Autism spectrum disorder (ASD) has no single cause. It arises from a combination of genetic predisposition and environmental influences during prenatal development, with genetics accounting for roughly 83% of the risk. The remaining portion comes from non-inherited factors, primarily experiences in the womb that shape how the fetal brain develops.
Genetics Play the Largest Role
A large population-based study published in JAMA estimated ASD heritability at 83%, meaning the vast majority of what determines whether someone develops autism comes down to their genes. Twin studies put the number even higher, around 87%. The remaining 17% is attributed to non-shared environmental influences, meaning experiences unique to one sibling but not another.
There is no single “autism gene.” Hundreds of genetic variants contribute to ASD risk, and they interact in complex ways. Some are common variants that individually have small effects but add up across the genome. Others are rare, spontaneous mutations (called de novo mutations) that arise for the first time in the child rather than being inherited from either parent. This genetic complexity is a major reason autism presents so differently from person to person.
What Happens in the Developing Brain
One of the clearest biological differences in autism involves how the brain prunes its connections during early childhood. In typical development, babies are born with an enormous surplus of synapses (the junctions between brain cells), and the brain gradually eliminates the ones it doesn’t need. This pruning process sharpens the brain’s ability to distinguish meaningful signals from background noise.
In autistic individuals, there is a dramatic increase in the growth of neural connections during the first year of life, while the pruning process that follows is significantly reduced. By late childhood, typically developing brains have pruned away about 50% of their synaptic connections. In autistic brains, that number drops to roughly 16%. The result is an excess of short-range connections, particularly in areas involved in social processing and executive function. This overabundance can effectively scramble long-distance communication signals in the brain, which may help explain the wide range of sensory, social, and cognitive differences seen across the spectrum.
The cells responsible for pruning are a type of immune cell in the brain called microglia. When these cells don’t function properly, or when chemical signals that guide them are disrupted, the pruning process stalls. This disruption can stem from the genetic variants described above, from prenatal environmental exposures, or from both working together.
Prenatal Environment and Maternal Health
The womb environment during key stages of brain development can meaningfully shift ASD risk. Infection during pregnancy, gestational diabetes, and maternal obesity are all established risk factors. The connecting thread appears to be the maternal immune response: when a pregnant person’s immune system activates strongly, the resulting flood of inflammatory signaling molecules can cross the placenta and interfere with fetal brain development. These molecules play a role in every stage of how the fetal brain wires itself, so disruptions during critical windows can have lasting effects.
Animal studies have confirmed that maternal immune activation alone, even without the infection itself reaching the fetus, produces long-lasting changes in brain chemistry and behavior in offspring. The inflammatory response also alters how the placenta processes serotonin, a chemical messenger that helps guide early brain wiring, increasing the amount delivered to the fetal brain at the wrong time.
Air pollution is another prenatal exposure with growing evidence behind it. A meta-analysis of 28 studies found that maternal exposure to air pollution was associated with a 1.4-fold increase in ASD risk. Several large-scale studies reported even higher numbers, with prenatal air pollution exposure linked to a 2.2 to 3.6 times increased risk depending on the type and timing of exposure.
Parental Age
Older parental age at conception is associated with a modest increase in autism risk. A meta-analysis of nearly 30 studies found that the oldest categories of maternal age carried about a 40% increased risk, while the oldest paternal age groups carried roughly a 50% increased risk, compared to parents in their early 30s. The likely explanation for paternal age is straightforward: sperm cells accumulate new genetic mutations with every division, and older fathers have had many more divisions. Maternal age effects may relate more to changes in egg quality and epigenetic regulation.
Premature Birth and Low Birth Weight
Babies born early face a higher likelihood of an ASD diagnosis, and the earlier they arrive, the greater the risk. Among babies born extremely preterm (22 to 27 weeks), 6.1% are later diagnosed with autism, compared to 1.4% of babies born at full term (39 to 41 weeks). Each additional week of gestation is associated with a 5% lower prevalence of ASD on average. Being born small for gestational age also increases risk independently, though prematurity itself is the stronger factor. These associations likely reflect both the underlying conditions that caused the early birth and the vulnerability of a brain that must finish critical development outside the womb.
How Genes and Environment Interact
Epigenetics offers one of the clearest explanations for how environmental factors translate into changes in brain development without altering a person’s DNA. Epigenetic changes are chemical modifications that sit on top of DNA and control which genes get turned on or off. The most studied type is DNA methylation, where small chemical groups attach to DNA and silence specific genes.
Research has found distinct methylation patterns in the brains of autistic individuals. In the frontal cortex, proteins that add and remove these chemical tags are present at abnormal levels, which appears to reduce the activity of genes involved in brain cell signaling and organization. Placental tissue from pregnancies that resulted in an ASD diagnosis also shows significantly different methylation patterns, suggesting these changes begin before birth.
Maternal lifestyle factors like smoking, alcohol use, obesity, and poor nutrition can all drive epigenetic changes during pregnancy. Even maternal asthma has been linked to altered methylation of immune-related genes in fetal blood. Additionally, variations in a gene involved in processing folate (a B vitamin critical for DNA methylation) are associated with increased ASD risk, which is one reason adequate folate intake during pregnancy is considered important for neurodevelopment.
Vaccines Do Not Cause Autism
Twenty epidemiological studies conducted across multiple countries have found no link between vaccines and autism. These include studies of the MMR vaccine and studies of thimerosal, a preservative once used in some vaccines. The populations studied have been large enough to detect even rare associations, and none has been found. A Finnish study tracking 1.8 million vaccinated children identified zero vaccine-associated cases of autism. A Danish national registry study of 537,303 children found no difference in autism rates between those who received the MMR vaccine and those who did not. In Canada, researchers actually found that autism rates increased during a period when MMR vaccination rates decreased. The original 1998 paper that proposed a link has been retracted due to serious ethical violations and methodological fraud.