Skip to content

COVID-19 is an emerging, rapidly evolving situation.

Get the latest public health information from CDC: https://www.coronavirus.gov
Get the latest research information from NIH: https://www.nih.gov/coronavirus

Zhi-De Deng: Transcranial Magnetic Stimulation: Physics, Devices, and Modeling

Watch on YouTube.

Transcript

Zhi-De Deng:  Hi. My name is Zhi-De Deng. I am a staff scientist at the National Institute of Mental Health Non-invasive Neuromodulation Unit, and today we're going to be talking about the physics, devices, and modeling of transcranial magnetic stimulation, or TMS. First, my disclosures. I am an inventor on patents and patent applications held by various institutes related to magnetic stimulation technology.

So let me paint a picture of the different electrical and magnetic stimulation techniques. Here, I have categorized many of these modalities into four categories ranging from subthreshold stimulation methods, to subconvulsive but suprathreshold stimulation methods, to convulsive therapy, to surgical implantation techniques.

The surgical techniques include vagus nerve stimulation and deep brain stimulation. These have to do with implantation of electrodes in the brain, and installation of the battery pack in the patient's chest area to provide focal and chronic stimulation in the brain and in the vagus nerve.

Less invasive, we have convulsive therapies, such as electroconvulsive therapy; and magnetic seizure therapy.

These techniques involve putting the patients under generalized anesthesia, and electrical pulses are delivered to their heads to induce a therapeutic seizure in the brain.

Less invasive still are the suprathreshold but subconvulsive techniques, and this is where we have transcranial magnetic stimulation and its variants. TMS, some form of it known as deep TMS, which I'll explain a little bit later. Finally, we come to the subthreshold techniques. These are techniques that deliver very weak electrical or magnetic stimulation to the brain. It is subthreshold, meaning it is sub-action potential threshold. They do not produce action potential to the neurons. And some of these techniques include transcranial direct current stimulation, or alternating currents stimulation, and low field magnetic stimulation, and transcranial pulse electromagnetic field stimulation.

The history of brain stimulation actually goes back to thousands of years. The ancient Egyptians were very well aware of the electrical properties of catfish.

The Greeks actually used stingrays like this to numb pain during childbirth. But of course, Greeks didn't know that water is what conducts electricity, and so they treated these creatures as mythical creatures, and it was said that for best results, you have to catch the fish in the moon of the Libra, and leave it out in the open for three days before you bring it into the room, but never in contact with the subject for numbing of the pain.

It wasn't until the 1800s, the late 1700s when people started to study bioelectricity more scientifically.

Starting with the seminal work of Galvani, an Italian physician and a physicist, who first conducted the frog leg twitch experiment. He observed that when dissecting frogs that -- whenever he placed a metal rod on the sciatic nerve of the frog, he would induce a muscle twitch in the frog leg.

And he hypothesized that there was some kind of energetic fluid that was intrinsic to the animal, which gets released when the metal probe was touching the muscle.

Galvani had a slightly younger rival, a man by the name of Alessandro Volta. Volta disagrees with this theory. He thought that it was the different composition of the metal that was generating the electricity and that the frog leg was merely a conductor facilitating the flow of the current. And in his own experiments, he sandwiched metals of copper and zinc into a pile that you see on the left and the so-called voltaic pile. This is the early form of an electric battery. And he thought that it is the different composition of the zinc and copper that generated electricity. One of the sides is the anode, the other metal would act as the cathode, and there's an exchange of ions that generated a current flow in the animal.

The debate between Galvani and Volta not only spawned the field of electrophysiology, but much of biochemistry, and also electromagnetism.

It was the nephew of Galvani Aldini, who first experimented with transcranial electrical stimulation in psychiatric patients. He was treating, at the time, a 27-year-old farmer with melancholic madness around 1801.

Of course, Aldini tried this technique on himself first, and he noted that he experienced a great shock in the inner surface of his skull, which rattled the inner fluids of his brain.

