Translating TMS from Autism Science to Autism Therapy

The following is a blog that I wrote at the request of John Elder Robison. John wanted to know how I became interested in Transcranial Magnetic Simulation (TMS) and what drove me to use this technique in autistic individuals. My initial answer was somewhat short and John asked poignant questions which made me expand my original explanations.  Hopefully it has met with his satisfaction.


I first became interested in the idea of using TMS to relieve some of the disabling aspects of autism years ago while unpacking some of my old electronic equipment stored in my basement. I had packed the equipment when I went into medical school thinking that I would be busy but that someday I would share this hobby with my daughters. Scroll forwards a couple of decades later and it became painfully clear that my daughters had no interest in electronics and that my old equipment would be having an inevitable appointment with the trash bin. As I was partaking in a nostalgia trip opening boxes I stumbled across an assortment of transformers, inductors and small motors that I had scavenged from junk equipment. While spreading all of my previous treasures on the basement floor an idea came to my mind. Based on evidenced derived from autopsy studies performed in my laboratory, the use of coils and the induction of electricity could provide a possible way for treating individuals with autism spectrum disorder (ASD).  In this blog I will try to summarize some of the results of our autopsy studies and how the same gave rise to the first clinical trial using TMS in autism.

The outer rim of the brain or cerebral cortex contains conglomerates of cells called minicolumns that allow for information processing.   Minicolumns are ecosystems of specialized cells (i.e., neurons) and their projections that repeat themselves several hundred million times throughout the cerebral cortex. Think of minicolumns as the microprocessors of a computer. The enormous number of minicolumns accounts for the advanced parallel processing capabilities of human brains and its advantage over computers.

Within minicolumns cells are compartmentalized so that excitatory neurons are located at its core while inhibitory cells are preferentially contained at its periphery. Some of the inhibitory cells framing the minicolumn stand in perfect geometrical orientation, that is, at ninety degrees to the surface of the cortex (see figure 1). Researchers have described the latter arrangement of cells as a “shower curtain” of inhibition enveloping the excitatory core of the minicolumn. Shower curtains keep water inside of the bathtub; without a shower curtain water would splash over the bathroom floor. In similar fashion, a flaw in the inhibitory surround of minicolumns would allow for signals to escape into adjacent minicolumns thus procreating a cascade of excitation.

radial organization

Figure 1: In the 1890’s Cajal described the presence of neurons in the cerebral cortex whose orientation was at ninety degrees to the surface of the cortex. Because of the salience of their orientation in the past these neurons were often called radial cells. These figures illustrate different areas of the cerebral cortex stained by immunocytochemistry to reveal the presence of these radially arranged cells. Recent work by the Spanish Neuroanatomist DeFelipe has shown that these radial cells frame the periphery of minicolumns. Note: The surface of the brain is towards the top of the figure. The small dots are the cell bodies of the neurons and the “wires” coming from them are the axons.

Our basic research in ASD suggested that the shower curtain of inhibition surrounding their minicolumns was thinned out throughout the different cortical layers. This faulty inhibitory surround has been proven correct by other researchers using EEG and is now the basis for a new diagnostic technique introduced by a group of investigators at the University of North Carolina. The faulty inhibitory surround of minicolumns has allowed us to explain many of the symptoms observed in ASD including both sensory phenomena and seizures ( For many affected individuals these symptoms are alarmingly disabling. This led us to consider whether the inhibitory defect observed in minicolumns of ASD individuals would be one amenable to treatment. Was there a way of increasing the inhibitory tone of minicolumns? If so, this would open the door to innovative therapies targeting symptoms (e.g., sensory overload) in ways never imagined before.

Previous research had shown that giving inhibitory agents or anticonvulsants to treat seizures in autistic individuals could, in some cases, reduce some of their disabilities. The problem is that these agents had major side effects. An overall increase in inhibition could cause stupor or coma. Furthermore, these drugs are non-specific for the type of cells that they acted upon. I wanted something specific that would act on the cells that framed the minicolumns. I wanted an intervention that would preferentially target those projections that stood at ninety degrees to the cerebral cortex and formed the shower curtain of inhibition to minicolumns.

Going back to high school physics, there is a principle of science (Law of Faraday) as to how electricity is induced by magnetism. Basically this principle states that a changing magnetic field will preferentially induce current through a conductor standing at ninety degrees to the same. Transcranial magnetic stimulation (TMS) makes use of this principle to when it applies an extremely strong magnetic field to the cerebral cortex. The coil of the TMS machine focuses the magnetic field while projections of neurons called axons take the place of wires or conductors. At the time, I thought that TMS could be used to rebuild the shower curtain of inhibition to the minicolumns (see figure 2). Several years later other groups developed a similar idea for using TMS in ASD but based on other possible mechanisms, including that of the broken mirror hypothesis ( Personally I do not believe that TMS can be used to target mirror neurons within the cerebral cortex. The magnetic field cast by TMS would cover all cells within a given area of which the mirror cells would only be a minority. Any effect of TMS would be due to the majority of cells being stimulated, some of which have antimirror effects.


