of TMS, I am kind of skipping a lot of the basic TMS stuff I would talk about. But if anybody out there would like to learn more about the basics of TMS we have a class and there are several very good classes across the country where we as a community are really trying to keep rigorous methods of development requiring TMS.
>> And I think that we will start to see that more in K01 training plans and S32 training plans, so I highly encourage you to contact me or frankly any of the people on the list of speakers coming to the seminar series.
>> Finally I would tell you about what exactly we are doing for our patients.
>> Okay, so again, dopaminergic disorders typically we think about the mesolimbic system, which is the drive for cues, the drive for drugs or rewarding stimuli, and then your executive control system, which are kind of the brakes to that process.
>> Cocaine and alcohol affect the striatum, and therefore change drug self-administration into the working hypothesis that we’re using our labs, to say that in general, potentially what causes some people to sort of be vulnerable to becoming substance dependent, is that they have a disruptive ratio between executive control circuits and the limbic circuits.
>> So the way I really like to sort of talk about this and explain to people is through an analogy with alcohol, because as we don’t societally endorse drugs, many of us have had experience self-administering alcohol. Potentially more than we should on some occasions, and hopefully we have also had occasions where we have said no, in the face of an alternative reinforcer.
>> And so if you think about a time if you were kind of let's say at a bar with friends and colleagues, and the time is getting late, and then somebody in the crowd says “One more round!” You might think to yourself, “That would be so good, it would taste great, I would get socially reinforced for it, it’d be free.” That is your reward system. That's the red system there.
>> Thinking about all the positive hedonic aspects of it.
>> And then hopefully you have another part of you that says “Oh, I shouldn't. I have to drive home. I have to give a talk tomorrow. My significant other will be mad at me.” And so you weigh the relative value, the incentive salience, of these two circuits.
>> And so it's not bad if sometimes you stay and you have another one, generally. But it is bad if you tend to always stay and have another one. And that's what leads you to sort of DSM criteria for substance dependence.
>> And so what we think is that there is at least two fundamental ways we can think about developing a treatment for cocaine and potentially for alcohol as well.
>> One option of course is to decrease the limbic arousal so in the presentation of a line of cocaine, or a crack pipe, or just another drink, the relative value of that reinforcer is lower.
>> So you don't seek it.
>> The other option is just to increase your ability to resist. So in some ways the incentive value, or the salience, of the reinforcer may remain the same. But your ability to walk away from the system is much stronger.
>> And if we think about those two general ways to conceptualize it, we can use TMS, which has frequency dependent effects on cortical excitability, as a tool to either increase activity in this executive control circuit, through something like 10 Hz or 20 Hz stimulation, or intermittent state-averse stimulation, or we can try to decrease activity in this limbic arousal craving circuitry with LTD like stimulation with something like 1Hz, 5Hz or maybe continuous state-averse.
>> Well if you look at the number of studies have been done for rTMS and addiction, and I apologize in advance, it used to be really easy to keep this table up-to-date. This is a review article we published in 2015. And now it's actually exponentially increasing.
>> So thank you all out there for your interest -- it's awesome and I apologize if your name doesn't appear on the slide if it’s not here.
>> But in fact it's a really good sign.
>> And if we just kind of take a snapshot at about 2015, if you just sort of look down this column, most places say DLPFC. That's great! So the left dorsolateral prefrontal cortex, some people doing the right dorsolateral prefrontal cortex, seems to be overwhelmingly the site the people are trying to treat. It also happens to be the site that we treat for depression. So the standard clinical treatment for depression is approximately 10 Hz, given over the left dorsolateral prefrontal cortex, and sometimes the right. For several weeks at a time.
>> Well that's great. We have a big brain out there and part of me just wants to explore other places.
>> So there's been a lot of great work coming out of Toronto for the last few years. Stanford groups are doing a lot of great work now where they’re exploring other aspects of the cortex.
>> Most of these studies are still being done in depression, because depression is leading the field. But in fact, there's lots of arguments that for addiction we should maybe stimulate the limbic circuit directly.
>> So we’re interested in trying some more of the medial wall studies. We are keeping our eyes open to the fact that a lot of people are already getting pretty good data with dorsolateral prefrontal cortex.
>> Sometimes I like to describe the dorsolateral prefrontal cortex as the magic brain button because it seems to share everything with our TMS and I say that tongue in cheek but it may actually be true.
>> But anyway, we will see if the field emerges.
>> And so my favorite scientist, perhaps, Sherlock Holmes said “It is a capital mistake to theorize before one has data. Insensibly one begins to twist fact to suit theories instead of theories to suit facts.” So with this, we need to be evidence based and we need to explore the neural circuits. We have these great tools, functional connectivity to do mapping, have task-based neuronavigation. And these tools will point us to sort of a smarter answer than just doing what everybody else does because it seems to work.