And then he became insomniac for days.

Now, that might have been his side effect to the stimulation, but his patient, who was suffering, remember from depression, was reportedly cured after several weeks of this electrotherapy.

Unfortunately, much of Aldini's work was overshadowed by some of his more unusual scientific endeavors.

Aldini was famous for doing public demonstrations of galvanized corpses. That is, he would take freshly executed prisoners, brought them into the lab, and tried to re-animate the dead body using electricity, which later inspired Mary Shelley's novel, Frankenstein. And so, you can see here, a depiction of the reaction of the ordinary folks to the experiments of Aldini.

And pretty soon, around the mid-1800s, you have devices here that are sold for home use, devices that generate electricity. And it came with an instruction electropathic guide for home use. And this device can cure everything. Fast forward to the 1930s. We saw the invention of electroconvulsive therapy, which today remains one of the most effective treatments for major depressive disorder.

And the problem with ECT, of course, ECT induces a lot of memory side effects.

And so, research in brain stimulation stalled until the invention of the modern-day transcranial magnetic stimulation. Now, TMS is not exactly a modern technique. We saw earlier versions of this technology already in 1910 by Thompson. So, illustrated on the very left. Thompson had a huge magnetic coil that was placed on both sides of his head. It wasn't very practical. It wasn't very portable.

The modern TMS machine was developed around the 1980s, generally credited to the three gentlemen you see in the middle: Jalinous, Freeston, and Tony Barker. And they made the first portable, practical version of the TMS machine.

And the original application uses a round coil, which you can see in Tony Barker's hand.

Modern-day TMS, which you can see on the right, uses a figure-eight coil for better focality. We'll get to that a little bit later. And today, TMS is approved by the FDA for treatment of major depressive disorder, and also OCD, a topic that you have learned from Dr. Lisanby earlier.

The underlying physiology of TMS.

The TMS coil that is placed near the subject's head induces a magnetic field that is of a 2-Tesla magnitude.

And this magnetic field, in turn, creates an electric field in the brain. When neurons in the brain are exposed to this electric field, the neuronal membrane potential changes, and when it reaches a certain threshold, this neuron underneath the coil will fire, causing action potential, and subsequently, behavioral effects.

The physics of TMS is very simple. It's nothing more than Faraday's Law of electromagnetic induction. A time-varying magnetic field is generating a secondary electric field in the brain. This is an illustration of an early Faraday experiment. He would use a canister to induce electricity in a spool of wire, labeled A. You can think of this canister as a TMS stimulator, and the winding, A, as the TMS coil. And the winding A is moved rapidly in and out of another spoil of wire labeled B, which then generates electricity in that loop of winding, which you can then measure using a meter. The analogy would be that this canister that generates the electricity is the stimulator, the spool of wire A is the TMS coil, and the head is B.

When you open up a typical TMS device, you can see a number of components, including a very large capacitor, which is in operation charged up to 2 kilovolts of voltage. The delivered current to the TMS current is upwards of 7,000 amps, 7 kiloamps, producing a magnetic field at the surface of the coil around 2 Tesla. However, it only generates an electric field in the brain on the order of one volt per centimeter, or 100 volts per meter, which is quite small.

The problem is that the human head and human tissue is not a spool of wire illustrated in the previous slide. The human tissue is actually not a very good conductor for the magnetic field. Hence, you need a lot of energy, 7-kilo amps of current, and 2-kilo volts of capacitor voltage to even generate a tiny amount of current flow in human tissue. And when you actually calculate the power by multiplying voltage and current, you obtain a value that is enough to drive a dreadnought class battleship around the turn of the 20th Century.

Because TMS uses very high power, some of the early designs are relatively simple in terms of circuit typology.

The first generation conventional TMS stimulator is basically an RLC circuit. Maybe something that you have learned in high school physics. A capacitor is charged through a charger, and then all of that charge is dumped onto the inductor, which is labeled L.

