Autism spectrum disorder (ASD) arises from a complex mix of genetic, neurobiological, and environmental factors, with no single cause identified. Current prevalence in the United States is about 1 in 31 children, based on 2022 surveillance data from the CDC. While the precise interplay of causes varies from person to person, decades of research have clarified the major contributors and ruled out several persistent myths.
Genetics Play the Largest Role
Twin studies consistently show that ASD is one of the most heritable neurodevelopmental conditions, with genetic factors accounting for an estimated 60 to 90 percent of overall risk. But “genetic” doesn’t always mean “inherited from a parent.” A substantial share of cases involves de novo mutations, spontaneous genetic changes that occur for the first time in the child rather than being passed down.
How much de novo mutations matter depends on the family’s background risk. In families with no prior history of autism (low-risk families), spontaneous mutations contribute to an estimated 52 to 67 percent of cases. In families where autism already runs in the family (high-risk families), de novo mutations account for only about 9 to 11 percent of cases, because inherited genetic variation is doing most of the work. This distinction helps explain why autism can appear “out of nowhere” in some families while clustering in others.
No single gene causes autism. Researchers have identified over a hundred genes that contribute to risk, each with a small individual effect. Some of the most commonly implicated genes affect how brain cells communicate at synapses, how the brain develops during pregnancy, and how proteins are regulated inside neurons. The genetic architecture is so varied that two people with autism may share almost none of the same risk variants.
Differences in Brain Development
One of the clearest neurobiological findings involves synaptic pruning, the process by which the brain trims excess connections during childhood and adolescence. In typical development, a burst of synapse formation happens in infancy, and the brain then eliminates roughly half of those connections in the cortex by late adolescence. This pruning sharpens neural circuits and makes communication between brain regions more efficient.
In children with autism, that pruning process slows dramatically. Research from Columbia University found that by late childhood, the number of synaptic spines (tiny structures where neurons connect) had dropped by about 50 percent in typically developing brains but by only 16 percent in the brains of children with autism. The result is a surplus of connections that can make neural signaling noisier and less organized.
Scientists traced this pruning problem to an overactive protein that interferes with the brain’s cellular cleanup system. When that protein is too active, brain cells lose much of their ability to break down and recycle their own components, a process essential for clearing out unneeded synapses. This finding suggests the difference isn’t that extra synapses are being created, but that normal synapses aren’t being removed on schedule.
Epigenetic Changes
Epigenetics sits at the intersection of genes and environment. Your DNA sequence stays the same throughout your life, but chemical tags attached to that DNA can change how actively certain genes are read. The most studied of these tags is DNA methylation, a process that typically dials gene activity down.
Children with autism consistently show altered methylation patterns compared to typically developing children. Overall, global methylation levels tend to be lower in children with ASD. But the picture is more nuanced at the level of individual genes. Some genes involved in cell communication and brain development show increased methylation (meaning they’re less active), while others show decreased methylation (meaning they’re more active than expected). These changes have been observed in blood samples, making them potentially useful as early biomarkers, though no single gene has been found to be consistently altered across multiple studies.
What makes epigenetics particularly relevant is that these chemical tags can be influenced by environmental exposures, nutrition, stress, and other factors during pregnancy and early life. This provides a plausible mechanism for how environmental risk factors might translate into changes in brain development without altering the underlying DNA.
Parental Age and Birth Spacing
Advanced parental age is one of the most consistent non-genetic risk factors. Fathers aged 40 or older have roughly a 40 percent higher chance of having a child with autism compared to fathers aged 25 to 29. The likely explanation is that sperm cells accumulate more spontaneous mutations with each passing year, increasing the odds of de novo genetic changes. Older maternal age also contributes some independent risk, though the effect is smaller and harder to separate from paternal age.
The interval between pregnancies matters too. Children conceived less than 12 months after the birth of a sibling face approximately double the risk of ASD compared to those conceived 24 to 47 months apart. Interestingly, very long intervals (more than 84 months) carry a similar twofold increase. The short-interval risk may relate to maternal nutritional depletion, particularly folate stores that haven’t had time to replenish. The reason for the long-interval risk is less clear.
Why Autism Is More Common in Boys
Autism is diagnosed roughly three to four times more often in boys than in girls. Part of this gap reflects underdiagnosis in girls, who more frequently develop social coping strategies that mask autistic traits. But biology also plays a role through what researchers call the female protective effect.
The core idea is that girls require a greater genetic burden to cross the threshold into observable autism. Evidence for this comes from sibling studies. When a girl does have autism, her siblings carry significantly more autistic traits than the siblings of boys with autism. In a combined analysis of two large population samples, siblings of girls above the clinical threshold were 38 percent more likely to also cross that threshold compared to siblings of affected boys. This pattern, replicated across nationally representative cohorts, suggests that affected girls come from families with a heavier overall genetic load, meaning whatever protective factor girls carry had to be overcome by more numerous or more disruptive genetic variants.
The specific biological mechanism behind this protection isn’t fully identified, but the genetic data show that rare mutations found in girls with autism tend to be larger and more functionally disruptive than those found in boys. In other words, it takes a bigger genetic hit to produce the same behavioral outcome in a girl.
Vaccines Do Not Cause Autism
The claim that vaccines, particularly the MMR vaccine, cause autism has been thoroughly investigated and definitively rejected. A major meta-analysis pooling five cohort studies (covering over 1.25 million children) and five case-control studies (covering nearly 10,000 children) found no association between vaccination and autism. The odds ratios were essentially 1.0 across every analysis, whether researchers looked at the MMR vaccine specifically, the preservative thimerosal, or mercury exposure. The original 1998 study that sparked the concern was retracted due to ethical violations and data manipulation, and its lead author lost his medical license.
Prenatal Environmental Exposures
Several environmental factors during pregnancy have been linked to modest increases in ASD risk, though none are considered primary causes on their own. These include maternal infections during pregnancy (particularly those causing significant immune activation), exposure to certain air pollutants, and use of the anti-seizure medication valproate. Prenatal exposure to valproate carries one of the strongest known environmental risk associations, increasing the chance of autism several-fold.
Complications during birth, such as oxygen deprivation or extreme prematurity, are also associated with higher rates of autism. However, it’s difficult to determine whether these complications independently cause autism or whether they co-occur because the same genetic factors that increase autism risk also increase the likelihood of pregnancy complications.
What ties these environmental factors together is their timing. The developing brain is most vulnerable during the first and second trimesters, when the basic architecture of neural circuits is being established. Environmental exposures during this window can interact with genetic susceptibility in ways that alter brain development, potentially shifting a child from below the diagnostic threshold to above it. For most children diagnosed with autism, the cause is likely a unique combination of inherited risk, spontaneous mutations, and prenatal conditions rather than any single factor acting alone.