>> Okay so that has been our goal. We have had six logical steps in order to get there.
>> Thinking about the possibility of stimulating that red circuit, the craving circuit. The first resistance that I had was from people that said well there's no way that you are going to be able to hit the ventromedial prefrontal cortex with TMS. Your frontal sinuses are in the way and the magnetic field decays very very fast.
>> So we thought about that and we addressed it. And then the next idea was to say, well is it even possible to try to distinguish the red circuit from the blue circuit.
>> Can we actually even parse them apart and be smart about stimulating them?
>> Once we did that, we wanted to say okay, now that we have these two circuits, are they different in cocaine users? Because we needed to know where to treat. Should we go to the dorsolateral prefrontal cortex like people have been doing or should we maybe go to the ventromedial prefrontal cortex?
>> Once we found an answer to that, we wanted to evaluate if it was possible to move to the circuit. In just a single day to induce transient LTP/LTD-like effects in the circuit that we chose, this connectivity between the frontal cortex node that we’re stimulating and the subcortical node.
>> And then finally, once we figure out that we can move the circuit around in a single day, we can evaluate whether or not it is feasible, safe, and if we have efficacy and durability of long-term TMS treatment.
>> Because again for the depression treatment protocols, you can simulate for at least six weeks typically.
>> And that's when you start to see behavioral change.
>> There’s a lot of reason to believe that in addiction particularly stimulants, it might not take that long. The preponderance of evidence in pain research suggest that you need about 10 days or 15 days.
>> So shorter is probably possible.
>> But we just wanted to see.
>> And then finally if I have time, I will talk a little bit about the area I think a lot of work is going on now, and a lot of the other seminar speakers actually have more expertise in this domain, is the role of individual variability. So in the spirit of precision medicine, and being able to choose individuals for states or sort of baseline cortical excitability states, their EEG rhythms, that might optimize treatment efficacy for patients.
>> What should we consider?
>> Okay, so, basic TMS 101. Stimulation depth. So it's a standard figure of eight coil -- the general dogma is that it takes about, standard figure of eight will penetrate about 2 cm, or 20 mm deep at motor threshold.
>> And generally the excitation is limited to the cortical mantle.
>> However, if you can get to the cortex, the cortex is only about 1.5 to 3 mm thick. Cortex is very thin. If you just take a moment to think about that. Everything that makes us human is only about 1.5 mm thick. I think that's actually phenomenal and worth thinking about. And so our idea, knowing these facts, is we just need to get to the cortex with TMS and once we get there we can penetrate all five layers do we need to do on a cellular level.
>> To sort of add to the group, recently I was speaking to a reporter and this person said, “Well, when you have a TMS coil, on the brain, about how much of the brain are you stimulating?” So I said “Oh, about 2 cm,” and she asked, “If that was the mass of the United States, about how much would you be stimulating? Like a region, or a city, or a county, or a state?”
>> So I thought “Well that is a testable question” so I ran back to office and I ran the numbers and sure enough TMS approximately 2 cm of stimulation through the surface area of the coil -- is approximately this great state of South Carolina if you were to put it on a map of the United States.
>> So now maybe you will never forget that because I said it.
>> And so that's a really good visual. So that's about how much, if we just remain very America-centric, then we can stick to about the state of South Carolina.
>> We actually just, I wrote a little letter to the editor in Brain Stimulation and I included some other countries in the diagram if you want to see. So that's kind of fun.
>> So back to science. So this was actually inspired by reviewers, because in fact reviewers tend to always know what they're talking about.
>> And these reviewers were very concerned about the depth, the distance from the cortex to the scalp. Particularly in the individuals with cocaine and alcohol use disorders, who have, you know, they’re using alcohol a lot -- there's a lot to suggest that alcohol changes grey and white matter density, it’s associated with atrophy -- so they were rightly concerned that maybe we wouldn’t be hitting the cortex. So we developed a tool to sort automatically unbiased calculation of the distance from the scalp to the skull and as you can see here, if you look at F3, this is the DLPFC position, and FP1, this is the frontal pole position, actually the distance is pretty close.
>> So the strength of the magnetic field falls off very quickly with figure of eight coils, maybe one over cube root. And so it's very fast and every little bit of distance matters. But thankfully the distance isn’t too different.
>> This is just an image again of a representative control and representative cocaine user, and how we measure the distance.
>> This is a cocaine user here with obvious atrophy and sulcal enlargement so if you're not used to looking at heavy drug users, this actually looks very similar to the kind of brain you would see in an alcohol user. You see these very large sulci.
>> And so the atrophy though, in this particular individual, still was kind of preserved. There wasn’t very much atrophy at the level of the frontal pole. So we were fairly convinced that in this guy, we still sort of hit the cortex with the TMS coil.