And the inductor is the TMS coil.

And during the second phase, all of the current is dissipated in the middle through the resistor, R, creating an overdamped sinusoidal wave form as illustrated on the top left -- top right. Excuse me.

And this is called a monophasic pulse.

Most of the energy is in the first phase of the temporal waveform. So, it is a monophasic pulse.

You can remove that resistor in the middle and have on the second phase most of the current returning to the capacitor. This generates a biphasic pulse. Basically, you get one complete cycle or multiple cycles of a cosine wave. And this second stimulator design is more efficient for repetitive stimulation because you don't have the internal resistor to dissipate heat. If you were to run repetitive protocols with high frequency using the monophasic pulse, the internal resistor, R, would heat up fairly quickly. And using the second design, one is allowed to recycle some of that energy back to the capacitor, getting ready for a subsequent pulse.

So, most of the modern-day repetitive TMS stimulators are designed using the second design.

Now, it turns out that the brain and neurons are not as responsive to sinusoidal waveforms.

Neurons in the brain much prefer square waves. This has something to do with the membrane of the neuron acting as a capacitor that is connected in series with extracellular and intracellular resistance, serving as more or less an RC circuit.

Given that the neurons prefer square waves, one can design a stimulator that can produce a near rectangular or near square pulses. And this is the work by Angel Peterchev at Duke University. This is the so-called controllable pulse width TMS, using insulated gate bipolar transistors to sharply switch on and off current so that you can produce very sharp edges in the electric field waveform. And, you can control the timing of the switching, which means that you can control the width of these pulses. And also, the polarity of the pulses can be somewhat controlled if you have different banks of capacitors in the stimulator. This is a commercial product now, sold by Rogue Resolutions, or Brainbox.

Another up and coming technology is the so-called synthesizer. It's a similar concept to the controllable pulse width TMS device but using a modular design to achieve very fine control over the voltage and electric field waveform. And one can produce an arbitrary electric field waveform using this kind of new design.

And this is also being implemented at Duke University by Stefan Goetz.

So, those are some of the devices that can generate the temporal waveform of TMS. Next, we are going to talk about the spatial aspects of TMS, spatial targeting, which has to do with the geometry of the head, the geometry of the TMS coil that is placed over the head. There are some constraints, though, before I go into the coil geometry. The fundamental physical constraint of TMS is that you cannot achieve a three-dimensional focus in-depth with TMS. So, the electric field that is induced by TMS is always maximum at the surface of the head and the surface of the cortex, and then it decays exponentially as you go away from the coil, deeper and deeper into the brain.

Conventional TMS coils, like the figure-eight coil that is illustrated here, has an electric field penetration about two to three centimeters from the head surface, hitting about the interface between the grey matter and the white matter, where the brain is being activated.

In fact, there are many other TMS coil designs. And several years ago, I went out to collect some of the coil designs that are produced by different manufacturers, and also theoretical proposals for TMS coils. And I have here a catalog, more than 60 of these TMS coils, and I have categorized them into loosely three categories, which I will tell you about in a second.

And you can see here that TMS coils come in various sizes, from the very small circular coil to a very large helmet type coil. And also, the same applies to figure eight type designs. They come in various sizes and shapes as well. So, the three categories are; number one, circular type coils. And these coils consist of a circular loop of winding, or stacks of circular loops, which produces a circular field pattern underneath the perimeter of the coil.

Included in the circular type coils are the so-called deep TMS coils, or the H coils, manufactured by Brainsway. Number 14, for example, coil number 14 is the H1 coil, which has been FDA approved for the treatment of depression. And you can see that many of these H coils also produce a circular type field pattern in the brain as illustrated by computer models on the right-hand-side.

The second category of coils is the figure-eight coils. These consist of two loops of wire next to each other. And the electric field is maximum at where the two loops join in the middle underneath the center of the coil. And so, you can see here that many of the commercially available treatment coils are figure-eight types, including ones from Magstim, from MagVenture, and Neuronetics, are figure-eight type coils.

