Noninvasive Deep Brain Stimulation – Has Neuroscience’s Holy Grail Been Found?

By Neuroskeptic | June 13, 2017 1:25 pm

A high-profile paper in Cell reports on a new brain stimulation method that’s got many neuroscientists excited. The new technique, called temporal interference (TI) stimulation, is said to be able to reach structures deep inside the brain, using nothing more than scalp electrodes.

Currently, the only way to stimulate deep brain structures is by implanting electrodes (wires) into the brain – which is an expensive and potentially dangerous surgical procedure. TI promises to make deep brain stimulation an everyday, non-invasive tool. But will it really work?

The paper comes from Nir Grossman et al. from the lab of Edward S. Boyden at MIT. Their technique is based around applying two electrical fields to the subject’s head. Each field is applied using two scalp electrodes.

It is the interaction between the two fields that creates brain stimulation. Both fields oscillate at slightly different frequencies, for instance 2 kHz and 2.01 kHz. Where these fields overlap, a pattern of interference is created which oscillates with an ‘envelope’ at a much lower frequency, say 10 Hz. The frequency of the two fields is too high to have any effect on neural activity, but the interference pattern does have an effect.

Crucially, while the electric fields are strongest close to the electrodes, the interference pattern is most intense at a remote point – which could be deep in the brain. Here’s an overview:


Grossman et al. present a lot of evidence validating the TI concept and showing that it does allow selective, deep-brain stimulation – in mouse brains. The most striking data comes from an experiment in which Grossman et al. used TI to stimulate the hippocampus of mice without stimulating the cerebral cortex. This is remarkable because the hippocampus is deep in the brain, and far from the electrodes:

grossman_cfosWhat’s more, TI stimulation is “steerable”. By varying the strength of the two electrical fields, Grossman et al. were able to move the stimulation zone. This holds out the exciting possibility that neuroscientists could easily stimulate different brain regions, without having to implant an electrode in each one.


There’s no doubt that this is one of the most exciting neuroscience papers to come along in a long time. It’s a cliché, but TI really could revolutionize neuroscience, as well as having clinical applications in the treatment of disorders such as Parkinson’s disease and more.

But will it work in humans? Grossman et al. imply that it will:

We anticipate that it might rapidly be deployable into human clinical trials, as well as studies of the human brain.

Grossman et al. seem so confident about the human applications of TI, they used a human brain in their ‘graphical abstract’ (reproduced above) even though only mouse brains appear in the paper.

There’s one obvious snag though: scale. The human brain is much bigger than the mouse brain. When Grossman et al. achieved “deep” stimulation of the mouse hippocampus, the hippocampus was about 3 millimeters away from the electrodes. In humans, the depth would be about ten times as high.

Presumably, it would be possible to compensate for the increased distances by using stronger electrical fields in humans, but this might create safety issues. Boyden and his group are reportedly working on human studies at the moment.

Another problem with TI could be that the stimulation zone won’t have “clean edges”: brain areas close to the stimulation target may get some degree of stimulation too. This would be undesirable, although it’s not necessarily a fatal problem, and optimization of the electrode placement could help to sharpen the stimulation zone.

Finally: how new is the idea of TI? Grossman et al. don’t cite any previous literature on the method. This lead me to assume, when I read the paper, that the idea of deep stimulation via two interfering electrical fields was a new one.

However, I learned on Twitter that it’s not a new concept. “Interferential Stimulation” (IS) was reportedly first proposed by Soviet scientists as early as 1965, and has since become an established tool in electrotherapy, i.e. the use of electrical stimulation on the nerves to treat pain. Here’s a 1996 patent (one of many) for an Interferential stimulator for applying localized electrotherapy stimulation.

As far as I can see, Grossman et al. are the first people to use interferential stimulation in the brain, but they didn’t originate the technique itself.

CATEGORIZED UNDER: animals, dbs, EEG, methods, papers, select, Top Posts
  • Marty

    I’m not sure how their technique compares (having only glanced over it), but this sounds very similar to “deep” transcranial magnetic stimulation ( which uses a similar triangulating principle, and has been shown to at least be able to target at a depth of 6cm. However, the design of the head coil in this case is specific to the target site, something that I think is similar to what Grossman et al. propose.

