The origins of the inhibitory/excitatory hypothesis in autism

Many years ago I wrote an article that described abnormalities of the cerebral cortex in autistic individuals. Because the same dealt with basic concepts of neuroscience I thought it best to have it published in a neurology journal as opposed to addressing mental health professionals through a psychiatry journal. In its first submission, the article was accepted with corrections in a rather prestigious journal (i.e., Neurology) and in span of a decade or so the article has been cited well over 500 times in the medical literature. In the end, I think I made the right decision in publishing it in Neurology. My only regret was that most of the requested corrections by the reviewers entailed reducing the length of the article to about half its original size. Detailed explanations about our validation of the methodology and the lengthy discussion section were significantly trimmed down.

One of my teachers in neuropathology at the Johns Hopkins School of Medicine was Dr. Arthur C. Clark. He now works at the University of Calgary. Arthur once told me that many of his articles had been rejected but later resurrected and published under different circumstances. I seem to hear him tell me, “Never waste a beautiful paragraph!” In the case of my Neurology article, that is exactly what happened. I took many of the deleted paragraphs and published them as their own article: “Neuronal density and architecture (Grey Level Index) in the brains of autistic individuals”.  Given the lengthy time towards publication in the journal Neurology, I was able to publish this second article the same year as my original publication (2002). Although published in a less prestigious journal, I thought the latter article made better sense than our original publication. In addition, many of the remaining paragraphs that were cut from our original article made their way into other publications, primarily book chapters.

I still regard having had to dissect our original contribution as a serious loss. Occasionally I will hear somebody complain, as an example, that we didn’t appropriately quantify the modules of the cerebral cortex (minicolumns) and, for that matter, using stereology would have been a better choice for quantitation. Then I have a vivid flashback that I had modeled our original results with stereology but was forced to cut out that part only to publish our methodology later on as a book chapter not readily accessible to many readers. Aaaaaaargh!

Well, at least one thing did go well with our Neurology article. I thought our findings were quite real and offered a different perspective on both autism and neuropathology in general. This mode of thinking made me the person most interested in wanting to validate the findings. Our group did so by reproducing the minicolumnar abnormality by using different techniques as well as other independent populations. Also, the specificity of the findings was proven when we studied the neuropathology of many other mental conditions.


Besides providing the validity of our original findings I soon found myself trying to explain their significance. In 2002, the same year as our Neurology publication, I published another article entitled, “Clinical and macroscopic correlates of minicolumnar pathology in autism”. It was the intent of the article to prove the significance of our findings by making a number of predictions.

In making predictions I was helped in that I understood a lot about the pathology of autism at least from a minicolumnar perspective. The minicolumns that we had reported as being pathological in autism, were specially abnormal (i.e., thinned out) in their peripheral compartment. This portion of the minicolumn is where many of the inhibitory cortical elements are found and has been denoted by some electrophysiologists as confering a shower curtain of inhibition to the conductive core (pyramidal cells) in the center of the minicolumn.  You could probably consider by analogy (to the minicolumn) an insulated copper wire, where the insulation is the peripheral inhibitory elements of the minicolumn.  A defect in the insulation of the minicolumn would cause for signals streaming through the core of this module to suffuse into adjacent minicolumns.  The end result would be the generation of a cascading or amplification effect. This simple explanation made a lot of sense to me and I publicized the same through lectures (well ahead of our publication) and a couple of articles.

One of the people that acquired a great deal of interest in our observations was John Hussman.  John is a well-known and quite successful mutual fund manager who has an autistic son.  Given his interest and financial status, John was interested in the potential of this idea for possible therapeutic interventions.  John did publish a short letter to the editor (circa 2001) in this regard but, unfortunately, we did not further pursue  the possibility of funding a clinical trial.

I expanded on these observations (and the inhibitory/excitatory hypothesis of autism) in two publications.  The first article was published in 2002 (Clinical and macroscopic correlates of minicolumnar pathology in autism).  Another article was submitted in 2002 but took an inordinate amount of time to get published: Disruption of the inhibitory architecture of the cell minicolumn: implications for autism, published in 2003.

In the publication on clinical and macroscopic correlates of minicolumnar pathology in autism we made several predictive observations. Not only did we believed that minicolumnar pathology could explain the presence of seizures (an inhibitory deficit) and other sensory abnormalities, we also believed it could: 1) account for abnormalities in fast frequency rhythms of the brain (so-called gamma frequencies) that themselves depend on the patency of inhibitory cells (see an earlier blog:, and 2) be responsible for changes in the blueprint of corticocortical connectivity of the brain which could be observed macroscopically by MRI.