>> So these are really important parameters to consider because the strength of the field falls off so quickly. So we now as a laboratory have begun to include scalp to cortex distance as a regressor in all of our analyses.
>> And I could definitely see that in general, not only would we use them in the future as regressed post-hoc, but we can use them as dosing metrics pre-hoc -- so you can figure out the dose of TMS that you need to give based upon their scalp and cortex in the various spots.
>> Of course this isn't an issue if you're studying motor control because your dosing is your MEP. But again, just because the motor threshold is at a certain value, that doesn't tell you how much the frontal cortex is going to, the distance will play a factor.
>> And so an outstanding engineer that used to do preclinical work with me, Phillip Summers, he did clinical with my colleague and he was looking at vessel dilation using two photon spectroscopy, and he developed a tool for us to calculate the surface fir and the volume of CSF that are present in various locations.
>> So this tool is called Brain Ruler, and it’s free, open source, and it's available on MathWorks. You can download it, it runs in MATLAB and is fairly easy to use and has a GUI.
>> Okay so another thing we wanted to figure out specific for our patients in developing an evidence-based treatment for these cocaine users, was can we actually get to the subcortical brain regions if we choose to stimulate the DLPFC or if we choose to stimulate the ventromedial prefrontal cortex.
>> So we know from a lot of great work that comes before us that TMS induces activity at the side of the coil as well as monosynaptic striatal afferents. There’s been a number of very elegant studies that have been done. Some of my favorites have been by Strafella’s group.
>> They’ve shown that with different frequencies of stimulation, you can change dopamine findings in the striatum with cortical TMS.
>> And this is great. When we move again to the prefrontal cortex, these things become a little bit harder. So the dorsolateral prefrontal cortex is again, the FDA-approved treatment for depression, and Strafella’s group also did some dopamine binding studies looking at the effects of DLPFC stimulation on striatal dopamine and he saw a frequency-dependent change. But we weren’t so sure about the medial prefrontal cortex.
>> Theoretically it should follow, but we have all these other issues and constraints that might happen.
>> Because they weren’t necessarily sure that if we put a coil on the frontal pole right here, that we would actually be able to activate the striatum because it just hadn't been done.
>> Additionally, we were wondering -- a lot of studies come before us that use TMS in the MRI scanner, to sort of apply a pulse at the level of the cortex and watch happens at the level of the striatum and they would see changes in BOLD signal in the striatum but they would also see changes in BOLD in the stimulus, in the motor cortex, in the temporal lobe, all over the place. And of course as good scientists, we want to make sure that that is a causal change and not just related to the other aspects of receiving TMS.
>> If you ever had TMS itself, just single pulses, you know that it’s a little weird. Especially if it comes at an unexpected time. It can be associated with movement, even if you put it over the DLPFC or the medial prefrontal cortex because you want to move away from the coil.
>> So there's all the sort of human subject factors.
>> And the field is generally very good and very rigorous about insisting that everyone in the field use good active sham controls. And yet we don't insist that people who do interloop TMS fMRI have a sham control condition, myself included.
>> And so an outstanding graduate student in my lab, Logan Doggel, set out to kind of take on this task.
>> The coil, just to tell you a little, so MUSC was the first place in the world to take a TMS coil and put it into the MRI environment and apply pulses and this is done by Mark George. He is the director of the Brain Stimulation Division at MUSC and my boss. And he did a lot of this pioneering work. He was actually in this building where he did a lot of his training before.
>> So he at that time developed this coil holder here. So this is an early TMS coil, this was Dansec who was later bought out by MagVenture, and this coil is actually really nice and light and small. And this is the coil holder so you can take this TMS coil, and you can position it wherever you want over the person’s head, and actually we’ll go all the way back to the parietal cortex. And the secret to this is our coil holder here.
>> So this is our TMS targeting system. I know that some of you are thinking of developing TMS in early TMS systems, and we are happy to have you call us or email us and we are happy to tell you what we know.
>> Okay. And so of course it's always nice to sort of see the data yourself. And so what Logan and I did, is we took the TMS coil and we put it over the dorsolateral prefrontal cortex. We wanted to see if in fact we could get a bone signal. And I, at the time, being relatively new to TMS thought, well clearly we will get a BOLD signal underneath the coil, because that’s where we’re activating, and we will get a BOLD signal in the striatum, because that’s the monosynaptic afferens.
>> Well it turns out that that doesn’t actually turn out to be true.
>> So what Logan did is he placed the TMS coil here on the dorsolateral prefrontal cortex, and he acquired data very fast with a TR of one second. The full story is that we had actually tried this already and we weren’t seeing BOLD signal underneath the coil. So we had a hunch that maybe it was because the hemodynamics underneath the coil were different from the hemodynamics at the monostriatal afferens. The way that TMS works, the mechanism of action, is really at the level of the axon rather than the initial cell body.
