Will We Discover What Causes Autism?

The appointment and recent confirmation of a new US Secretary of Health and Human Services has rekindled a long-standing debate about the cause(s) of autism and may launch a new round of studies on this important topic. More than three-quarters of a century’s worth of ideas, research, and data have not, as yet, yielded conclusive answers on the cause(s) of autism. (I pluralize “cause” because it is very highly likely that there are multiple causes.) Although we can all applaud a national effort to better understand autism, there is an important pitfall to be considered.

The first question that must be addressed is why specific cause(s) remain stubbornly elusive despite already spending tens (if not hundreds) of millions of dollars on studies seeking answers for more than 50 years; there are now thousands of credible scientific publications on autism.

Before delving into this topic, it is perhaps useful—and humbling—to remember that Down Syndrome was first described (as least in modern medical terms) in 1862 by the British physician John Langdon Down,[1] but the actual cause was not discovered until nearly 100 years later, in 1959, by Jérôme Lejeune, Marthe Gautier, and Raymond Turpin.[2] It is also instructive to review the hypothesized “causes” of Down Syndrome prior to the discovery of the actual cause—a third copy of chromosome 21—to better understand what happens when cause(s) are unknown.

It is fair to say that pseudoscience, superstition, and even outright racist and discriminatory beliefs abounded as to what caused Down syndrome. As an example, some scientists and physicians evidently believed that the shape of the eyes of many people with Down syndrome was due to expression of “inferior” “mongoloid” genes. Other scientists hypothesized—and asserted with authoritarian force—that the most frequent causes are emotional shock, anxiety, or disease in the mother during the first months of pregnancy. Others speculated that Down syndrome was caused by syphilis in the mother.[3] One can only imagine the pain and distress—and quack treatments—such proclamations caused.

The current situation in autism in some ways parallels Down Syndrome prior to 1959. There have been many important discoveries as to what increases the risk of autism and but there have also been false causes promulgated that have been harmful (e.g. anxious and emotionally cold “refrigerator moms”).[4] Importantly, in the absence of answers, pseudoscience and superstition have flourished, often persisting even when overwhelming evidence shows these “causes” to be dubious or even outright fraud.

It is crucial that "autism" be defined in detail in future studies. This is challenging because, over time, the autism “spectrum” has expanded to the point that people with relatively mild social awkwardness who self-identify as autistic as adults are sometimes aggregated with nonverbal autistic children and with “Rain Man” forms of high functioning autism. The movie Rain Man includes a skillful portrayal of high functioning autism by Dustin Hoffman, who won an Academy Award for the role. While acknowledging—and supporting—the right of anyone to self-identify as autistic, and that causal studies should focus on these autistic people too, it is also important to fundamentally understand that “pooling” self-identified and “mild” high-functioning cases with autistic children whose developmental trajectories are marked by reduced motivation for social interaction and significant reliance on routines and repetitive behaviors and associated disruptions in speech, language and social developmental and learning that have always been hallmarks of autism will make it virtually impossible to discover the cause(s). These are fundamentally different typologies—with likely distinct causes.

As Laurent Mottron of the University of Montreal asserts: “In research, the inclusion of individuals categorically defined by over-inclusive, polythetic criteria in autism cohorts results in a population whose heterogeneity runs contrary to the advancement of scientific progress.”[5] In plain English, the ongoing lack of precision in defining autism in published studies makes discovering the cause(s) scientifically difficult—if not impossible.

As a practical example on how this impacts discovery, autistic children who meet the full diagnosis first promulgated by Leo Kanner in 1942 have great difficulty participating in neuroimaging assays in which one has to lay motionless in a noisy tube. This means that the many of nuanced studies of neurological correlates of autism include “high functioning” autistic people who can cooperate in the studies, but may in fact share relatively few of the traits that are hallmarks of autism. This also means that we do not know as much as is needed about an autistic brain in order to discover the neurological cause(s) of classic autism, although these studies may help us understand the neural underpinnings of high-functioning autism. Similarly, in an effort to boost sample sizes, population studies of genetic links to autism have often included a “Broad Autism Phenotype (BAP)” that pool cases with only a few autistic traits with those that meet all the diagnostic criteria. Although there can be advantages to phenotyping subtypes of autism in genetic linkage studies, failing to focus on “index” cases of autism typologies makes it difficult to discover genetic causes.

Moreover, discovering causes will be directly dependent on adopting a stable and consistent definition of autism and gaining an understating of the specific causes in index cases of Kanner’s (classic) autism as distinct from high-functioning autism or self-identified autistic adults. This can also be done while simultaneously but independently focusing on high-functioning autism and self-identifying autistic adults as additional distinct autistic form(s). It is perhaps instructive that insulin-resistant[6] (type 2) diabetes was only discovered after Frederick Banting and colleagues reported that insulin deficiency was a primary cause of diabetes (type 1) by focusing specifically on juvenile diabetes[7] in well-characterized patients. What if these scholars had pooled type 1 and type 2 cases (along with cases of prediabetes) into a broad “diabetes spectrum” when searching for causes? Our nation would be wise to prioritize understanding the cause(s) of “classic” autism and high-functioning autism as distinct types as we once again embark on this important quest and avoid adopting definitions “whose heterogeneity runs contrary to the advancement of scientific progress.”[8]