Several weeks ago a good friend of mine, Alycia Haladay, asked me to write a criticism of Steven Chance’s recently published article in Brain. The criticism would be posted as a blog for the Autism Science Foundation (https://autismsciencefoundation.wordpress.com/). Steven and his group studied a large series of autistic individuals and controls looking for abnormalities of the cerebral cortex. In the end they reported that the basic unit of information processing of the cerebral cortex, called minicolumns, are abnormally wide in autistic individuals. This, at first, seemed to contradict my own previously published findings reporting reduced size of minicolumns in autism spectrum disorders (ASD).

Minicolumns are the basic template that serves to organize cells within the cerebral cortex. They are arranged vertically in cylindrical fashion with excitatory cells at their centers and inhibitory ones at the periphery. We had previously reported narrower minicolumns in autism with the major defect within its inhibitory surround. This has given rise to the so-called shower curtain analogy. In effect, a shower curtain keeps water inside of the bathtub. A defective shower curtain would provide for water splashing all over the bathroom floor. In a similar way, having a defective shower curtain of inhibition for minicolumns in ASD would cause the information within the minicolumn to cascade over to adjacent minicolumns. This runaway cascade of excitation helps explain some of the symptoms that truly handicap some autistic individuals, e.g., sensory phenomena, seizures. Moreover, the defect in cortical surround inhibition which we described anatomically in autism has now been reproduced using other techniques, i.e., EEG and tactile vibration recordings.

Steven Chance’s results were based on the largest postmortem series to date. Furthermore, he used the same computerized algorithms that we have applied in our past studies. Yet, Steven described that on average, minicolumns were wider. The difference between studies is related to an aging effect. Steven’s postmortem series was large enough to provide for an age correlation. In effect, minicolumns in autism are larger during childhood but become smaller with aging. By tabulating his results and analyzing our previous reports, he was able to reproduce our original findings in older patients. The new findings in the postmortem literature of ASD help explain one of the more often reproduced neuroimaging results. Thus, according to neuroimaging studies brains in ASD are larger during childhood (due to larger minicolumns) but plateau in their growth (due to shrinkage of minicolumns) and acquire a size similar to controls during their teenage years.

This is what Steven wrote to the ASF: “Casanova and colleagues have conducted several previous studies on mini-column structure in autism and, overall, their measurements indicated a narrowing of mini-columns. If the recent study indicates widening, then how can this apparent difference between studies be explained? Certainly, the difference must be taken seriously because our new study is the largest so far conducted (ie. measurements were taken from more subjects than any previous study). There are differences between the studies in the selection of brain regions that were measured and it is well established that brain regions differ in the width of mini-columns, so some regions might be affected more than others. However, at least one brain region was the same as in previous studies, so is there another factor that could explain the discrepancy? Yes, we think so… it depends on the age of the individual. Re-analysis of the past data from Casanova and colleagues indicates the same effect of development and aging as we found in McKavanagh et al 2015. In other words, both studies suggest that minicolumns are wider in youth in autism but then become narrower in later life. So while the average effect might differ between studies depending on the average age of the participants, the same relationship is found with age in both studies.”

The Autism Science Foundation blog dissected both Steven and my own response in order to make them more readable to the lay person (https://autismsciencefoundation.wordpress.com/). This is my original write up to ASF:

Autism Spectrum Disorder and Minicolumns

Many years ago I found myself wondering as to the scant evidence of pathology in mental conditions such as autism. It occurred to me that pathologists classically ascribe the presence of abnormalities to disorders with evident cell loss, gliosis, and/or when cell bodies are reduced in size or stain differently. What would happen if the pathology escaped this level of resolution, i.e., abnormalities that involved ensembles of cells rather than single neurons? This is when I decided to take a look at minicolumns in order to study the pathology of autism.

A minicolumn is a conglomerate of cells whose architecture is repeated several hundred million times throughout the cerebral cortex. Some people think of minicolumns as the microprocessor of a computer. Indeed, Mountcastle thought it was the basic unit of information processing in the brain. More recent electrophysiological experiments have proven that the genesis of executive functions and other aspects of cognition reside in the workings of minicolumns.

Our initial studies in autism have shown abnormalities in the structure of minicolumns that seem specific to the autism spectrum disorder (ASD) when compared to many other mental disorders (e.g., schizophrenia, Down syndrome). Early on, this basic abnormality allowed us to suggest that autism was characterized by an excitatory/inhibitory bias, an alteration in the blueprint of white matter connectivity, and a high prevalence of gamma frequency abnormalities. More importantly it lead us to initiate a therapeutic trial with repetitive transcranial magnetic stimulation (rTMS) to correct the excitatory/inhibitory imbalance caused by this minicolumnopathy. Several other groups have now reproduced our initial positive results with this intervention.

The minicolumnopathy of autism appears to involve changes in columnar size and the variability of the same. Previous studies have shown how these abnormalities vary by brain parcellation (sometimes being larger or smaller) and how the findings relate to the various clinical findings observed in autism spectrum disorder. Joining our published data with that of Chance (McKavanagh et al., 2015) it appears that the minicolumnopathy of autism bears an age dependent profile wherein these neuronal ensembles are wider early on but diminish in size with aging. This is an important observation that demands further studies.

The findings of minicolumnar abnormalities in autism should be taken in the context of other reproducible neuropathological findings; e.g., heterotopias, supernumerary cells in both the gray white matter junction and in the molecular layer. These findings indicate that the minicolumnopathy of autism is a malformative process better known as a cortical dysplasia. The presence of cortical dysplasia provides a common explanation for intractable seizures. The pathology may be selective as to whom it affects as, on occasion, involved individuals have either a specific genetic or immunological profile. More importantly, it seems that some individuals with cortical dysplasias develop a degenerative course the basis for which is poorly understood at present. Not coincidentally other neurodegenerative disorders like Alzheimer’s disease often manifest thinning of minicolumns. It may be that the age-related changes in the minicolumnopathy of autism may constitute an adaptive response meant to correct for changes in the cerebral cortex (e.g., losing cells in malformed excitatory inhibitory cell dyads) or the beginning of a neurodegenerative process whose manifestation may appear later on in life.

Manuel F. Casanova, MD
SmartState Endowed Chair in Translational Childhood Neurotherapeutics
University of South Carolina School of Medicine Greenville Campus
Greenville Health System