>> And therefore, the BOLD signal at the area that is directly stimulated is likely not the same as the shape of the BOLD signal in the striatal region. So if you naïvely do this, and run your data through FSL or SPM or AFNI, which assumes a canonical hemodynamic response function, you’re not going to see anything convolve with that.
>> So we did high resolution, so the one second TR. And we fired the TR, fired the pulse within the gap between each measurement at each volume that is collected.
>> We had 20 TMS pulses per session, and the inter-TMS pulse interval was 10 to 15 seconds.
>> And this is because we wanted enough time to see the BOLD signal rise and fall nearly completely.
>> So for those of you out there that are neuroimagers, this is like the world's slowest event related design.
>> And we modeled it that way.
>> And what Logan found, and hopefully this will be submitted soon, is that when he applied TMS to the dorsolateral prefrontal cortex, and you have sham stimulation, so in this case -- his sham condition or his control condition, in this case we exploited the idea of distance from the scalp to the cortex as our way to create a control. We took open cell reticulated foam and we pressed it down so it would convey the mechanical force associated with the TMS coil but it wouldn't have -- it would be far enough that the magnetic field wouldn’t penetrate the foam and get into the cortex to cause excitability.
>> So this sort of sensory aspect of TMS in the scanner would be consistent. But the direct TMS effects would not be, would be able to be isolated. So what Logan showed here in this percent signal change graph is this is the percent BOLD signal change on the y-axis, and this is time after the bold signal on the x-axis. So as you can see here if you're familiar, that the dotted line here is a typical hemodynamic response function. The dotted line was the hemodynamic response function that was associated with the sham. So in face sham itself caused a BOLD signal in the striatum.
>> But, the BOLD signal in the striatum associated with real stimulation was greater.
>> And it was greater; the difference was enough that we could actually detect it.
>> So if you look here, he also found that there was a dose-dependent relationship between amplifier output and the evoked BOLD signal in the striatum.
>> Suggesting that there is a causal link between the amount of energy that is going in and the amount of BOLD signal that is happening in the striatum.
>> So that got us very excited.
>> And here I won’t spend too much time on this, hopefully the paper will come out and you can read it. He did a region of interest analysis and what he found was that out of all these regions of interest, the only regions that were increased in BOLD signal associated with the TMS pulse in many regions, regions which were differentially activated by single pulses of TMS to the dorsolateral prefrontal cortex. Or the anterior cingulate cortex and the caudate.
>> To many of us in the field – a sigh of relief because that is great. Those are the brain regions that we depend upon, the brain regions that we build our conceptual models upon, and probably very relevant to neuropsychiatric disease.
>> So the next step was, great. So we think we can actually stimulate the dorsal striatum and that there may be dose-dependent effects on the bold signal.
>> We wanted to know, okay, can we discriminate this sort of mesolimbic from mesocortical, hot from cold cognition circuits.
>> We started with just healthy control. This is done with a team of people in the lab. We also developed a tool for rigorous positioning of the coil for this particular experiment. And you can find the tool -- it is freely available as well. And basically it's just a nice handy tool to use, to sort of find and locate the EEG 1020 system locations for various places you want to put the coil.
>> So as we expand out to applying TMS to different brain regions, these kind of tools will be increasingly useful in the field.
>> At least I hope so.
>> And again this is standard TMS BOLD experiment. Individuals are laying in the scanner and essentially it's a slow event related design where TMS pulse occurs every in this case about 10 seconds.
>> And what we find here is that when we position the coil over the dorsolateral prefrontal cortex, we seem to evoked BOLD signals on the dorsal aspects or lateral aspects of the prefrontal cortex as well as in some caudate on the ipsilateral side. And when we place the coil over the medial prefrontal cortex, in this case, the left frontal pole, we get evoked BOLD signal in the area directly underneath the coil or maybe even a little bit further back from the coil, as well as in the striatum itself.
>> I will say that we now have upgraded to a prisma scanner at MUSC and Logan has been sort of leading this project and playing with multi-echo imaging, so both multi-site and multi-echo, and I actually would love to redo this experiment with those parameters. So I do think that multi-echo imaging probably is going to really advance our ability to look at the brain through interleave TMS fMRI.
>> Okay great. So now we think that we can separate the circuits. We think that if nothing else, that putting the TMS coil over the DLPFC, we can get more dorsal aspects of the frontal striatal system. And by putting the coil over the medial prefrontal cortex you can get more ventral aspects of the limbic system.
>> And this is published as well. Okay so that's great. So now we are pretty convinced that we can apply TMS to the ventral medial prefrontal cortex, the frontal pole, and get changes in BOLD signal. There have now been a lot of groups that have done this. And, so it seems like it's sort of obvious now but at the time it wasn’t really obvious so we were happy to sort of see it.
