Autism has no single cause. It arises from a combination of genetic factors, prenatal environmental exposures, and differences in early brain development that interact in complex ways. Heritability estimates from twin studies range from 64% to 91%, making genetics the strongest contributor, but genes alone don’t tell the whole story.
Genetics Play the Largest Role
Autism runs in families, and twin studies consistently show that identical twins are far more likely to both be autistic than fraternal twins. A meta-analysis of twin studies found that genetics account for roughly 64% to 91% of the variation in autism risk. That’s an enormous genetic influence, comparable to conditions like height or schizophrenia.
The genetics involved are not simple. There is no single “autism gene.” Instead, hundreds of genes contribute, each adding a small amount of risk. Some of these genetic changes are inherited from parents, while others are brand new mutations (called de novo mutations) that arise spontaneously in the egg, sperm, or early embryo. De novo mutations are estimated to contribute to 30% to 39% of all autism cases. In families with no prior history of autism, that figure jumps to 52% to 67%, suggesting that spontaneous genetic changes are especially important when there’s no inherited predisposition.
In families where multiple children are autistic, inherited common genetic variants play a larger role, and de novo mutations are less frequent. This distinction helps explain why autism can appear “out of nowhere” in some families while clearly running through generations in others.
How Brain Development Differs
The genetic differences linked to autism affect how the brain wires itself during development, particularly in how brain cells form and maintain connections called synapses. One process that appears disrupted is synaptic pruning, the brain’s natural way of trimming excess connections to create more efficient neural circuits.
Research on mice carrying an autism-linked gene mutation found that the brain’s immune cells (which normally help with pruning) became overactive, stripping away too many connections at key developmental stages. This excessive pruning didn’t happen at birth but kicked in during a middle developmental window and persisted into adulthood, reducing the density of mature, functional connections in brain regions important for learning and memory. The immune cells responsible showed physical signs of heightened activity: swollen cell bodies, shorter branches, and increased markers of cellular digestion.
This over-pruning was selective. It targeted one side of the connection point while leaving the other intact, suggesting the process isn’t random destruction but a specific imbalance in how neural circuits are refined. While this research used animal models carrying one particular mutation, it illustrates a broader principle: autism-linked genes often converge on the same developmental processes, disrupting the timing, pace, or precision of how the brain builds its networks.
Prenatal and Birth-Related Risk Factors
Several factors during pregnancy and birth are associated with increased autism risk, though none are strong enough on their own to “cause” autism. They likely interact with genetic susceptibility.
- Parental age: A meta-analysis of nearly 30 studies found that the oldest categories of maternal and paternal age increased autism risk by approximately 40% and 50%, respectively. Advanced paternal age (typically over 40) carries a slightly higher risk than advanced maternal age (over 35), possibly because sperm accumulate more spontaneous mutations over time.
- Preterm birth and low birth weight: Babies born before 32 weeks of gestation have about 2.5 times the odds of an autism diagnosis compared to full-term infants. Very low birth weight (under 1,500 grams, or about 3.3 pounds) raises the odds roughly threefold. Low birth weight under 2,500 grams increases odds by about 60%. These associations hold even after adjusting for maternal age, smoking, and other factors.
- Certain medications during pregnancy: Valproate, an anti-seizure medication, is one of the clearest environmental risk factors. Prenatal exposure significantly increases autism risk even after accounting for the mother’s epilepsy and psychiatric history. Importantly, other common anti-seizure medications (carbamazepine, lamotrigine, and others) do not carry the same risk, making this specific to valproate rather than to epilepsy treatment in general.
Epigenetics: Where Genes and Environment Meet
Epigenetics offers the most compelling explanation for how environmental exposures translate into changes in brain development. Epigenetic modifications don’t alter your DNA sequence. Instead, they act like dimmer switches on genes, turning them up or down. These modifications include chemical tags added to DNA or to the proteins that DNA wraps around, and they can be influenced by everything from air pollution to nutrition to stress.
Valproate is a clear example of this mechanism in action. At therapeutic doses, it directly interferes with the molecular machinery that controls gene expression, altering both the chemical tags on DNA and the proteins packaging it. This means a pregnant person taking valproate isn’t just exposing the fetus to a drug; the drug is actively reshaping which genes get turned on or off during critical periods of brain formation.
Prenatal exposure to air pollution, particularly compounds from burning fossil fuels, may similarly affect brain development through epigenetic changes. Early life experiences, including the quality of caregiving, can also alter gene expression patterns in brain regions involved in stress response. These findings reinforce that “genetic” and “environmental” causes of autism are not separate categories. They are deeply intertwined, with the environment literally shaping how genes behave.
The Gut-Brain Connection
A growing body of research points to differences in gut bacteria composition among autistic individuals. Most autistic people experience gastrointestinal symptoms, and the gut produces roughly 90% of the body’s serotonin, a chemical messenger that plays roles in mood, sleep, and sensory processing. The gut communicates with the brain through the nervous system, hormones, immune signals, and the byproducts of bacterial digestion.
A multi-national study of over 5 million births across five countries found that children born by cesarean section had a 23% higher risk of autism compared to those delivered vaginally, possibly because vaginal delivery seeds the newborn’s gut with beneficial bacteria from the birth canal. Early antibiotic use, which disrupts developing gut bacteria communities, has also been linked to altered neurodevelopment. These associations don’t prove that gut bacteria cause autism, but they suggest that the gut microbiome may be one piece of a much larger puzzle, particularly for the gastrointestinal and sensory symptoms many autistic people experience.
Why Autism Is More Common in Males
Autism is diagnosed in males roughly four times more often than in females, with estimates ranging from 2.8:1 to 6.4:1. This isn’t simply because girls are underdiagnosed (though that’s part of it). There appears to be a genuine biological “female protective effect” that raises the threshold of genetic risk needed for autism to develop in girls and women.
Evidence for this comes from multiplex families (those with more than one autistic child). When researchers measured autism-related traits in these families, females showed a distinctive split: they tended to cluster at either low or high trait levels, with fewer in the middle. Males showed a more even spread. This pattern suggests that females can carry a substantial genetic load for autism without crossing the diagnostic threshold, something that “breaks through” only when risk factors accumulate past a higher tipping point. Notably, a large family study found that many unaffected sisters of autistic children had histories of early language delay with autistic qualities of speech that later resolved, offering a glimpse of how this protective mechanism might work in practice.
The exact biology behind this protection remains unclear. It likely involves some combination of sex differences in gene expression, brain anatomy, hormonal influences, and even socialization patterns that may buffer girls’ social development.
Vaccines Do Not Cause Autism
The claim that vaccines cause autism has been thoroughly investigated and rejected. Reviews by the Institute of Medicine and the Agency for Healthcare Research and Quality concluded with high confidence that there is no association between the MMR vaccine and autism. This finding is based on extensive observational evidence across large populations. The original 1998 study that sparked the concern was retracted due to ethical violations and fraudulent data, and its author lost his medical license.