Autism has increased from 1 in a 1000 in the 1990s to 1 in 31 in 2025, a  3000% increase!1 Why is autism increasing SO much?

Ever since Bob Wright, former president of NBC, became the grandfather of a child with autism and created Autism Speaks, awareness of and research on the condition has skyrocketed. Practically everyone knows someone who has a child with autism.

Welcome to my world. I am a developmental and behavioral pediatrician who has specialized, over the last 30 years in diagnosing and caring literally thousands of children and adolescents with ASD. 

In this blog I’ll start with what we know from science about autism’s steady increase and then, at the end, I’ll share a new hypothesis about autism’s rise based on my office observations, confirmed by recent research.

 

First, autism is increasing because we have defined it more broadly. This is reflected in the term autism spectrum as defined by the DSM 5.  This relatively new definition includes children that we, in the medical profession, never would have included before. At first, only children with ‘classic autism’—no words, in their own world, and focused solely on repetitive behaviors—would have been included. Then we added children with some language but with serious challenges socializing whose special interests—cars, trains, wheels, Thomas the Tank Engine, LEGOs, YouTube—took them out of social interaction. Now we have expanded the spectrum to include those with no language delays at all but who are ‘quirky’ and ‘different’, struggles with making or keeping friendships, and have dominating interests like knowing all the sports stats for the Yankees or the train schedules for the B&O.  No question, a broadening of the definition is one major cause of autism’s increase.

Second, there is a hunt on. We’re looking for it more and earlier. The media is making parents hyper-aware that their children might have symptoms of autism. I recently evaluated an infant who was only 6 months old because his mother had seen on Instagram that autism’s symptoms could be diagnosed as early as 6 months by visual tracking devices. I told her that these tracking tools were still experimental but I took her concerns seriously. Even without a formal diagnosis, I was able to make recommendations to help this mom engage with her hard to engage son using The P.L.A.Y. Project’s methods.  Then, at 14 months—the age at which science has determined most children CAN be reliably diagnosed3—I confirmed the mother’s suspicions. Her son, though he had improved in his engagement and interaction, still met criteria for autism according to the DSM 5.  The American Academy of Pediatrics now recommends that all pediatricians screen for ASD at 18 and 24 months4, which is a very good thing given how effective intensive early intervention can be.  So, even as we are broadening the definition to include milder cases, we are looking for it earlier, thus boosting prevalence.

Third, most of us in the field suspected that environmental factors had to be a major contributor to autism’s increase. In fact, studies indicate that the environment indeed has contributed to autism increase but accounts for only about 20% of the rise5. The suspected list of environmental causes is growing and includes such things as exposure to environmental toxins and pesticides, to antidepressant medications during pregnancy, to the survival of very premature babies, having older parents, and maternal obesity among others.  For several years, at the turn of the millennium, it was thought that immunizations and/or mercury in immunizations might be a cause of ASD but 20+ large-scale, well-done studies have concluded that immunizations do not cause autism, period!5 Dietary causes have also largely been disproven too7 (though children on the spectrum have more than their share if g.i. issues!). No doubt other environmental causes will be discovered. 

Finally, autism is now known to be largely genetic, with genetic factors accounting for 70-90% of cases8. In a recent analysis of very large samples of the DNA of families who have children with autism, researchers have identified between 200 to a 1000 ‘autism genes’ with half due to single gene mutations9. These genes code for the many different kinds of neuronal (brain cell) connections and when there are abnormal additions or deletions to these genes, the brain cells don’t connect as they should, leading to the diagnostic symptoms of autism. Twin studies are confirmatory. Siblings are at higher risk for autism.

So, now we know that the cause of autism is mostly genetic with some environmental contributions, but, and this is important, no other ‘genetic’ conditions (e.g. Down Syndrome) have shown the same enormous increase as autism spectrum disorders! In other words genetics can’t explain the autism’s increased prevalence. 

I collect three generation genograms (family trees) on all of my families as part of a good medical history. Over many years, I began to see patterns in the genograms. Not only was autism showing up in some generations of the family tree, but I noticed certain traits appearing including shyness, social anxiety, obsessive/compulsiveness/OCD, perfectionism, neat freaks, the detail-oriented, rigid and/or orderly, collectors and hoarders, gamers and fanatics. These traits, I saw, are often inherited. Then I began to notice certain occupations in my genograms: accountants, engineers, and IT specialists, electricians, mechanics, welders, and tool and die workers, which are all detail oriented occupations. Then I began to notice that the genograms were showing these traits and occupations on both sides of the family. 