>> Okay great. So now we have the healthy controls -- so we wanted to see what the circuit is like in cocaine users.
>> So this was led by Melanie Canterberry who was a postdoc of mine at the time. We had 20 cocaine dependent individuals, all whom have had failed quit attempts which is a very important aspect of cocaine dependence, and 20 age and gender matched controls. In this case, in the same way, we put the TMS coil over the dorsolateral prefrontal cortex and the medial prefrontal cortex.
>> In the controls, and we learned a little bit from the first study, now we’re doing many more TMS pulses. So it does in fact take probably at least 20 to 30 to observe a consistent signal.
>> And so here, we apply the TMS pulse to the dorsolateral prefrontal cortex and we see evoked bold signal in the controls in the dorsal striatum as well as in the cingulate, and in the controls we also see BOLD signal in the orbital frontal cortex or ventromedial aspect of the prefrontal cortex as well as ventral striatum. So we kind of replicated the same pattern that we observed in the first study which feels good.
>> In the cocaine users, DFPLC simulation evoked a similar pattern. So you see here, activity in both the dorsal striatum and in the anterior cingulate cortex.
>> So that's pretty interesting.
>> In the cocaine users when you put the coil on the medial prefrontal cortex, so you apply stimulation to the brain region which is typically mobilized by rewarding stimuli or evocative stimuli, the cocaine users have a much lower response.
>> So in a group comparison, you see there's actually not too much that's different between the users and the controls and their brain reactivity through a single TMS pulse when you stimulate the DLPFC. But when we stimulate the MPFC, there's a big difference. And that is, the cocaine users don't seem to mobilize the frontal striatal limbic circuit in the same way that the controls do.
>> So that was the area that was most different. So based upon some resting state data which also suggests that the circuit between the medial prefrontal cortex and the striatum was aberrant in cocaine users and the functional connectivity was a little bit too high than usual, we decided to use the circuit as the target for TMS treatment development.
>> Next step.
>> So we have a target, feel like we were fairly evidenced-based getting there. We have a lot of preclinical data to help grow our hypothesis and refine the ways we’re going to do it. So now our idea is can we move this circuit at all, can we induce a transient state of LTD in the circuit.
>> And so we did a study where we tried to decrease the brain response to cocaine cues. So craving is one of many factors that leads to relapse and it’s in fact one of the most important factors, and consistent factors, not only in cocaine users but transdiagnostically appropriate.
>> So we wanted to test two questions -- can we use TMS and LTD-like form of TMS to dampen activity in the frontal striatal circuit, and can we use TMS to dampen cue reactivity, so the way in which the brain naturally responds to these cocaine cues.
>> The protocol that we developed is called frontal pulse stimulation. And this is theta burst stimulation, an LTD-like form of theta burst stimulation, so those in the audience that spend a lot of time thinking about this, you may criticize me for saying that it’s LTD-like, because it’s all pulse number dependent. But for now we’ll just go with the facts, and the facts are that we use 3600 pulses of TMS to the frontal poll. It was delivered in continuous state adverse stimulation fashion. There were two trains, 1800 pulses per train and there was a 60 second inter-train interval. So essentially we had 1800 pulses, and then we paused, and then we did 1800 pulses.
>> The whole protocol can be conducted within five minutes.
>> In these studies, we did ramping of the TMS dose, which I will tell you about a second. We call this kind of a sandwich design where we put them in the scanner and we show them pictures of cocaine cues or neutral cues and we also use interleave TMS fMRI to look at their baseline brain reactivity.
>> We then gave them a dose of real or sham cTBS to the ventromedial prefrontal cortex and we put them immediately back in the scanner.
>> These studies -- there were 49 individuals enrolled in the study. And it was a thin subject design, so everyone received both real and sham stimulation.
>> So there's a lot of scanning and a lot of data.
>> This is actually what the protocol looks like so hopefully the sound will work and these data were recently published, available now online.
>> So we just recently published and we published in the supplementary material, our standard operating procedure for this procedure.
>> So I think that is really important actually. As I have had the pleasure of talking to you guys, and the people at NIDA yesterday, we are all in this together and we all learn a lot that we never actually publish very much because you don’t write everything up. The more we can start, like in our supplementary materials, publishing standard operating procedures, or kind of creating a forum or community where we can post tips that we learned, I think the stronger that we’ll become.
>> So this is published in the supplementary material of our recent paper and the video is published as well. So feel free to look it up.
>> So this is Oliver, my brave research assistant. And this is Scott, my other research assistant, and they will be demonstrating what we do for our patients.
>> What you’ll see in this video is, this is the TMS coil, this happens to be a MagVenture coil. It’s a sham coil, so it’s real on one side and sham on the other side. And it’s placed over Oliver’s left frontal pole. The cords to the AC running underneath it are electrodes, and the electrodes are engaged to help mimic the sham if the case is sham.