Figure-eight type coils also come in various sizes and shapes. Some of the coils that were used to stimulate deeper in the brain include magnetic seizure therapy coils, illustrated in coil number 50, and also the double cone coil used to stimulate the leg area, which is slightly deeper than the hand motor representation on the cortex. And that is illustrated in coil number 47. And so, you can see here, compared to the circular type coils, the electric field produced by figure-eight type coils are generally more focal.

And then, the third category of coils are less easily categorized. They include multi-loop coils, MRI-like coils. And you can see here that even though some of these coil geometries look very complicated, they produce electric field shape that are similar to either a circular coil or a figure-eight type coil.

And when you place all of these coils in one diagram, some very interesting trends start to emerge. So, this plot is the so-called depth-focality plot.

On the X-axis it plots the depth of stimulation. That is the depth at which the electric field decays to half of its maximum. Recall that the maximum of the electric field always occurs at the surface of the brain, and then it decays exponentially in depth. And the characterization of the depth of stimulation here is the distance at which the electric field falls to half of the maximum value at the surface.

The Y-axis plots the spread or tangential spread of the coil. So, the higher the spread, the more non-focal is the stimulation. In other words, more brain is being stimulated.

And so, you can see here that there are, out of all of these 60 or so coils, they lie nicely on two lines, the circular type coil on top, and the figure-eight type coil on the bottom; which means that the figure-eight type coils are intrinsically more focal compared to the circular type coils. That means that for a given depth of stimulation, you always stimulate less brain.

The spread is always smaller with the figure-eight compared to a circular type coil.

You can also see that no matter what the coil type is, both of these lines are trending upward, meaning that the deeper you try to stimulate in the brain, it is also necessary that you stimulate more brain. So, that is the so-called focality-depth trade-off. And at the deepest point in the -- of the depth curve, both coil types converge to this coil, zero, which is -- if we go back one slide -- is what we called in physics a flux ball. This is a configuration in which you wrapped a whole head with hire. So, don't try this at home. But that represents the theoretical limit to the depth of stimulation of TMS.

And you can also see on this plot that there -- the conventional figure-eight TMS coils, such as the ones that are FDA approved for the treatment of depression, the ones from Neuronetics, from MagVenture, and Magstim lie in the blue region. They all have quite similar depth of stimulation and focality profiles.

In fact, the FDA approval of the Magstim and MagVenture coils is based on substantial equivalency of the technology to the already approved Neuronetics coil. And that was approved back in 2008.

Talking about targeting a little bit. So, suppose you now selected a coil. Suppose you have a figure-eight coil, and you want to use it for treatment of depression.

How do you place this coil over the head? There are several strategies for coil placement. The very rudimentary one is the so-called 5-centimeter rule. The first step is to determine the hot spot on the head that would induce a thumb twitch. This is the so-called motor hotspot. And you move, then, forward anteriorly along a parasagittal line from your motor hot spot, five centimeters, to arrive at the left dorsolateral prefrontal cortex. And that's where you place the coil for depression treatment.

You can imagine that this 5-centimeter rule is not very accurate, because people have different size heads.

And so, a more sophisticated strategy is the use of an EEG system for coil placement, which at least accounts for the size of the patient's head, and it is scalable across patients. And this has to do with some mathematical calculation, or placement of an EEG, 10-20 EEG grid, on the subject's head, and locating where the F3 EEG electrode site is. And that marks where you might call left dorsal lateral prefrontal cortex.

And even more sophisticated, nowadays, if you can afford to get an MRI or a functional MRI, you can define your dorsal lateral prefrontal cortex that way. And there are even more sophisticated targeting techniques nowadays, using either diffusion tensor imaging, or functional MRI, as a way to localize the stimulation hotspot.