 

So, I began to think that mating patterns might be a missing piece of the cause of autism’s increase. I noticed that, in my practice at least, when ‘smart people married smart people’ (with ‘autistic traits’ in their family tree) the risk of autism went up. I was puzzled, however, because there was no research evidence to support this observation of mine. Then, in 2020 a large-scale well done research study on men in the Swedish military revealed that the smarter and more detailed oriented the men were, the more likely they were to have a child with autism10!  There is a genetic term for this very modern phenomenon—positive assortative mating—defined as a form of sexual selection in which individuals with similar phenotypes (i.e. traits) mate with one another more frequently than would be expected under a random mating pattern.

Why is there more assortative mating? Are more people with autistic traits marrying each other? Here’s my working (i.e. not yet proven) anthropological hypothesis: For the last 300,000 years, homo sapiens selected their mates within a very small group of hunter gatherers. As society over the last 10,000 years evolved to living in small towns and farms, humans mated within the limited confines of those in the community. If someone lived in a traditional society, their parents often arranged their marriages to complete strangers. In none of these societies did one did pick one’s partner based on personality traits; but, with the extraordinary mobility in all modern societies over the last 50-75 years (viz. over the time of autism’s steep rise), the number of eligible men and women traveling off to school and/or workplaces far from home has markedly increased the pool of mates to choose from. We modern homo sapiens tend, both statistically and naturally, to marry someone like us—based on similar personalities and interests, similar intelligence and even traits—more than ever before in human history. Now we can even pick our mates, who can live anywhere in the world, by menu on dating apps! In short, there has been, in the last 50-75 years, a large-scale and global change in the way we choose our mates due human mobility. This would be both an environmental AND genetic explanation for autism’s rise. 

Is it possible that the dramatic increase in autism has to do with the fundamental changes of a highly advanced society including more toxic exposures (NOT immunizations!), older parents, more surviving preemies, better public health surveillance, more sophisticated diagnostic tools, AND the genetic characteristics of parents who are linear thinkers and detail-oriented people who work in technical and mechanical fields leading to a genetic-environmental risk for autism? Time will tell if my hypothesis is right. 

In the meantime, it is a fact that autism prevalence is at an all-time high. It is a fact that because of this, there are too many children waiting for therapist based services. In my next blog entitled ‘Don’t wait! Participate!’ I will argue for the importance of empowering parents to be their child’s best helper so we can help this growing number of children on the spectrum reach their full potential. Stay tuned.  

References

1Autism prevalence 2025, CDC:

https://www.cdc.gov/autism/data-research/index.html

2FeinD, et al (2013). Optimal Outcomes in Individuals with a History of Autism. J Child Psychol Psychiatry.54(2): 195-2-5

3AAP Study Recommending Screening at 18 and 24 months for ASD: https://pediatrics.aappublications.org/content/136/Supplement_1/S41

4 Pierce, et al, Evaluation of the Diagnostic Stability of the Early Autism Spectrum Disorder Phenotype in the General Population Starting at 12 Months  https://pubmed.ncbi.nlm.nih.gov/31034004/

5Bai D, et al. (2019) Association of Genetic and Environmental Factors With Autism in a 5-Country Cohort, JAMA Psychiatry.76 (10): 1035-1043 https://jamanetwork.com/journals/jamapsychiatry/fullarticle/2737582#227982497

6CDC report on Immunizations and Autism: https://www.cdc.gov/vaccine-safety/about/autism.html?CDC_AAref_Val=https://www.cdc.gov/vaccinesafety/concerns/autism.html 

7Sathe N, Andrews JC, McPheeters ML, et al. (2017) Nutritional and Dietary Interventions for Autism Spectrum Disorder: A Systematic Review. Pediatrics. 139(6):e20170346

8Satterstom FK et al (2020) Large Scale Exome Sequencing Study Implicates Both Developmental and Functional Changes in the Neurobiology of Autism, Cell. 180(3): 568-584)  

9 Genovese A, Butler MG. The Autism Spectrum: Behavioral, Psychiatric and Genetic Associations. Genes (Basel). 2023 Mar 9;14(3):677. doi: 10.3390/genes14030677. PMID: 36980949; PMCID: PMC10048473. (https://pmc.ncbi.nlm.nih.gov/articles/PMC10048473/

10Gardner R, et al (2020). The Association of Paternal IQ with ASD and Its Comorbidities: A Population-Based Cohort Study, Journal of Am. Acad. Of Child & Adol. Psychiatry. 59 (3) 410-421