>> In this case it's real. And you’ll see Scott in the background turning the knob, and what Scott is doing is critically important to delivering TMS to the frontal pole. He’ll be ramping the dose -- so in our case the protocol does 1800 pulses followed by a break, followed by another 1800 pulses. Takes about 5 minutes.
>> So we have some time to ramp up the dose, versus the standard 600 pulse protocols, which are very quick. And so delivering them on the forehead can be very painful because you can't ramp the dose.
>> But if you have the ability to ramp the dose, you can an individual to tolerate up to sort of 110% or whatever it is you want to do, of motor threshold. So in this case, you will see Scott ramping up the dose slowly. And you will observe Oliver receiving the treatment.
>> I will say for full disclosure I’ve definitely also received this on my head many times. And they volunteered freely and willingly.
>> So you see Scott is ramping the dose in the background. I think I will zoom in in a second and you’ll be able to see Oliver’s forehead moving a little bit so because of the frontalis muscle on the forehead, you’ll start to see a little bit of eyebrow movement. So that’s Oliver saying he can kind of feel the difference, as an experienced TMS person, of the TMS pulses versus the sensation of electrodes.
>> And now we have reached the dose so that was the ramping protocol. And then we’ll continue for the full minute and a half for the 1800 pulses.
>> So that's what we do. And we have done this now on over 100 substance dependent individuals. We have had no serious adverse events. We have occasionally had people who will come in and they will come in for a couple of visits and they’ll decide that they don't want to do it anymore.
>> We certainly haven't had like the typical adverse events that you worry about with our TMS procedures.
>> So we feel good, it's always good to be cautious. So we are always cautious with watching and monitoring so patient compliance seems to be pretty good. The feasibility seems to be pretty good and I'm happy to report that the outcomes seem to be pretty.
>> The other key thing that we do in this particular experiment is to something while the participant is receiving TMS.
>> So relatively, I mean it's not really new, but a relatively new thing that people are talking about in the literature, is the role of task dependent plasticity. So we know from our pre-clinical friends that a primed neural circuit has a much higher plasticity potential than an unprimed neural circuit.
>> And you can verify that statement through a conversation about dendritic spines or treadmilling rates or cocaine self-administration profiles.
>> We know that engaging the neural circuit in the task makes it much more likely to manipulate that or that memory or whatnot.
>> So while people are receiving TMS for addiction, again our idea is to dampen down brain response to cues so we have them imagine their cocaine cues, or in this case their smoking cues or alcohol cues, whatever it is of interest.
>> So we go to the standardized script induction of the last time that they used. Describing the positive attributes of the experience, and we have them view the images while we are receiving the TMS.
>> We have them keep them to keep your eyes open as well which is sometimes a challenge, but it's generally pretty good.
>> Okay so what are the results? Initially a lot of people thought I was a little crazy to do this over the ventromedial prefrontal cortex because they had tried it maybe on people and can be very painful. But the secret is in the ramp. If you start low, you can bring people up to a clinically relevant dose. So we started small with a small pilot study of 11 individuals. These are cocaine dependent individuals and what we found, in this case we did interleave TMS fMRI as our probe. cTBS and LTD-like form of TMS to the frontal pole and real or sham within the subject. And TMS as a probe afterwards.
>> And what we found were specific decreases, this kind of blew my mind, in evoked bold signal in a accumbens, ventral striatum, and in the orbital frontal cortex. This is phenomenal. To anyone who’s actually looked through their neural imaging on a subject by subject basis, you would be surprised to see that we actually found this in the majority of the individual participants, so that was kind of cool and maybe surprising. But it's a small study and there's only so much we can learn from a small study.
>> So we moved to a larger study where we had 49 individuals, 25 of whom were chronic cocaine users with usage history of about 15 years or so. And 24 of whom were chronic alcohol users. The reason we chose alcohol users initially was because our cocaine users had really strong alcohol use history and so, in some ways, we were trying to use them as a control group. But of course, it turns out that they are very interesting in their own right.
>> So in the cocaine users, the same design -- we see here that in the sandwich design using TMS as a probe and then as a modulating tool and a probe again, that cTBS within the subject sham controlled showed specific decreases in evoked BOLD signal in the ventromedial prefrontal cortex, the cingulate, the striatum, and the insula.
>> So if you see here these are all very classic, even the subgenual cingulate, these are all very classic limbic areas that we think of when we think of psychiatric disease and substance abuse, and craving. That was very exciting.
>> These are the actual data here. So our significant clusters were here in the caudate, the accumbens, and cingulate, as well as some precuneus and posterior parietal cortex. And that has been published.
>> And in fact, we actually see a similar pattern in alcohol users, which I have here. The alcohol users also have a decrease of evoked BOLD signal to the ventral surface of the brain, medial frontal cortex here as well as in some insula and posterior inferior frontal gyrus. So again here, you see the sort ventral aspects of the brain changed by cTBS.
