Autism arises from a combination of genetic predisposition and environmental influences during critical windows of brain development. No single cause has been identified. Instead, researchers have found that dozens of genes, prenatal exposures, and differences in how the brain wires itself during early life all play interconnected roles. Current CDC data shows autism is diagnosed in about 1 in 31 eight-year-olds in the United States, making it one of the most common neurodevelopmental conditions.
Genetics Carry the Largest Share of Risk
Twin studies provide the clearest picture of how much genetics matter. When one identical twin is autistic, the other twin is also on the spectrum in 60 to 90 percent of cases. For fraternal twins and non-twin siblings, that rate drops to around 20 percent. Because identical twins share virtually all their DNA, that high concordance rate points to genetics as the dominant factor. But the fact that it doesn’t reach 100 percent tells researchers something else is involved too.
The genetic picture is complex. Hundreds of genes have been linked to autism, and most of them are involved in how brain cells communicate with each other. Some of the best-studied genes produce proteins that help neurons form and maintain connections at synapses, the tiny gaps where one nerve cell signals to another. Mutations in genes like SHANK3 affect scaffold proteins that hold synaptic structures together, while changes in genes for adhesion molecules like neurexin and neuroligin alter how neurons physically link up. Many of these mutations are “de novo,” meaning they appear for the first time in the child rather than being inherited from a parent. Others are inherited variants that individually carry small risk but may combine to push development in a different direction.
Importantly, most of these genetic changes don’t work like an on/off switch. A single mutation in the SHANK3 gene, for example, doesn’t always cause a loss of function. Some mutations actually increase the activity of specific protein interactions, creating a gain of function that changes synaptic behavior in unpredictable ways. This helps explain why autism presents so differently from one person to the next.
Prenatal Environment Shapes Early Brain Development
What happens during pregnancy can influence whether genetic susceptibility leads to autism. Several maternal health conditions have been associated with higher likelihood of an autism diagnosis in the child. These include gestational diabetes, infections during pregnancy (such as rubella or cytomegalovirus), and prolonged fever or significant inflammation. Maternal depression and severe emotional stress during pregnancy also appear to play a role, likely through the effects of stress hormones on fetal brain development.
Certain medications taken during pregnancy carry risk as well. Valproic acid, an anti-seizure drug, has one of the strongest associations. Thalidomide and possibly misoprostol have also been linked to autism when used during pregnancy. Heavy alcohol use, cocaine exposure, heavy smoking, and folic acid deficiency round out the list of substances that researchers have connected to increased risk.
Air Pollution and Environmental Toxins
Fine particulate air pollution, the tiny particles known as PM2.5 that come from vehicle exhaust, industrial emissions, and wildfires, has emerged as a notable environmental risk factor. Research from Harvard’s School of Public Health found that exposure to an additional 10 micrograms of PM2.5 per cubic meter of air during early childhood increased autism risk by 64 percent. Prenatal exposure raised risk by 31 percent, with the third trimester being the most sensitive window. These effects appeared even at pollution levels below current regulatory limits.
Heavy metals and pesticide exposure during pregnancy have also been associated with increased risk, though the evidence is less precise about which specific compounds matter most and at what doses.
How the Brain Wires Itself Differently
Regardless of the initial trigger, autism involves differences in how the brain builds and refines its neural circuits during early life. One key process is synaptic pruning, the brain’s way of trimming excess connections to make the remaining circuits faster and more efficient. Every baby is born with far more synapses than they’ll ultimately need. The brain then selectively eliminates the ones that aren’t being used, following a “use it or lose it” principle.
This pruning happens in three major waves. The first occurs between birth and age two. The second takes place during adolescence, roughly ages 11 to 15. The third focuses on the prefrontal cortex, the brain’s planning and decision-making center, and continues into the mid-twenties. Research has found evidence of brain overgrowth in the first year of life in autistic children, suggesting that the earliest phase of pruning may proceed differently. Specialized immune cells in the brain called microglia carry out the physical work of pruning by engulfing tagged synapses. If this tagging or removal process is disrupted, the brain may retain more connections than typical, potentially contributing to the sensory sensitivity and different information processing patterns seen in autism.
Epigenetics: Where Genes and Environment Meet
Epigenetics offers one of the most compelling explanations for how environmental exposures translate into changes in brain development. Your DNA sequence is fixed, but chemical tags attached to your genes can dial their activity up or down without changing the underlying code. The most studied of these tags is DNA methylation, a process that can silence or activate genes depending on where it occurs.
Researchers have identified distinct DNA methylation patterns that can distinguish autistic individuals from non-autistic controls across multiple tissue types. These patterns frequently appear on genes that are important for neurodevelopment. Some of the genes affected, like MECP2 and GABRB3, have been independently implicated in autism through genetic studies, suggesting that epigenetic changes converge on the same biological pathways as inherited mutations.
What makes epigenetics particularly interesting is that some of these chemical marks may be passed across generations. Certain regions of the genome resist the normal “reprogramming” that happens when sperm and egg cells form, meaning methylation patterns from a parent’s environmental exposures could potentially influence their child’s gene expression. This is still an area of active investigation, but it provides a plausible mechanism for how a parent’s exposure to toxins or nutritional deficiencies could affect their child’s neurodevelopment even before conception.
Birth Complications Are Not a Standalone Cause
Preterm birth and delivery complications have long been discussed as potential contributors to autism, but the picture is more nuanced than it first appears. A large Israeli study analyzing nearly 115,000 deliveries over 12 years initially found higher autism rates among extremely preterm infants (1.6 percent) compared to full-term babies (0.7 percent). However, once researchers adjusted for other factors like ethnicity, maternal age, infant sex, and whether the baby was small for gestational age, the association between preterm birth and autism was no longer statistically significant.
The researchers concluded that a single obstetric factor is unlikely to cause autism on its own. A more plausible explanation involves multiple factors occurring simultaneously. In other words, the same genetic or environmental conditions that contribute to preterm birth may also independently contribute to autism, rather than prematurity itself being the cause.
Vaccines Do Not Contribute to Autism
A 2025 analysis by the World Health Organization reviewed 31 studies published between 2010 and 2025, drawing on data from multiple countries, and confirmed no causal link between vaccines and autism. This included specific analysis of aluminum adjuvants used in some vaccines and thimerosal, the mercury-containing preservative that sparked the original concern. A large Danish cohort study covering all children born between 1997 and 2018 was among the evidence reviewed. The WHO’s conclusion reinforced findings it had previously reached in 2002, 2004, and 2012. This is one of the most thoroughly investigated questions in modern medicine, and the answer has been consistent every time: vaccines do not cause autism.