Can early childhood social deprivation cause autism?
Orphanages and screen time | nature or nurture
There’s an old video that depicts an orphanage. The orphans are toddler-age, and many had been abandoned when they were only a few months old. Some were moved between institutions, never settling down—until they entered the orphanage in which the video is filmed. There, the then toddlers are finally given a chance to bond with a single gentle and attentive caregiver. Remarkably, when they entered the orphanage, some of these children showed autism-like symptoms: social deficits, lack of play, and stereotyped movements like rocking back and forth. These traits diminished as the children bonded with the caregiver.
I don’t know the origins of the video nor whether it is a reliable source. But it is what first raised the question in me whether autism could be caused by social deprivation.
When autism was first identified by Kanner, it was thought to be caused by cold parenting styles. A debate insued—was autism due to nature or nurture?
Enter Michael Rutter who, along with his co-author Susan Folstein, conducted a twin study that showed that there was more concordance in autism symptoms between monozygotic than dizygotic twins. That, in addition to other studies, established autism as a heritable condition. The research field as a whole shifted from studying early childhood social environments to studying genes.
Unlike the video of the orphanage, Michael Rutter is a reliable source. When he was alive, he was not only a world-leading expert in autism but also in attachment theory. His twin study was one of the first to establish the strong genetic component in autism. (Interestingly, I was reading that study, and even back then it was understood that autism has multiple causes, which I wrote about in my last essay). He also wrote the book Maternal Deprivation Reassessed, which reframed the earlier studies on orphanages into our more modern understanding of attachment theory.
Michael Rutter is the ideal person to study the relationship between extreme social deprivation and autism.
He had just this chance when children were rescued from truly terrible conditions in Romanian orphanages in the early 1990s. In brief: the children were typically kept in cots with little adult contact or playtime with each other, some of them had a bottle attached to their cot for self-feeding. Georgette Mulheir describes these orphanages in a Ted Talk. When she entered one, there was silence—no laughter, no playing, no crying.
Some of the children were rescued to various countries, including the UK, where they entered the foster care system. There, Rutter took the opportunity to study the effects of institutional deprivation on the children. He did not expect to see signs of autism.
Indeed, in terms of what is known about the organic basis of autism and the very strong genetic component that is involved, autism would not be expected to arise as a result of severe privation.
– Rutter et. al. (1999) Quasi-autistic Patterns Following Severe Early Global Privation
Nevertheless, when signs of autism began popping up, he dutifully evaluated those that showed symptoms. For the initial study, which followed the children from the ages of 4 to 6, he found that about 6% of the Romanian orphans qualified for an autism diagnosis, which is much higher the general population.
The finding that some 6 % of a group of 111 children from Romania adopted into English families showed a pattern of behaviour that closely resembled childhood autism is remarkable. The resemblance to autism was clearly demonstrated through the use of a well-validated, reliable, standardised diagnostic interview employed by an investigator very well experienced in its use and in the clinical manifestations of autism. It is evident, too, that the number of cases is far too high to be dismissed as a coincidence. The first question to be addressed, then, is whether the syndrome seen in these children constitutes childhood autism or, rather, a different clinical disorder that just happened to mimic it.
– Rutter et. al. (1999) Quasi-autistic Patterns Following Severe Early Global Privation
In addition to social deficits, these children showed intense and restricted interests. Interestingly, the children didn’t tend to display stereotyped movement, like hand flapping and rocking back and forth, which is common in both autism and in institutional deprivation. This is both a surprise to me and, it seems, to Rutter as well. Two children were fascinated in watches. Another held a peculiarly strong interest in vacuum cleaners. Yet another had substantial knowledge about the water sewage system. Importantly, the possibility that these children entered the orphanage due to their autism was ruled out—they all entered within the first few months of life. There also appeared to be a lessening of the symptoms over the two year period of the first study.
Firm in his belief of his previous findings that autism is primarily a genetic disease, and hopeful that the symptoms would continue to lessen, he called the condition quasi-autism.
The group of children were followed up on three more occasions at the ages of 11, 15, and 24. The results of which were published in 2007, 2010, and 2023.
In the 2007 follow-up, it appeared that some of the adoptees had lost their autism symptoms. However, the researchers also identified more children with autism symptoms—up from 6 to 11%. Rutter returned to the question of whether to call the condition autism or quasi-autism:
The follow-up findings indicate that the ‘quasi’ qualifier was justified. A quarter of the children with quasi-autism lost their autistic-like features by age 11…On the other hand, the term ‘autism’ also appeared justified. The diagnostic features of quasi-autism overlapped greatly with ‘ordinary’ autism and, despite the general trend for lessening autistic features over time, there was a substantial degree of persistence.
– Rutter et. al. (2007) Early adolescent outcomes of institutionally deprived and non-deprived adoptees. III. Quasi-autism
In the 2010 follow-up, when they were 15 years old, the autism symptoms continued to persist. Additionally, the fraction of orphans with quasi-autism now sat at 14.9%. The authors reneged slightly on their previous conclusion that some of the children lost their autism:
Although the 2007 paper noted that three children had seemed to lose their autistic manifestations, it was noted that they continued to have problems in other areas. Moreover, the subsequent follow-up at age 15 indicated that the loss of autistic features was less complete than had been inferred to be the case at age 11.
– Kumsta, Kreppner, Rutter et. al. (2010) III. Deprivation-Specific Psychological Patterns
The final follow-up in 2023 (for which Rutter did not take part) confirmed that the autism features persisted into adulthood.
In a 2013 overview, Rutter maintained that quasi-autism was saliently different than other kinds of autism because 1) extreme institutional deprivation, which was the main risk factor, was not found in ‘ordinary’ cases of autism and 2) the Romanian adoptees had an atypical presentation. With the idea that quasi-autism is different than autism, I disagree. The causes of autism are diverse—there are about 1000 genes implicated in the disorder—and the presentation is diverse, too. There is no ‘typical’ or ‘ordinary’ autism. Why should we call what these English Romanian adoptees experienced something different just because we have a probable cause—and it isn’t primarily genetic?
In the same overview, he continued to tout the high heritability of autism, stating that it was over 90%, which is really the upper limit of the heritability estimates. A pretty good (and large) Swedish study estimated the heritability of autism to be between 29-91%.
However, at the same time, Rutter was open to the idea that the environment was a risk factor towards autism, including the possibility of gene cross environment interactions, and that this direction merited further study. With this, I agree.
There are now a few studies that investigate the role of the social environment (but most studies still focus on straight genetics or non-social environmental exposures like pollution). A study on a cohort of nurses found that autism-like traits correlated positively with a history of childhood abuse. A study at Drexel University found that screen time in infancy correlated with autism-like symptoms later on. And there is an ongoing conversation on whether autism when present with blindness may be, in some cases, caused by the lack of social input through the eyes. These are correlative studies, and I still don’t know enough about them to have formed a strong opinion. As for blindness—I wonder if the same conversation can be had about deafness, or muteness?
My own hypothesis is that there may be genetic predisposition for or protective effects against this disorder given certain social environments. Only 15% of Romanian adoptees developed autism. This suggests that not all people exposed to that extreme environment would develop it.
An even bolder hypothesis is that the increase in autism cases in the world is due to the changing social environment. The same way that obesity prevalence increases as our food environment changes, though obesity is also a heritable condition (just look at twins). This is, of course, highly speculative, and I’m not the first to suggest it.