Autism spectrum disorder arises from a convergence of genetic predisposition and environmental exposures during fetal and early development, not from any single cause. Twin studies place the heritability of autism somewhere between 64% and 91%, making genes the dominant factor, but the remaining risk comes from prenatal and perinatal conditions that interact with that genetic background in ways researchers are still untangling. The picture that has emerged over the past two decades is one of extraordinary complexity, with hundreds of genes, dozens of environmental suspects, and biological pathways that connect them in ways no one fully predicted.
Genes Are the Biggest Single Factor
The strongest evidence for a genetic basis comes from twin studies. A meta-analysis pooling data from multiple twin cohorts estimated autism’s heritability at 64% to 91%, with shared environmental factors accounting for anywhere from 7% to 35% depending on how broadly autism was defined.1PubMed Central. Heritability of autism spectrum disorders: a meta‐analysis of twin studies That range is wide, but even at the lower end, genetics plays a larger role than any other single category of risk.
The genetic architecture is not simple. There is no “autism gene.” Instead, risk accumulates from two broad classes of genetic variation acting together. One class involves rare mutations, many of which appear for the first time in the affected child rather than being inherited from a parent. These so-called de novo mutations include deletions, duplications, and rearrangements of stretches of DNA, and they are highly varied from one person to the next.2PubMed Central. Strong association of de novo copy number mutations with autism The other class involves common genetic variants, each contributing a tiny sliver of risk on its own. Research on families with multiple autistic children has confirmed that these common variants add to risk on top of rare mutations, rather than one type canceling the other out.3PubMed Central. Polygenic transmission disequilibrium confirms that common and rare variation act additively to create risk for autism spectrum disorders In multiplex families, both rare inherited variants and broad polygenic risk contribute, consistent with an additive model where many small pushes combine into a large effect.4PubMed Central. The contributions of rare inherited and polygenic risk to ASD in multiplex families
This additive architecture helps explain why autism runs in families without following a neat inheritance pattern. Two parents can each carry a load of common risk variants and pass different combinations to different children, so one sibling may be autistic while another is not. A fresh de novo mutation on top of that inherited load can push a child over a threshold that a sibling narrowly stayed below.
The Prenatal Environment
Genes set the stage, but the conditions a fetus develops in can shift the outcome. Several prenatal exposures have been linked to increased autism risk, though none of them are anywhere close to “causing” autism on their own.
Maternal immune activation is one of the more biologically plausible environmental pathways. When a pregnant person’s immune system mounts a strong inflammatory response, the resulting surge of signaling molecules can alter fetal brain development. Animal models have shown that triggering an immune storm during pregnancy produces offspring with behaviors and brain changes that parallel autism.5PubMed Central. Cytokine changes associated with the maternal immune activation (MIA) model of autism: A penalized regression approach In humans, the connection is harder to nail down, but severe infections during pregnancy do appear to modestly raise autism risk. A related line of research has identified maternal antibodies that react against proteins in the developing fetal brain, and the presence of specific combinations of these antibodies has been associated with a diagnosis of what researchers call maternal autoantibody-related autism.6PubMed Central. Maternal Anti-Fetal Brain IgG Autoantibodies and Autism Spectrum Disorder: Current Knowledge and its Implications for Potential Therapeutics
Certain medications taken during pregnancy also matter. Valproate, an anti-seizure drug sometimes used for epilepsy or bipolar disorder, carries the clearest risk. In a large Danish cohort, children exposed to valproate in utero had roughly 2.7 times the risk of autism compared to unexposed children, even after adjusting for the mother’s epilepsy.7PubMed Central. Risk of Autism after Prenatur Topiramate, Valproate, or Lamotrigine Exposure An earlier study from the same registry found a similar elevated risk, with an absolute rate of about 4.4% among valproate-exposed children.8PubMed Central. Prenatal valproate exposure and risk of autism spectrum disorders and childhood autism By contrast, other anti-seizure medications like lamotrigine and topiramate showed no meaningful increase in autism risk in the same analyses. This specificity is important: it is not “taking medication during pregnancy” that raises risk, but specific drugs with specific effects on fetal neurodevelopment.
Pesticide and air pollution exposures have also drawn attention. A review of the epidemiological evidence found associations between autism and traffic-related air pollutants, certain metals, and several pesticides.9PubMed Central. Environmental chemical exposures and autism spectrum disorders: a review of the epidemiological evidence A population-based study in California found that prenatal proximity to applications of several common agricultural pesticides, including glyphosate and chlorpyrifos, was associated with modestly elevated autism risk, with odds ratios generally in the range of 1.10 to 1.16.10BMJ. Prenatal and infant exposure to ambient pesticides and autism spectrum disorder in children: population based case-control study A systematic review reached a similar conclusion, reporting high rates of association between early pesticide exposure and autism across multiple study designs.11Revista Paulista de Pediatria. Early exposure to agricultural pesticides and the occurrence of autism spectrum disorder: a systematic review These effect sizes are small individually, but in populations with widespread exposure they could matter at a public-health level.