    Still, the more options there are the better, and from my observation this could be implemented in humans far easier than H-coils, something that likely contributes to the fact that we don’t see them much out there beyond proof-of-principle.

    • Neuroskeptic

      Interesting, thanks! I think the TI idea is quite different from the deep TMS. Deep TMS seems to be based on summation of several magnetic fields, which can produce deep stimulation, but primarily produces shallow stimulation:

      “The electrical fields induced by various types of coils are always greater in cortical regions (closer to the coil placement) but that the decrease in electrical field within the brain is markedly slower for the H-coil [Deep TMS coil]”

      Whereas TI is based on interference, not summation, and it is supposed to produce deep stimulation without any shallow stimulation.

      • Kate K

        We are going to know many more orders of magnitude about this in a decade or two. Advances in processing of data from much higher resolution brain scans, comparing large numbers of subjects will push this to levels of understanding well beyond what current researchers are even hypothesizing.

        Consider pathology. in a few years AI and processing of cancer cells will make human staging by 2017 expert pathologists look like stone age superstition.
        Imagine if you will comparing every human with that cancer in your closest genetic groups for those cells, at every stage, with every therapeutic method and processed by AI. Even radiological chemical and surgical methods of dealing with cancer will seem medieval soon.

        With brain science the advances right now in blunt stimulation are showing efficacy. Properly mapped and directed they will be extremely efficacious.

        The problems, especially ethical ones are not well discussed though. The “zero sum brain” theory may or may not be valid, but stimulation already shows some effects similar to chemical addictions. Then there are questions as to what to do if we can moderate behavior of dangerous criminals, or potentially dangerous persons. Or the fact that a lot of tech, same was with directed evolution tech is going to give competitive advantages that make attempts to control it moot.

        • OWilson

          “Then there are questions as to what to do if we can moderate behavior of dangerous criminals, or potentially dangerous persons”

          That, I’m afraid is a taboo question, impossible to answer in a Politically Correct way! :)

      • TheArgonaut

        As far as I understand it, yes, that’s right. Deep TMS is itself similar to ‘gamma knife’ therapy in cancer, and you could, at a stretch, also claim that HD-tES is similar (using >2 electrodes to specifically target an area). The interference stuff is certainly interesting, and will hopefully be replicated outside of this group soon!

    • Neuroskeptic

      Another Twitter find

      In 2006, someone patented a system for using two magnetic fields at different frequencies to produce low-frequency electric “beats” at a deep site. In other words, very similar to Grossman et al.’s TI except with magnetic fields:

      “Enables a target region of interest to be magnetically stimulated, without necessarily stimulating adjacent regions or regions that lie between the surface and the target region. Some embodiments of the invention utilize at least two time-varying magnetic fields that create intersecting electric fields in the target region”

    • pip010

      “far easier than H-coils, something that likely contributes to the fact
      that we don’t see them much out there beyond proof-of-principle.” because they DONT work ! Even the inventor of TMS publicly shamed it in front of a big audience in this year BrainStim conf in Barcelona.

      To understand why the idea is dead fish in the watter you need to run only 1 FEM simulation :) Et = -Ep – Es that is the total elxtic field is the sum of the incident/primary (from coil) and induced/secondary (due to complex domain, that is you head). When talking about induction, the primary is dominant near the coil while the secondary anywhere further than 2cm from coil surface. PERIOD.

      • Marty

        I’ve been skeptical about H-coils for a while :). You’ve probably seen that they ARE marketing a type of “deep” TMS out there now, via Brainsway, that works with Magstim units. I mentioned it to a distributor a while back, asking if it really worked. They laughed and shook their head, saying it was just a “really powerful pulse”.

        • pip010

          heh, Brainsway have FDA Clarence but no CE medical class label, for which you DO need evidence of your claim !!!

  • Uncle Al

    Move over, waterboarding, enhanced interrogation just sprouted wings.

    • OWilson

      As does it’s doppleganger, beautiful orgasms, without leaving your room.

      That “full’ feeling when you are overweight, at the touch of a button!

