MPS-TMS: world’s first fully customizable wireless brain stimulator
Background
Trancranial magnetic stimulation, or TMS, works by firing a magnetic pulse over your head to wirelessly excite the neurons in your brain.
The exploration of electrical stimulation of animals can trace back to several hundred years ago. Luigi Galvani (1780s) observed that frog legs twitched when struck by an electrical spark.Giovanni Aldini (early 1800s) applied electrical currents to human corpses, including heads. Sorry, I didn’t warn you that this pathway is not full of humanitarianism, similar to every other early scientific discovery.
So far, these practices involve direction contact of stimulation, i.e. a few electrodes have to be placed on your skin or scalp to perform the stimulation. We still have such things nowadays. TElectroconvulsive therapy (ECT) is the most known one – placing probes on your brain to treat mental disorders. I have a friend who received such treatments. It works well on his drug addiction. However, it has a significant drawback – it impairs your short-term memory. For example, he would forget the fact that we just met yesterday. And this is a well-known fact among ECT receivers. The cause is that to work on your deep brain (the part that is deeper beneath your scalp), the stimulation has to pass through all the way from the contact on your scalp. First of all, it hurts, of course. More importantly, the brain sections that are not related to your symptom but is in the way will have to be stimulated. Like, when you are trying to dig out a carrot, there is no way to leave its stem intact, unless we have magic.
And this magic is TMS. It all starts with one guy named Faraday, who found that an alternating magnetic field can induce electric potential. And those crazy experiments already concluded for us that the electric potential can activate neurons! And bang, we have a chain of reaction: TMS fires a magnetic field; the field wirelessly penetrates through the scalp and skull and induces electric potential; the potential activates neurons – TMS contactlessly activates neurons.
People have done lots of studies with this new tool of stimulating neurons. The most exciting one is that TMS can treat depression, a disorder that bothers 280 millions people worldwide. The conventional treatment for major depression is pills, yet, quite a large percentage of patients are resistant to medicine. But nobody can resist magnetic field – unless you have the same helmet as the one Magneto has.
However, we don’t actually have a mathematical explanation for this process. We know that it works simply because we see that it works. I know it is a common phenomenon in biomedical works – if it works, it works. I know the best we can do with medicine might be to find a proper amount of intake, or the cocktail-like prescription. But with TMS, essentially a magnetic field, it’s hard to not imagine that through manipulating the field (intensity, shape, direction, repetitive rates), we can do better.
Unfortunately, TMS machines used to have very limited functionalities. A TMS machine can only fire a fixed-shape (monophasic with low repeating rate, biphasic) pulses, and we can’t manipulate when and how to repeat the pulses, which is even more important than pulse shapes. Our dream of having a TMS playground so we can find a better solution for the whole group of patients or each individual, is not feasible.
Until now. Thanks to modular power electronics, we built a 2.8 kV / 6 kA stimulator, which can be used to synthesize TMS pulses of flexible shape, arbitrary firing rate and timing.
- If you are an electric engineer, the closest thing you can relate might be an digital-analog-converter, but on a mega watt level.
- If you are a neuro-scientist, imagine it as your Aladdin: capable of any pulse shape (mono-/bi-phasic, half-sine) of any protocol (single / paired pulse, theta-delta, burst) at any repetitive rate you may want.
Here is a short intro video.
Implementation
Pipeline for closed loop TMS experiments.
A closer look into the stimulator hardware.
Detailed look of a single board.
A more detailed introduction of this stimulator can be found here.
The biggest challenge here is the combination of power level and the output spectrum. A typical TMS device provides an output at 3 kV and ~kA. Traditional TMS waveforms are usually resonant or decaying resonant sinusoidal waveforms, whose period is around 200 µs, i.e. 5 kHz signal, and we are trying to synthesize them! To reliably synthesize a signal, we need the output bandwidth to be at least ten times higher. For better qualities and less distortion, we need even a higher bandwidth. Our eventual configuration is 300 kHz.
Next Step
SiC MOSFETs are still MOSFETs – it’s not comparable to bipolar devices speaking of short-circuit capability, which is what actuall happening in a TMS device – just putting 3 kV on a piece of wire. Although, the short-circuit lasts only a very short time, around 100~200 us.
A better fit of power devices is IGBT. However, the problem with IGBT is that it switches very slowly. Get a high quality with limited switching, is it possible?
The gist is to make modules have different voltages to create even more output levels, i.e. a higher resolution without crazy switching. I have demonstrated this idea with a simple experiment. You can find the demonstration here.
I will develop a more competent hardware soon in the future.