Paternal Age and Preeclampsia
Older fathers have long been flagged as a risk factor for autism, with the assumption being that sperm accumulate new mutations over time. The reality is more nuanced. An analysis combining genetic sequencing data with Danish registry records found that the new mutations arising from older fatherhood do confer a small amount of autism risk, but the broader epidemiological link between delayed paternity and autism likely reflects other factors that do not necessarily increase with age.12Nature Communications. Paternal-age-related de novo mutations and risk for five disorders In other words, men who become fathers later may carry traits or genetic loads that independently raise autism risk in their children, separate from the accumulation of fresh mutations.
Pregnancy complications also play a role. Preeclampsia, a condition involving high blood pressure and organ stress during pregnancy, has been linked to increased autism risk in offspring, with early-onset and severe cases appearing to carry the strongest association.13PubMed. Preeclampsia: The silent catalyst for autism spectrum disorder in offspring The proposed mechanism involves reduced blood flow to the placenta, which could deprive the fetal brain of oxygen and nutrients during critical developmental windows. This is one of several perinatal complications, including preterm birth and low birth weight, that epidemiological studies have repeatedly linked to autism, though none of them are strong enough risk factors on their own to be predictive.
What Happens in the Developing Brain
Knowing that genes and prenatal exposures contribute to autism is one thing. Understanding what actually goes wrong in the brain is another, and the answer varies depending on which genes and which exposures are involved.
One line of research points to problems with synaptic pruning, the process by which the brain trims excess connections during development. In mice and human brain organoids carrying a mutation in a gene called SCN2A (one of the genes linked to autism), the brain’s immune cells, called microglia, were found to be overly aggressive in eliminating synapses. This excessive pruning reduced the density of connection points between neurons and impaired learning and memory. Blocking microglial activity partially reversed these effects.14PubMed. Microglial over-pruning of synapses during development in autism-associated SCN2A-deficient mice and human cerebral organoids
At a larger scale, some autistic individuals show differences in how brain regions communicate. Research on the default mode network, a set of brain regions active during rest and self-referential thought, has found atypical connectivity patterns in autism. These disruptions appear to be rooted in imbalances between excitatory and inhibitory signaling and may contribute to differences in social attention and the ability to shift focus toward socially relevant information.15PubMed Central. The Default Mode Network in Autism A notable subgroup of autistic individuals also shows brain overgrowth, with an enlarged head and brain volume, particularly in early childhood, and this subgroup tends to present with more pronounced symptoms.16PubMed Central. Mechanisms of brain overgrowth in autism spectrum disorder with macrocephaly
Epigenetics as a Bridge
If genes are the blueprint and prenatal exposures are the construction conditions, epigenetics is one mechanism by which those conditions alter how the blueprint gets read. Epigenetic changes do not alter the DNA sequence itself but affect which genes are turned on or off at particular times. Preliminary evidence suggests that differences in one type of epigenetic mark, DNA methylation, contribute to autism risk.17PubMed. Leveraging epigenetics to examine differences in developmental trajectories of social attention: A proof-of-principle study of DNA methylation in infants with older siblings with autism
Genome-wide studies of brain tissue from autistic individuals have found that methylation changes tend to cluster around genes involved in neurodevelopment and occur in regions of the genome that are regulated by specific types of chemical markers on chromosomes. The pattern suggests an epigenetic delay in the normal developmental trajectory of the brain, as if certain genes that should have been switched on or off at particular stages were on a different schedule. This altered timing was associated with disruptions in how genetic instructions get processed into proteins, affecting genes already known to be relevant to autism.18PubMed Central. Epigenetic Delay in the Neurodevelopmental Trajectory of DNA Methylation States in Autism Spectrum Disorders Epigenetics offers a plausible way for prenatal environmental exposures like immune activation, medications, or toxicants to leave lasting marks on gene expression without changing the underlying DNA.
Mitochondria and Energy Production
A strand of research that gets less public attention involves the cell’s energy-producing machinery, the mitochondria. A systematic review and meta-analysis confirmed that a subset of autistic individuals show signs of mitochondrial dysfunction, and that this dysfunction is linked to other common physiological features of autism, including gastrointestinal problems, oxidative stress, and immune irregularities.19PubMed. Biomarkers of mitochondrial dysfunction in autism spectrum disorder: A systematic review and meta-analysis Studies have also reported reduced activity of the molecular machinery that mitochondria use to generate energy, alongside reduced expression of genes encoding those components.20PubMed Central. Mitochondrial Dysfunction in Autism Spectrum Disorders
Whether mitochondrial dysfunction is a cause, a consequence, or simply a co-traveler with autism is not settled. The developing brain is one of the most energy-hungry organs in the body, so it makes biological sense that even modest impairments in energy production could disproportionately affect brain development. But it remains unclear whether these mitochondrial problems are driven by the same genetic variants that create autism risk or whether they represent a separate vulnerability that compounds the effect.