>> Then a very bright postdoc in my lab Tanisha Kearney Ramos, analyzed the data where we did cue reactivity immediately before and after the same paradigm.
>> So again, in one case, to the basic scientist in me, really likes the interleave TMS data because that’s the principle, that’s the fundamental mechanism we’re looking at. But of course if we’re trying to develop a clinically relevant procedure we probably want to look at cue reactivity because we know that cue reactivity leads to relapse. We could have done attentional bias, there are many tasks we could do, but task specific activation is probably very important. And what Tanisha found briefly, is that in cocaine using individuals, real TMS specifically changed the connectivity, task based connectivity to cocaine images in our cocaine users.
>> And this was also true in alcohol users.
>> And I will unpack this for you in the last few minutes of the talk. So what she did is using psychophysiological interaction, so PPI tasks, this is a way to measure functional connectivity but in this case we are looking at task modulated functional connectivity.
>> So how is it that your brain oscillates between looking neutral pictures and then basically brought online when you see cocaine pictures and off-line when you see neutral pictures again.
>> So it is task modulated connectivity. She did a region of interest analysis. And she compared the block for individuals who are looking at cocaine cues versus when they are looking at neutral cues.
>> And what she found were in the contain users, baseline functional connectivity was high when they were looking at cocaine cues before they receive TMS.
>> After they received real TMS, but not after they received sham TMS, so these are real TMS specific effects, there was a significant decrease in connectivity between the medial prefrontal cortex and the caudate, and the putamen, and the insula. These are tremendously exciting data and I'm very excited for her. Because as you may recall from the initial slides, these are the same brain regions as you know that are associated with relapse in cocaine users.
>> So these data really suggest that in cocaine users, we can apply cTBS to the medial prefrontal cortex and potentially make the brain look like the brain of somebody who won’t relapse with cocaine.
>> So that's pretty cool.
>> Okay and in our alcohol users, data is good, and then if you can generalize it to another drug using class, or a whole new population, it’s even better.
>> So even more excitingly, she found the same pattern was true in the alcohol users.
>> So alcohol users, their engagement of their brain, of the medial prefrontal to the striatal systems when they see alcohol pictures, was specifically decreased following cTBS administration, but not sham.
>> So it suggests that cTBS may be a transdiagnostically appropriate tool to change the brain response to drug cues. And what we didn’t find was a change in self-reported craving.
>> So why is this -- well we are not too worried about it because self-reported craving is not a super robust measurement, and in fact generally you need a lot of sessions in order to get a behavioral change.
>> And so the last step that I will talk about here is the newest study that we are doing which is ongoing. And it's looking at sustainable changes in brain reactivity with cocaine cues in treatment-engaged individuals.
>> So like I said, this is a single site, double blinded sham -controlled cohort study in individuals that have enrolled in an outpatient treatment program where they are largely receiving cognitive behavioral therapy for four weeks, so 28 days.
>> When they enroll, we randomize them to receive 10 days of real or sham cTBS and the sessions occur typically before their outpatient CBT sessions.
>> We apply the TMS between week 1 and week 4 of their outpatient treatment and then we follow-up one month and two months later. And we're looking at cue reactivity at these different timepoints.
>> Right now the study is still blinded, so it's hard to know what to tell you all. But it's very exciting. And what I can say, is that since it's blinded, we have the ability to kind of look at group A and group B. And it does in fact seem as though group A and group B are starting to separate -- we are about halfway through the data collection for the study.
>> So if we're lucky, they are separating in the right way. But you know, they may be separating in the wrong way. And that’s honestly the fun thing about science -- either way we will learn something.
>> And so hopefully the data will be ready to present at ACNP this December so I will have more for you then.
>> So I just kind of want to end with a little inspiration. So “the theories which I have expressed to you here, may appear to you to be so chimerical are really extremely practical – so practical that I depend upon them for my bread and cheese.” You know all the time, especially as an extramural program, we become very good at using all the buzzwords of language. Like translationally relevant, evidence-based treatment. And they sound good and we believe it. But in this case, we really might be able to make a difference. We really have a lot of really strong preclinical data that suggests that these circuits have a causal effect on drug self-administration. And we have more and more rigorous sham control studies in cocaine users with TMS that show we can move these circuits. So they’re buttressed by this huge body of knowledge that suggests these circuits are important to relapse. So I think with a community of us working together, we really could bring this to sort of an FDA-approved treatment for cocaine dependence sometime maybe in the next 3 to 5 years.
>> That's very exciting because there is currently no FDA approved treatment for cocaine dependence at all.
>> So our vision is to be able to deliver rTMS as the first translationally-relevant, neural circuit based brain stimulation treatment for cocaine dependence in the next 5 years, if we're lucky.