Figure: (A) Poor spatial contrast across minicolumns as postulated in the brain’s of autistic individuals. Minicolumn 2 receives input but is not capable of adequate inhibition of surrounding columns for optimal contrast. (B) TMS activation of intracortical inhibitory neurons within minicolumn 2 leads to enhanced surround inhibition and better spatial contrast. This leads to a greater discriminatory capacity for the involved cortical unit.

The law of Faraday has been used in order to build machines that generate current. These machines convert the mechanical energy generated by moving a conductor within a magnetic field into a current. In a generator the faster you move the conductor the larger the output of electricity. A simple experiment in high school many years ago showed me that the speed by which I cranked a generator altered the brightness of a light bulb to which it was connected. The faster I went, the brighter the bulb became.

In biological systems the relationship of how fast a magnetic field moves in regards to a conductor (i.e., axons of neurons) may not necessarily apply. You can increase the frequency of rotation within a generator to produce more electricity. In the brain, however, lower TMS frequencies are primarily inhibitory while higher ones are excitatory. We believe this may be due, in part, to the differences that exist between electrons in conductors as opposed to the flow of ions in axons. In essence, ions do not behave as perfect electrical particles and face a time lag when forced to change their direction. This delay or lag in the effect of a magnetic field on a conductor is called “hysteresis”.

Back then we believed that lower frequencies could target the multiple horsetail axons of radial cells that frame the minicolumn. Higher frequencies, with the proposed inefficiencies caused by hysteresis, would no longer be able to take advantage of the specificity offered by the geometrical orientation of cells and end stimulating all or most cells. Since the majority of neurons within the cortex are excitatory, the end effect of higher frequencies would be one of stimulation. This set of assumptions allowed us to select some of the parameters for our first clinical trial. In the case of ASD, we elected to use low frequencies in order to increase the inhibitory tone of the cerebral cortex. Given the high prevalence of seizures and abnormalities of EEGs in ASD, using higher frequencies seemed irresponsible.

After selecting low frequencies for our initial clinical trial the question still remained as to what brain region we would target. TMS focuses most of its energy in an area smaller than a quarter within the cerebral cortex. The selection of a target area was fairly simple, as our own autopsy studies had led us to suggest that the prefrontal cortex was the site of salient minicolumnar defects in ASD. Several hundred studies by now have implicated a specific brain region (i.e., dorsolateral prefrontal cortex) in the generation of executive functions (e.g. theory of mind) that appear to be affected in autistic individuals. Targeting the dorsolateral prefrontal cortex also offered the advantage of using the large connectivity of this brain region to our advantage. Back then we thought that correcting the inhibitory defect within the dorsolateral prefrontal cortex could lead to a cascading effect in improving many other areas of the cerebral cortex.

Thus far (2016) we have reported 14 clinical trials within the medical literature using TMS in ASD individuals. These trials have involved treating close to 200 individuals with little, if any, side effects. Our first trial was Epub in 2008. Another group led by Peter Enticott from Australia similarly reported beneficial effects 2 years later and many others have joined since. Benefits have been reported using different outcome measures and targeting different areas of the brain. Outcome measures in our own studies have included behavioral screening techniques, measures of attention, self-monitoring for mistakes, and unbiased electrophysiological parameters.

When we introduced TMS as a possible therapeutic intervention for ASD it was the first attempt at targeting some of the brain abnormalities reported in this condition. Applying this technique has been done as an outpatient procedure lasting not more than 30 minutes depending on the individual. Now we are combining the technique with neurofeedback and finding synergistic effects with the same. Our future studies will focus on examining the long-term effect of TMS and in developing potential biomarkers with predictive value as to who would benefit from receiving the same.

Additional information regarding TMS and autism in general can be found in other blogs at




5 responses to “Translating TMS from Autism Science to Autism Therapy

  1. I find the function of minicolumns to be very interesting. I’m wondering when I was given botox to reduce my forehead’s muscle activity, if the reduction in activity also affected the minicolumns’ surround inhibition.


  2. For many years, it’s the first time when I’m not `skeptical at first sight` when hearing about a possible treatment for autism that directly addresses the brain (either drugs or neurological). I used to trust only in sensory diets and educational-behavioral approaches (not ABA, but TEACCH). But now, who knows…


    • I think that you have to be careful hen talking about treatments for autism. RIght now, at least in our experience, rTMS has only been of benefit in higher functioning individuals. If somebody has structural defects, relentless seizures, etc. rTMS may not be of help.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s