Does it make a difference?

What the difference between these coil localization methods? And to do this, we resort to computational electric field modeling. The typical workflow has to do with taking an individual MRI, segmenting the MRI into different tissue compartments, and building a three-dimensional representation, or a mesh model from that individual subject, on which you can draw in electrodes or TMS coils.

And you would then solve for the electric field distribution using a number of numerical methods. For example, finite element or boundary element methods. This way, you end up with an electric field distribution in a three-dimensional representation of that individual subject's brain. With this technique, then you can perform a region of interest analysis and quantify the maximum electric field at various parts of the brain and compare different strategies for dosing.

We have done so in some adolescent depressed patients. This is a data set obtained from the Mayo Clinic in the lab of Paul Croarkin. In ten adolescent subjects receiving an open-label course of repetitive TMS for the treatment of depression, they actually marked the spots on the subject's head. You can see on the top row, the three-dimensional representation of the subjects, the ten subjects that were enrolled in this study.

The yellow markers mark the spot of the motor hotspot. This is the spot on the surface of the head where you can get a thumb twitch. The blue dots represent the location you obtain from the five-centimeter rule. The red dots represent the EEG F3 method. And then finally, the green dot is an expert-determined structure MRI localization of the dorsal lateral prefrontal cortex.

And so, you can see here, first of all, there is a lot of variability across subjects in terms of where these markers ended up. And also, if you were to stimulate the electric field using these different coil localization methods, you can see that the induced electric field also varies quite a bit across individuals.

One can qualify the electric field strength at various parts of what you might call the dorsal lateral prefrontal cortex, the middle frontal gyrus, the middle frontal sulcus, and also the inferior frontal sulcus. Here, showing the median electric field strengths on the Y-axis. And in comparing the five-centimeter rule, the F3 localization method, and also the MRI targeting method.

One can see here that the five-centimeter rule consistently underdosed the electric field in all three parts of the brain compared to the MRI and the F3 localization method, perhaps suggesting that the clinical standard, which is the five-centimeter rule, is sub-optimal in terms of providing adequate dosing to the DLPFC target.

We have also examined interindividual variability in another study, and this is the data set collected at Cornell from Marc Dubin in 26 depressed patients receiving a standard 10-hertz rTMS protocol at the F3 target. In all of these subjects, the coil was placed on the head using the EEG localization method. And we also simulated the electric field. And when you plot the electric field magnitude at the middle frontal gyrus as a function of the stimulator output, one can see a scatter plot that is illustrated on the top right.

If everybody's heads were the same, this line would be perfectly linear, meaning the electric field is linearly proportional to the stimulator output. The fact that you're seeing deviations of these dots off of a diagonal line implies there is a lot of variability across individuals in the induced electric field in the middle frontal gyrus. One can use these computational techniques to study the dosing of TMS.

Next, I would like to discuss other forms of stimulation. One of these techniques is called low field magnetic stimulation. It uses an MRI-like head coil. And this technique was discovered at the [correction: McLean Hospital] where Michael Rohan, the inventor shown here, he scanned a number of bipolar patients using a novel echo-planar imaging sequence. And these subjects came out of the scanner reportedly feeling better. And Dr. Rohan thought that there must be something in the stimulation protocol, or the imaging protocol, that is causing a low-intensity magnetic stimulation in these subjects, producing a clinic effect.

He started to build this technology and using the MRI coil for low field magnetic stimulation. In a trial that was conducted at Cornell Medical Center, the clinical results of using this low field magnetic stimulation in the intent to treat a sample of 65 patients. There is a significant difference in the visual analog scale rating of mood after only three sessions of active stimulation compared to sham stimulation. The same pattern is observed in the per-protocol patients.

So, one can simulate again, using the technique that I showed earlier, the electric field distribution induced by low field magnetic stimulation. And one can see what the electric field distribution is like in a typical cortex. You can see that the maximum electric field is over the midline prefrontal region.

The electric field is quite diffused and non-focal. And the induced median electric field is on the order of 0.1 volt per meter, which is orders of magnitude less compared to the conventional TMS, which is on the order of 100 volts per meter.

In this study, we also sought to design a more portable version of this LFMS coil. You can see the novel design of what's called the cap coil, which resembles a helmet that produces a similar electric field in the brain while being more energy-efficient because the coil elements are closer to the head. In other words, you're not wasting energy stimulating the air above the head.

Another form of magnetic stimulation actually uses rolling permanent magnets. And this is the so-called synchronized transcranial magnetic stimulation. The theory, according to Andy Leuchter, is that the synchronized stimulation, if you were to set the frequency of the rotation of these magnets to individualized alpha frequencies, then the magnets can entrain this alpha oscillation if the magnets are placed on the frontal pole, the frontal region, and also the parietal brain regions. And by resetting these cortical oscillators, Dr. Leuchter hypothesized that you can treat depression.

This is an electric field model of this sTMS technology. I will not go into the details of how this works, but I will show you the electric field distribution on a spherical representation of the head as one of these magnets is rotating over the head. You can see here that the electric field, which is coded by the color heat map, switches between what looks like a circular field pattern to something like a figure-eight coil pattern. It's very interesting.

This is coming up is the figure eight right now, shifting towards a circular field pattern now. And when you put three magnets on top of the head, one against the frontal pole, one against the frontal cortex, and the other one over the mid-line parietal cortex, you can see it generates a moving electric field like so.

The magnitude of the induced electric field is very small. In the brain, the generated electric field is on the order of .02 volts per meter, which is about 10 times smaller compared to transcranial direct current stimulation.

One can also do TMS in a rodent model. This is a collaboration with Rachel Sherrard from France. We looked at magnetic stimulation, low-intensity magnetic stimulation, in a rodent model, and also in an in vitro brain set up. In this model, they did a very interesting experiment in which they surgically removed the innervation between the inferior olive and the cerebellum in mice.

You can see here that on the left side, showing the left and right cerebellum. And on the left hemisphere, because the innervations have been cut off, you don't see these white stripes, which are the climbing fibers of the axons connecting the olive and the cerebellum. And this is intact in the right hemisphere because the surgery was only performed unilaterally.

But after two weeks of a high frequency but low-intensity stimulation, we see structural growth.

We see reinnervation on the left cerebellum. The white stripes start to grow back after several days of this high-frequency stimulation, meaning that the stimulation, albeit very low intensity can induce structural changes in the brain.

And in also quantifying the number of cfos-positive cells, we see an almost double the amount of cfos-positive cells using a high-frequency stimulation protocol compared to sham. There is cellular activation in addition to observed structural changes.

In this paper, there was also a very interesting experiment in which they knocked out the Cry1-/-Cry2-/- genes in mice, and that completely abolishes their response to magnetic stimulation. But the axon regeneration pathway remains intact because when they treated the sample with BDNF, the axons do grow back. We think that Cry1-/-Cry2-/- genes are specifically responsible for the response to magnetic stimulation. And Cry1-/-Cry2-/- genes in some species code for magnetoreceptors, which is very interesting.

Finally, I would like to talk about the combination of TMS with imaging modalities. Starting with TMS in PET or SPECT. It's actually quite easy to combine TMS and PET because the magnetic field in the TMS does not interfere with the photon emission in the PET scan. And there's no constraint as to the timing of the TMS pulses that can interact with any kind of electromagnetic field in the PET scan. So, it is not constrained in terms of the temporal dynamics of these pulses that you deliver, making TMS-PET a very convenient technique.

The only thing you have to be careful of in terms of safety and equipment is that you might have to shield the photo multipliers so that it doesn't get damaged by the high magnetic field of the TMS. One can do TMS-PET easily.

A second technique that people are doing more and more these days is the combination of TMS in EEG. And there are several safety features related to that. One is induced heating in the underlying EEG electrodes. Originally, the EEG electrodes are ring-shaped, and when you put a TMS coil on top of it, it will couple to the magnetic field of the coil, and it will result in Ohmic heating.

And so, the EEG electrodes in a TMS compatible system is slotted, which means that they do not form a closed circle, but more of a complete C shape. This way, you reduce the eddy current generation in the electrodes.

And also, the amplifiers need to be protected as well from the TMS artifact.

The most difficult thing to handle in combining TMS EEG is the large artifact that is created by the magnetic stimulation. And so there have been several hardware modifications to the EEG amplifier. For example, allowing for high dynamic range so that you can record a signal. And also, sometimes, they use a sample in hold circuit to blank out the time period during the delivery of the TMS pulse so that the amplifier does not saturate. Those are some of the hardware implementations to try to get around this large TMS artifact problem.

And there are a number of software solutions for subsequently cleaning up the TMS-EEG signal using independent component analysis, and also space projection methods. Nowadays, one can perform TMS-EEG quite easily.

One can also combine TMS with functional near-infrared spectroscopy because, again, there is no interaction between the magnetic field emitted by TMS and the optical sensors that are placed on the head. The only thing to be careful of is when you place these optodes on the head, it creates a natural spacer between the TMS coil and the surface of the scalp, which then the induced electric field by the TMS would not be as strong as if the TMS coil was placed directly next to the scalp. One would have to account for the distance introduced by the optodes.

By far, the most difficult is to perform TMS in an fMRI scanner. And there are a number of technical challenges in combining TMS-fMRI mainly due to the large magnitude magnetic field of the scanner, and the large magnitude magnetic field produced by TMS. It can generate a torque in the TMS coil, and so the TMS coil design needs to be very sturdy. And also, you can only use planar figure-eight coils to cancel out the torque. If one were to use a circular coil, there's too much torque, and the coil will try to rotate and perhaps rip apart in the scanner.

The stimulator needs to be far away from the scanner magnet. And sometimes, the stimulator is placed outside of the scanner room, which then places a greater distance between the stimulator and the TMS coil, which means that you would have to then use a much longer cable connecting the TMS coil to the stimulator. And that creates a lot of losses, and along the way, there can be RF artifacts introduced by the scanner. Everything needs to be shielded, and perhaps thicker cables need to be used.

And also, there are constraints on the pulse waveform of TMS and the timing of these TMS pulses in relation to the gradient pulses produced by the MRI scanner. Those are some of the technical challenges in combining TMS-fMRI.

Another new application is the combination of TMS and single-unit recording. And this was successfully done at Duke University. We designed a coil to be compatible with a non-human primate that has an implantable chamber on the skull of the primate, and where the recording needles would go in. One can see here that using an active stimulation produces more action potentials compared to sham. One can record single-neuron activity while performing TMS. This is a very exciting project, and it is now a Brain Initiative funded project conducted at Duke University.

To summarize, I would like to go over our multiple brain stimulation techniques again and place them along the electric field gradient.

Starting with the very strong techniques, techniques that induce a very strong electric field in the head.

These are electrical stimulation techniques, such as electroconvulsive therapy. Followed by TMS, which also produces an electric field in the brain, which is sufficient to cause an action potential in the brain. Those electric field strengths are on the order of 100 volts per meter.

One order of magnitude down, we have magnetic stimulation in rodents. Because rodents and animals have smaller brains, the induced electric field is generally weaker. Following that, we have techniques in the sub-threshold domain, and these are transcranial direct current stimulation, and also low field magnetic stimulation.

And at the very low end of the spectrum, we have techniques such as the synchronized transcranial magnetic stimulation, using those rotating magnets that I showed.

This is the end of my lecture. I put up my contact information over here so that if you have any questions related to the content of today's lecture, feel free to contact me by email or call me by phone.

Thank you very much.