      Radio Shack will have a selection of virtual brain enhancing helmets. The Einstein model will be a big seller! :)

      (apologies to Monty P)

  • Alex

    This is a very intriguing study, and a clever utilization of physical principles in controlling brain stim. But I think a few issues will hinder its ability to be a meaningful therapy for people.

    – Scale/size (as aforementioned): But if one kept his or her focus on targeting the outer layers of the human cortex, great spatial control could be achieved with this technique.
    – Strength: Doing some back of the envelope calculations, the field strength applied in this paper is roughly 100 times that currently tolerated by people (in A/m^2). The strengths they apply to get motor movement are even stronger. (ouch!)
    – Specificity: They assert that the high frequency activity is not neuroactive in the non-interfering areas, but others have shown that high frequency stim can lead to changes in brain function (Chiaeb, Restorative Neurology and Neuroscience, 2011). Given the field strengths they use, I can’t imagine the field isn’t influencing brain activity substantially under each electrode..

    • Neuroskeptic

      Thanks for the excellent comment. When you say:

      “the field strength applied in this paper is roughly 100 times that currently tolerated by people”

      Are you taking into account that the field has a 2 kHz frequency? Doesn’t the tolerability depend on the frequency?

      • Alex

        That’s a good point – I do use the tDCS reference as my standard for safety/tolerability, as that’s what’s most well studied (e.g. Bikson, Brain Stim, 2016).

        That being said, there was a poster at the Soc for Neuro meeting last year from Mihaly Voroslakos that was investigating a somewhat similar multi-electrode, steerable, pulsed tCS approach. I think he used even higher frequencies (upwards of 10 kHz) and with current strengths of 9 mA (though I don’t remember the electrode size). I asked him what the experience of this stim was like, and he said it wasn’t very pleasant – some vertigo, metallic tastes in the mouth, etc.

        It’ll be intriguing to see what people develop to get around this issue. I’m not sure if it’s inherent to the electric modality or not..

  • Pingback: 1 – Noninvasive Deep Brain Stimulation – Has Neuroscience's Holy Grail Been Found?()

  • Don Crawford

    My question is: What exactly is being stimulated? And, what is the overall result?
    My concern is that modern science is interfering with Nature in ways they do not understand the consequences. The brain is not the mind, and it is the mind that is the Thinker. The brain is merely a receptacle which holds the memory of things past. The mind is translating the events in one’s current level of consciousness, which it then stores in the brain.

    • pip010

      I see, so the mind is some sort of cloud computing :)

  • David Kra

    It seems to me that what is being created here is called a Phased Array, especially as the term is used in Radar, Sonar, and Radio Astronomy. Multiple low power wide-dispersion sources (antennas) are fed with pulses timed (or phases adjusted) so that the signals all converge (are in-phase) at a particular angle and distance. The same technique can be used with the receiver to sum the signals from only one angle and distance. See

    This is different from CT & MRI’s scans and gamma knife treatments. With these, the transmitters do not all have to be active at the same time. Often one source is physically moved around the target over the course of the procedure.

    • Neuroskeptic

      Yes – in the TI technique described in this post, it is necessary to have two sources active at the same time.

      The difference I think is that in TI, the signals are at different frequencies and thus never come “in phase”, rather they create a low-frequency “beat” instead.

      But the principle is otherwise similar to a phased array.

    • pip010

      The main problem is, hardly addressed in the paper, that you brain is not an empty space (air). Thus all bets are off IMHO.

  • pip010

    rat brain is not deep!

  • pip010

    hah, another one, tACS is a modulator not stimulator. So it makes 0 sense to compare tACS to DBD. Who reviews those papers … !?

    Ahh,but it is not what you know in science it is who you know in the publishing business 😀

  • Pingback: Is It Possible To Stimulate Deep Inside Our Brains Without Implanting Hardware? | Gizmodo Australia()



No brain. No gain.

About Neuroskeptic

Neuroskeptic is a British neuroscientist who takes a skeptical look at his own field, and beyond. His blog offers a look at the latest developments in neuroscience, psychiatry and psychology through a critical lens.


See More

@Neuro_Skeptic on Twitter


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar