“…the precipitating factor in infantile autism is the parent’s wish that his child should not exist.”

 That may sound outrageous, but when the famous child psychologist Bruno Bettelheim wrote that in his influential 1967 book, The Empty Fortress, it wasn’t that far from a mainstream view.  Leo Kanner, the psychiatrist credited with first describing autism, had proposed in the 1940s that parents of autistic children were distinguished by their lack of warmth. The children, he wrote in 1949, “were kept neatly in refrigerators which did not defrost. Their withdrawal seems to be an act of turning away from such a situation to seek comfort in solitude.” And thus was born the notion of the “refrigerator mother”—a woman whose cold and distant style of mothering created children who were profoundly impaired in their ability to speak or interact socially. Thankfully, advances in genetics and neuroscience have firmly debunked this idea, but for years it  caused enormous shame and guilt for families coping with an already difficult situation.

 By the 1990s, twin studies had found that the heritability of autism spectrum disorders was as high as 60-90 percent--that is, 60-90 percent of the risk of autism in the population is attributable to genetic variation. Compare that to the heritability of breast cancer, estimated at about 27 percent. So its been clear for some time that genes play an important role in autism but until recently little progress had been made in actually finding any of DNA variations involved. It was known for some time that a number of rare classic genetic disorders—like Fragile X syndrome and tuberous sclerosis—could cause autism in children who inherit the mutations that cause these diseases. But that was about it.


In the last several years, a massive effort to identify the genetic basis of autism has begun to bear fruit. Some of the first breakthroughs came from studying a kind of genetic variation that had, until 2004, been almost invisible to geneticists. These copy number variations (CNVs) are segments of DNA sequence spanning a few hundred to a few million letters of the genetic code that are either missing or duplicated. The deleted or duplicated stretches may contain several or even dozens of genes. If our genomes are the book of life, that means that we differ not just in a letter here or there but in how many paragraphs we have. Many of these genetic differences are common and seem to be benign, but some can have profound effects on how the brain functions.  

A growing number of rare CNVs has now been linked to autism, increasing the risk by anywhere from about 5- to more than 20-fold.  One of the most intriguing findings has been that many of these same CNVs also increase risk for a range of neurodevelopmental disorders, including epilepsy, intellectual disability, ADHD and schizophrenia.

 The newest discoveries have come from major advances in the cost and feasibility of actually sequencing the human genome.  In the last few months, scientists have reported the results of sequencing the whole exome—that is the 1-2 percent of the genome that actually contains instructions for making proteins. In April, three papers published in the journal Nature identified several rare mutations that can cause autism. Like many of the CNVs that have been associated with autism, these mutations were present in those with autism but not their parents. Such “de novo” mutations can occur in the germ cells (sperm and egg) that combine to produce a child. How do they happen? We don’t really know, but the answer probably includes some combination of DNA copying errors, chance, and perhaps the mutating effect of some environmental exposure(s). This is just the beginning of new era of discoveries to come from large-scale DNA sequencing studies that could ultimately unravel the complex genetic underpinnings of the disorder. At this point, though, the vast majority of the heritability of autism remains unexplained.

But there’s also been another emerging genetic story that has surprising implications for how we think about autism. Family studies have shown that relatives of people with autism spectrum disorders tend to score higher than other people on traits that make up what’s called “the broad autism phenotype,” which includes socially awkward speech and behavior.  But autism-related traits also appear to be normally distributed in the general population. It’s clear that many people with autism are profoundly disabled. But there is a broad spectrum of severity, and, at the milder ends of the spectrum, no sharp borders between “normal” and “abnormal” social and cognitive functioning.

Could genes that increase the risk for autism spectrum disorders also influence variation in these autism-related traits in the general population? The answer seems to be “yes.” A recent study of 6000 identical and non-identical twin pairs (12 years old) found that the heritability of autism-related traits (social and communication problems and restrictive, repetitive behaviors) was about the same for those who scored at the high extreme end as it was for those in the rest of the population. In other words, genes contributing to autism itself overlap with those contributing to milder autistic traits in the broader population. The implication is that autism-related traits lie on a continuum and exactly where we draw the line to define the disorder is, at some level, a judgment call.

And so, the search for the genes that contribute to this spectrum may tell us something about how all of us become who are.

 Further reading:

1.   Malhotra D, Sebat J. CNVs: harbingers of a rare variant revolution in psychiatric genetics. Cell 2012;148:1223-41.

2.   Robinson EB, Koenen KC, McCormick MC, et al. Evidence that autistic traits show the same etiology in the general population and at the quantitative extremes (5%, 2.5%, and 1%). Archives of general psychiatry 2011;68:1113-21.

Recent Posts in The Other Side of Normal

Your Brain on Psychotherapy

Is it possible to cure a psychiatric disorder with therapy?

Genes and the Autism Spectrum

Could the spectrum be broader than we thought?

Diagnosing Childhood: Can We Do Better?

The five warning sIgns of childhood and the future of psychiatric diagnosis.

What's Normal Got to Do With It?

Understanding the "biology of normal" can help demystify mental illness.