>> There's lots of questions out there, there's lots of opportunity for you guys and everybody out there doing viewing to make a difference in the field. Tons of questions. And just like the depression field is still working on all of these optimization protocols now, a decade after FDA approval, and many years after Medicare and Medicaid have been reimbursing patients for it, I think we too, as substance dependence researchers, have a lot of time to work out all of these details. And it's very great and it’s exciting for our future.
>> And we just need more people to participate. So come on. Join us.
>> Okay, I’m going to skip through this because I think they're out of time. But individual variability is very important. There are some suggestions that various factors contribute to it. And Ridding and Ziemann have done an excellent review article talking about this and I highly recommend it.
>> And finally this initial cascade of evolution of addiction that I talked about, the vulnerable rewards seeking individual and the habitual drug taking behavior that manifests. Those two different phenotypes or different portions of the addiction spectrum may in fact respond better to one form of TMS stimulation than the other.
>> It may be the case that individuals that have more reward-based mechanisms that drive their cocaine use are going to respond better to medial prefrontal cortex stimulation. In individuals that have more habitual related reasons for drug use potentially like we think about with cigarettes and smoking, there have been great multisite trials with TMS as a tool, with the BrainSight coil, of developing treatment for nicotine and in that case they’re simulating sort of a more motor habitual executive control area.
>> So that would be really exciting to see as well.
>> And then one last thing, as last word of inspiration. So John F. Kennedy gave this speech a while ago and we thought at a time when it was very odd to think of going to the moon. We had never really, we had left the planet maybe a little bit, but we certainly hadn’t done anything as great as putting a man on the moon. That seemed so crazy.
>> And I think to many of us it seems crazy to actually be in these rooms on a daily basis, doing the jobs we do every day, and actually think we can make a difference. But we can.
>> And so listen to his words. “We chose to go to the moon! We chose to go to the moon in this decade and to do other things. Not because they are easy, but because they are hard. Because that goal will serve to organize and to measure the best of our energies and skills. Because that challenge is one we are willing to accept. One that we are unwilling to postpone and one that we intend to win.”
>> And to that, I wanted to say I have a lot of friends who do TMS, I have a lot of friends who try and get discouraged because of the giant parameter space and nobody knows what to do. But don't be discouraged.
>> We have an opportunity to really come up with new treatments. And there's a lot of evidence.
>> And I think we can make a difference.
>> And these are all the great people that help me all the time.
>> [ Clapping ].
>> Do any of you all in the room or does anybody in Internet land have a question?
>> You can type them into your window and I will receive them. In the chat window.
>> [ Indiscernible audience question ]
>> Absolutely, that's a great question. So the question was, “Why is it for cTBS we’re picking such a high intensity?”
>> Let me go back to that slide.
>> And the answer, so we're using 110% of resting motor threshold.
>> And I should say, Downer’s group uses resting motor threshold for the foot when they stimulate the medial wall. They are actually not too far from the motor cortex, so it actually seems pretty smart. It is really hard, I'm not that good at getting motor evoked potentials from the foot.
>> And I was trained by people who use the arms and hand and we use 110% mainly because of the distance.
>> Mainly because the distance from the frontal pole to the surface of the skin is much greater.
>> So that's really the reason.
>> Another related question there is why did we use 3600 pulses initially.
>> So the reason we use that, again, we use that because we wanted to use state averse stimulation. It's biologically relevant. It's a natural endogenous brain rhythm that involves the learning and memory. And that's great.
>> We wanted to use continuous state averse stimulation.
>> But because I was changing the frequency, the burst frequency, I wanted to minimize the number of other parameters we were giving.
>> So clinically, to treat using TMS, people often use about 3000 pulses, we wanted to keep the pulse number approximately the same.
>> Somewhere between 3000 and 4000 pulses.
>> cTBS, kind of off the shelf, comes in about 600 pulse units, so to speak, so we needed to do a multiple of six so we chose 3600.
>> The question was, do we adjust the stimulation intensity based upon the distance between the scalp and the brain or do we apply 110% to everyone?
>> The answer is, right now, we apply 110% to everyone and then we use the distance between the scalp and the cortex as a covariate in our analysis. So that’s what we’ve been doing but I think it’s completely reasonable to dose the TMS based on the distance and I could see that the field might potentially want to move in that direction in terms of translating this to a clinical treatment.
>> Refining it for your individual.
>> So that question was, do you study how to maintain durability of TMS over time, i.e. maintenance protocols? So again in addiction, we are far more immature that depression so we really are at square one.
>> I do know there’s a lot of emphasis in depression literature, all around the world, actually, groups are looking at durability of treatment effects. So no, we are not because we are not there yet, we don’t actually know if we can change the disease yet. But when we get there, we will definitely think about maintenance.
>> Thank you all very much.