Sometime in 2003 two distinguished researchers, Rubenstein and Merzenich, published a hypothesis about the inhibitory excitatory imbalance in autism.  The article provided a wealth of good ideas.  However, the supposition of an inhibitory excitatory imbalance in autism was not based on any data of their own but rather on a literature review, for which my paper was aptly quoted (and much appreciated).  Apparently, the reputation of the researchers, swayed later authors in giving them credit for having spoused the original idea of such a signal imbalance in autism.

Ever since our original publication we have pursued our findings in many different ways.  In terms of neuropathology our published data indicates the presence of a migratory disturbance during brain development in autism.  In effect, it appears that primitive cells migrating radially to the cortex become desynchronized from those that migrate tangentially. Radially migrating cells are primarily excitatory (future pyramidal cells) and those that do so tangentially are inhibitory (future interneurons).  In terms of interventions we decided on applying Transcraneal Magnetic Stimulation (TMS) as a way of building the inhibitory surround of otherwise defective minicolumns (see blog:  The trial was recognized by an EUREKA award from the NIMH as being highly innovative and significant and our results were published over the years in numerous publications and book chapters.  Overall we have treated over 200 patients with this technique with good results. Given our belief of an inhibitory deficit in autism and how gamma synchrony depends on the patency of this signaling system, we have used gamma frequencies as an outcome measurement (publication entitled: Effects of low frequency repetitive transcranial magnetic stimulation (rTMS) on gamma frequency oscillations and event-related potentials during processing of illusory figures in autism, 2009).  We have found that abnormalities of gamma frequencies are a core deficit of autism and believe that they may provide a severity dependent measure that allows us to judge the results of medical as well as educational interventions.


Last year the SFARI foundation published a blog giving credit to Rubesntein and Merzenich for having described the inhibitory excitatory theory of autism ( The blog elaborated on the significance of these findings: “Electroencephalography studies have also shown that gamma oscillations are altered in individuals with autism. Studies further suggest that these are the pivotal GABA cells (Note by MFC: GABA is the neurotransmitter for inhibitory neurons) responsible for orchestrating the timing of critical periods in which the brain undergoes development.” In the comment sections to the blog a staff person for SFARI posted: ” Tomorrow’s webinar may shed some light on this discussion. Vikaas Sohal will discuss his work showing how excitation-inhibition imbalance in distinct subtypes of neurons may contribute to aspects of autism. And more to the point, his results suggest how the E/I hypothesis might be updated to be useful for designing new experiments and, eventually, therapeutics. ” This all came about 11 years after our original publication (quoted several hundred times), multiple awards, numerous clinical trials as well as publications from our part elaborating on the subject!  Que sera, sera.

10 responses to “The origins of the inhibitory/excitatory hypothesis in autism

    • There were several case reports where anticonvulsants used to treat epilepsy (e.g., a patient with tuberous sclerosis who also had an autism phenotype) iproved autistic behaviors. Unfortuntely these agents are non specific and tend to act on all inhibitory cells regardless of type, i.e., chandlier, double bouquet, basket cells. It is possible that considerably increasing the inhibitory tone would improve some of the symptoms that handicap autistic individuals but unfortunately they would also cause serious side effects (e.g. lethargy or stupor). We thought that rTMS would be more specific by targeting the inhibitory elements at the periphery of the cell minicolumn.


  1. I thought that actually GABA is a neurotransmitter for inhibitory neurons, not the single neurotransmitter for inhibitory neurons. Norepinephrine also serves as an inhibitory neurotransmitter. Interestingly, purkinje cells in the cerebellum which have been implicated in autism use both norepinephrine and GABA as inhibitory neurotransmitters. As I recall in one autopsy study there were suggestions of some defects of the locus coereleus which is were norepinephrine is manufactured, so it seems it may be possible that norepinephrine could be implicated in inhibition problems of autism as is GABA .


    • Thanks for the comment. GABA is the neurotransmitter used by a majority of inhibitory cells, although as you point out, it is not the only one. Curiously, it may have excitatory effects during brain development. Some people believe that a lag in the conversion of the action from excitation to inhibition for GABA may be the basis of some developmental conditions.

      I remember your interest in the locus coeruleus from a previous blog that you wrote. We may have touched upon the subject when I wrote about the brainstem. I do believe you are correct in your assessment. Abnormalities in the same may also be related to sleep disturbances observed in some ASD individuals.

      Ojala todo te siga bien. Saludos.


  2. They have been studied in schizophrenia as well as several neurodevelopmental conditions (e.g., tuberous sclerosis, Rett sybdrome, rubella babies). Although there have been abnormalities reported for these conditions (e.g., schizophrenia) the changes related to ASD remain specific to the condition.


    • That’s great news. The problem in autopsy studies is the inconsistancy and non specificty of the findings. If reduced minicolum cells is specific to autism I hope more groups will replicate these findings since autism specific findings are so rare.


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