Autism is not caused by any single factor. It develops from a combination of genetic predisposition and environmental influences during pregnancy, with genetics playing the dominant role. Large-scale studies estimate that hereditary factors account for roughly 83% of autism risk, while environmental factors make up the remainder. There is no one thing that “makes” a baby autistic, and in most cases, the process begins long before birth.
Genetics Drive Most of the Risk
The strongest evidence for autism’s genetic roots comes from twin studies. Identical twins, who share 100% of their DNA, both have autism between 62% and 96% of the time when one twin is diagnosed. For fraternal twins, who share about 50% of their DNA like any siblings, that rate drops to between 5% and 40%. That gap points clearly to genes as the primary driver.
A large analysis combining data from multiple studies estimated autism’s heritability at 83%, meaning genetic differences account for most of the variation in who develops the condition. Shared environment, the factors siblings experience in common like household income or parenting style, contributed almost nothing to risk. The remaining 17% came from what researchers call “nonshared environment,” which includes unique prenatal exposures, random biological variation, and other factors that differ even between twins in the same womb.
Hundreds of genes have been linked to autism, and no single gene accounts for more than a small fraction of cases. Many of these genes are involved in how brain cells form connections with each other. Some people carry rare mutations with large effects, while others carry common gene variants that each nudge risk slightly upward. This is why autism runs in families but doesn’t follow a simple inheritance pattern like eye color.
How the Brain Develops Differently
Many of the genes associated with autism are involved in building and maintaining synapses, the junctions where brain cells communicate. During typical development, the brain overproduces these connections in early life and then “prunes” the excess ones, keeping the circuits that get used and discarding those that don’t. In autism, this pruning process appears to work differently, sometimes leaving more connections than usual in certain brain regions.
One visible sign of this shows up in head size. Babies who are later diagnosed with autism often have normal or slightly small heads at birth, then experience unusually rapid head growth starting around four to six months of age. About 60% of children with autism show this pattern to an extreme degree, compared to just 6% of typically developing children. By ages two to four, 90% of boys with autism in one study had larger than average brain volumes. This early overgrowth phase is followed by a slowdown, and by toddlerhood, growth rates return to normal. The pattern suggests that something in early brain development is running on a different timeline.
Prenatal Environment Matters
The 17% of autism risk that isn’t genetic comes largely from the prenatal environment. The developing brain is sensitive to disruptions during pregnancy, and several specific exposures have been linked to modestly increased risk.
Maternal immune activation is one of the better-studied pathways. When a pregnant person’s immune system mounts a strong response, whether from an infection or other inflammatory trigger, the signaling molecules involved can cross the placenta and affect fetal brain development. Immune signals play a role in every stage of how the fetal brain wires itself, so disruptions to this process can alter how neural circuits form. Animal studies consistently show that triggering a strong immune response during pregnancy produces offspring with autism-like behaviors and brain changes.
Certain medications taken during pregnancy also carry measurable risk. Valproate, a drug used to treat epilepsy, is the clearest example. Children exposed to valproate in the womb had roughly three times the risk of an autism spectrum diagnosis and five times the risk of childhood autism specifically, even after accounting for the mother’s epilepsy. Other common epilepsy medications did not show this increased risk, which suggests the effect is specific to valproate rather than epilepsy itself.
Parental Age and Autism Risk
Older parents are somewhat more likely to have autistic children. A meta-analysis covering nearly 30 studies found that the oldest categories of mothers and fathers had approximately 40% and 50% higher risk, respectively, compared to younger parents. The connection to paternal age likely reflects the fact that sperm cells accumulate new genetic mutations over a man’s lifetime. Each year adds roughly two new mutations to the DNA a father passes on, and some of these spontaneous changes affect genes involved in brain development.
Interestingly, one study of families already at high genetic risk for autism found that younger fathers (under 30) actually had higher odds of having an autistic child than fathers in their early thirties. This suggests the relationship between age and autism risk may be more complex in families with existing genetic susceptibility.
Epigenetics: Where Genes Meet Environment
Genes and environment don’t operate in separate lanes. Epigenetics describes how environmental exposures can change the way genes are read without altering the DNA itself. Think of it like highlighting certain pages of an instruction manual while crossing out others. Prenatal stress, infections, smoking, and nutritional deficiencies can all leave these marks on a developing baby’s DNA, turning genes up or down in ways that affect brain development.
Researchers have found that both valproate exposure and maternal immune activation can alter epigenetic marks on specific genes involved in brain function. One gene of particular interest, called MECP2, showed changed methylation patterns (a key epigenetic mark) in mouse offspring after maternal immune activation. Mutations in this same gene cause Rett syndrome, a condition that shares features with autism, which underscores how tightly linked these epigenetic changes are to neurodevelopmental outcomes.
Folate metabolism is another area where epigenetics, environment, and autism risk intersect. Prenatal folic acid supplementation is associated with a 30% reduced risk of autism in offspring, and prenatal multivitamins with a 34% reduction. These nutrients play a direct role in the chemical reactions that place epigenetic marks on DNA, which may partly explain their protective effect. This is one of the few modifiable factors with strong, consistent evidence behind it.
What Does Not Cause Autism
Parenting style does not cause autism. The “refrigerator mother” theory, which blamed cold or emotionally distant mothers, was thoroughly discredited decades ago. Autism is a neurodevelopmental condition rooted in how the brain forms before and shortly after birth, not in how a child is raised.
Screen time, diet after birth, and the number of toys in a nursery have no causal relationship with autism. These factors may influence behavior in any child, but they do not cause the underlying neurological differences that define autism.
Most Cases Have No Single Explanation
For the vast majority of autistic children, there is no identifiable single cause. Instead, a unique combination of inherited gene variants, spontaneous mutations, and prenatal environmental factors converge. Two autistic children can have entirely different genetic profiles and prenatal histories. This is part of why the condition is called a “spectrum,” with wide variation in traits, strengths, and challenges from one person to the next.
What is clear is that autism begins before birth, shaped primarily by the genetic blueprint a child inherits and fine-tuned by the biological environment of pregnancy. For expecting parents, the most evidence-backed step is ensuring adequate prenatal nutrition, including folic acid, and working with a healthcare provider to review any medications that might affect fetal development.