Autism spectrum disorder (ASD) has no single cause. It develops from a combination of genetic factors, differences in early brain development, and prenatal environmental exposures that interact in ways researchers are still working to untangle. Heritability estimates consistently fall between 77% and 85%, making genetics the dominant factor, but they don’t tell the whole story. About 1 in 31 children in the United States is now identified with ASD, based on 2022 surveillance data from the CDC.
Genetics Play the Largest Role
Twin studies provide the clearest window into how much genes matter. When one identical twin has autism, the other is far more likely to also have it compared to fraternal twins, who share only half their DNA. Across multiple statistical models, broad-sense heritability lands between 77% and 85%, with some models pushing even higher. Shared environmental factors, meaning things siblings experience in common like household income or neighborhood, contribute very little to risk on their own.
No single “autism gene” exists. Hundreds of genes have been linked to ASD, many of them involved in how brain cells communicate with each other, how synapses form, and how the brain organizes itself during fetal development. Some of these genetic variants are inherited from parents, while others arise spontaneously as new mutations in the child. This genetic complexity is one reason autism looks so different from person to person.
How the Developing Brain Differs
One of the most concrete findings in autism neuroscience involves synaptic pruning, the process by which the brain trims unnecessary connections between nerve cells. During typical development, a burst of new synapses forms in infancy, especially in the cortex (the brain region most associated with autistic traits). By late adolescence, roughly half of those cortical synapses are pruned away, keeping circuits efficient.
In children with autism, that pruning process stalls. Research from Columbia University found that by late childhood, synapse density in typically developing brains had dropped by about 50%, while in autistic brains it dropped by only 16%. The brain cells of autistic children were also filled with old and damaged components, showing a severe deficit in the cellular recycling system that normally clears out worn-out parts. Researchers traced this pruning failure to an overactive protein that, when too ramped up, shuts down the brain’s self-cleaning machinery. The result is a brain with excess connections, which may help explain differences in sensory processing, social communication, and behavioral flexibility.
Paternal Age Is a Significant Risk Factor
The age of the father at conception is one of the strongest non-genetic risk factors identified. Children born to fathers aged 40 or older are 5.75 times more likely to have ASD compared to children of fathers under 30, even after accounting for socioeconomic status and the mother’s age. This likely reflects the fact that sperm accumulate new genetic mutations over a man’s lifetime, and older fathers pass on more of these spontaneous changes.
Maternal age tells a more nuanced story. After adjusting for the father’s age, mothers between 30 and 39 show no meaningful increase in risk. Mothers over 40 do show a moderately elevated risk, roughly 1.9 times higher, though this is smaller than the paternal age effect and harder to separate from other factors that correlate with later parenthood.
Prenatal Environment and Immune Activity
What happens during pregnancy can shift risk in measurable ways. The maternal immune activation hypothesis suggests that when a pregnant person’s immune system mounts a strong inflammatory response, whether from infection, autoimmune conditions, or high body mass index, it can affect fetal brain development. Recent research shows familial clustering of autism, autoimmunity, and infections, along with genetic overlaps between autism and immune function. This makes it difficult to separate purely environmental immune triggers from underlying genetic susceptibility, but the association is consistent across studies.
Certain medications taken during pregnancy also carry documented risk. The anti-seizure drug valproic acid is a potent example. Children exposed to it during the first trimester face significantly higher rates of ASD. Animal and human studies show it disrupts synaptic signaling, reduces the formation of new brain cells, triggers neuroinflammation, and alters gene expression patterns during critical windows of fetal development. This doesn’t mean every child exposed will develop autism, but it illustrates how chemical exposures can interfere with the same brain-building processes that genetics influences.
Air Pollution and Chemical Exposures
Prenatal and early childhood exposure to fine particulate air pollution (the tiny particles produced by vehicle exhaust, industrial emissions, and wildfires) is linked to increased autism risk. A Harvard study found that exposure to a specific increase in fine particulate matter raised ASD risk by 31% during the prenatal period and by 64% during early childhood. These particles are small enough to cross the placenta and reach the developing brain, where they can trigger inflammation and oxidative stress.
Heavy metals also appear to play a role. Exposure to lead, cadmium, and manganese has been linked to changes in how certain genes are chemically tagged, a process called epigenetic modification. These tags don’t alter the DNA sequence itself but can switch genes on or off, sometimes affecting genes already associated with autism. Maternal smoking during pregnancy has similar effects, altering the chemical tagging of genes involved in social bonding and neurodevelopment. These epigenetic changes can be long-lasting, persisting well beyond the original exposure.
How Genes and Environment Interact
The most accurate way to think about autism’s causes is through a gene-environment interaction model. A child may carry genetic variants that create susceptibility, and prenatal exposures can then amplify or activate that risk. Research across multiple countries has found that differences in epigenetic patterns among autistic individuals are best explained by this combined model, not by genetics or environment alone.
This framework helps explain why autism runs in families without following a simple inheritance pattern. Two siblings can share many of the same risk genes, but differences in prenatal exposures, parental age at conception, or even random mutation events can lead to very different outcomes. It also explains why autism prevalence varies across populations with different environmental exposure profiles, even when genetic backgrounds are similar.
Vaccines Do Not Cause Autism
The claim that vaccines cause autism has been investigated extensively and rejected by every major health authority in the world. In December 2025, a WHO expert committee reviewed 31 primary research studies published between 2010 and 2025, covering data from multiple countries, and concluded there is no causal link between vaccines and autism. This reaffirmed the same conclusion the committee had reached in 2002, 2004, and 2012.
The review specifically examined concerns about aluminum adjuvants (used in some vaccines to boost immune response) and thimerosal (a preservative). High-quality evidence showed no association between the trace amounts of aluminum in vaccines and ASD, and no link between thimerosal-containing vaccines and autism. The original 1998 study that sparked vaccine fears was retracted due to ethical violations and data manipulation, and its author lost his medical license. Decades of research involving millions of children have consistently confirmed that childhood vaccines do not increase autism risk.