The Gut-Brain Axis
Gastrointestinal symptoms are common among autistic individuals, and over the past decade, researchers have looked at whether differences in gut bacteria play a role. Studies have found that autistic children tend to have an altered gut microbial profile, including reduced levels of bacteria that produce butyrate, a short-chain fatty acid that helps maintain the gut lining and appears to influence brain function through signaling pathways connecting the gut and the brain.21PubMed Central. Altered gut microbiota and short chain fatty acids in Chinese children with autism spectrum disorder At the same time, elevated levels of another short-chain fatty acid, propionic acid, have been observed in some autistic individuals, and comprehensive reviews have documented a significant shift in overall gut composition in people with autism.22PubMed Central. The Role of Short-Chain Fatty Acids and Altered Microbiota Composition in Autism Spectrum Disorder: A Comprehensive Literature Review
The promise here is tantalizing but the science is still early. Animal experiments have shown that restoring butyrate levels can alter social behavior in autism mouse models, and the proposed mechanisms by which gut bacteria might affect the brain, through immune modulation, gut barrier integrity, and direct chemical signaling, are biologically plausible.23Scientific Reports. Altered gut microbiota and short chain fatty acids in Chinese children with autism spectrum disorder But human evidence for a causal role remains thin. Correlation between gut differences and autism does not tell you which came first, and dietary differences, medication use, and stress, all of which are common in autism, can independently reshape the gut microbiome.
Why Males Are Diagnosed More Often
Autism is diagnosed roughly three to four times more often in males than in females. One explanation that has gained traction is a “female protective effect,” meaning that females may require a higher genetic burden to cross the threshold into autism. Research has confirmed that autism risk genes interact with sex-related biological processes, contributing to the male bias in prevalence and suggesting a neurobiological basis for this protective effect.24PubMed Central. Impact of autism genetic risk on brain connectivity: a mechanism for the female protective effect Efforts to find a single genetic location responsible for this protection have come up empty, indicating that it is not driven by one gene but is instead a distributed effect across many genetic and hormonal pathways.25PubMed Central. The female protective effect in autism spectrum disorder is not mediated by a single genetic locus
There is also a diagnostic dimension. Autistic females more often present with subtler social difficulties or develop compensatory strategies that mask their traits, leading to later diagnosis or missed diagnosis altogether. The sex ratio has narrowed somewhat as clinicians have become more aware of how autism presents differently in girls and women.
Why Autism Rates Have Risen
The prevalence of autism has climbed dramatically over the past few decades, a trend that feeds public concern about environmental causes. While genuine increases in risk from environmental exposures cannot be ruled out, a significant portion of the rise is explained by changes in how autism is diagnosed and recognized. An analysis of California’s caseload found that about a quarter of the increase between 1992 and 2005 was attributable solely to diagnostic substitution, where individuals who would previously have received a diagnosis of intellectual disability were reclassified as autistic.26PubMed Central. Diagnostic change and the increased prevalence of autism A twenty-year study in the United Kingdom reached a similar conclusion, finding that rising diagnosis rates, especially among adults, females, and higher-functioning individuals, likely reflect increased recognition and broadened diagnostic criteria rather than a true surge in the underlying condition.27PubMed. Time trends in autism diagnosis over 20 years: a UK population-based cohort study
Vaccines Do Not Cause Autism
No discussion of autism’s origins is complete without addressing the myth that refuses to die. The claim that vaccines, or the mercury-containing preservative thimerosal once used in some vaccines, cause autism has been thoroughly investigated and rejected. Large-scale epidemiological studies involving millions of children have found no association. The original 1998 paper that sparked the scare was retracted by the journal that published it, and its author lost his medical license. As one review summarized, the myths linking vaccines or mercury to autism were amplified by flawed science, organized parent advocacy groups, and political pressure, but the scientific evidence has consistently shown no connection.28PubMed Central. Vaccination as a cause of autism-myths and controversies. The persistence of this myth has had real consequences, contributing to declining vaccination rates and outbreaks of preventable diseases, while diverting research attention and public trust away from the actual causes being studied.
Autism-Linked Genes and Cognitive Tradeoffs
An intriguing strand of research asks why so many autism-associated genetic variants persist in the human population if they carry developmental costs. One study used genomic analysis to identify signatures of natural selection acting on autism-linked genes and found that many of these genes are involved in cognition, learning, memory, and nervous system signaling.29Scientific Reports. Genomic selection signatures in autism spectrum disorder identifies cognitive genomic tradeoff and its relevance in paradoxical phenotypes of deficits versus potentialities The implication is a cognitive tradeoff: the same genetic variants that in certain combinations contribute to autism may, in other combinations or at lower doses, be associated with cognitive strengths. This could help explain why autism-linked genes have not been weeded out by evolution and why many autistic individuals show areas of exceptional ability alongside their challenges. The idea remains speculative at the population level, but it reframes autism-associated genetics as part of normal human cognitive variation rather than purely as pathology.