Autism begins before birth. The earliest cellular signs appear during the first or second trimester of pregnancy, when the brain is rapidly building its foundational architecture. There is no single cause. Instead, autism results from a convergence of genetic factors, prenatal conditions, and differences in how the brain wires itself during development. About 1 in 31 children in the United States are now identified with autism, according to the CDC’s most recent data from 2022.
It Starts With Genes, but Not One Gene
Autism has a strong genetic component, but the genetics are complex. There is no “autism gene.” Instead, the genetic architecture involves a wide range of variants: spontaneous mutations that appear for the first time in a child (not inherited from either parent), rare inherited variants passed down through families, and structural changes in chromosomes where large segments of DNA are duplicated or deleted. Whole genome sequencing has also revealed contributions from mutations in non-coding regions of DNA (the vast stretches that don’t directly produce proteins but regulate how other genes behave), as well as repetitive DNA sequences called tandem repeats.
What this means practically is that two autistic people can have entirely different genetic profiles. Some cases involve a handful of high-impact mutations, while others likely result from the combined effect of many small genetic variations, each contributing a modest amount of risk. This diversity in genetic causes helps explain why autism presents so differently from person to person.
The Brain Builds Differently Before Birth
Between 10 and 20 weeks of gestation, the brain undergoes an explosive period of neuron production. Cortical neurons multiply exponentially during this window and are not produced again after birth. Research on postmortem brain tissue has found that some autistic individuals have an excess of neurons in certain brain regions, which points to disruption during this critical prenatal phase.
About 68% of high-confidence autism risk genes are active during the first through third trimesters, when they influence cell production, the migration of new neurons to their correct positions, and the formation of the brain’s layered structure. Studies have found focal patches of disorganized cortical layers and clusters of misplaced neurons in the prefrontal and temporal cortices of autistic brains. These are regions responsible for social reasoning, language, and decision-making. The remaining 32% of risk genes are active in the third trimester and early postnatal life, when the brain shifts to building connections between neurons.
Synaptic Pruning and the “Noisy Network” Effect
After birth, a typically developing brain goes through a process of aggressive pruning. Think of it like sculpting: the brain starts by growing far more connections between neurons than it needs, then trims away the ones that aren’t being used. By late childhood, the density of synaptic connections in a neurotypical brain drops by about 50%. In autistic brains, that reduction is closer to 16%.
This leaves a large number of redundant, short-range connections intact. The result is a kind of signal-to-noise problem. With too many overlapping local connections, distinct signals from neighboring brain areas compete and interfere with each other, creating “noise” within the network. At the same time, long-distance connections between distant brain regions develop more weakly. The abundance of short-range wiring can effectively scramble long-distance communication signals.
This pattern of local over-connectivity and long-distance under-connectivity has practical consequences. It may explain difficulties with shifting attention, the tendency toward repetitive behaviors, and challenges processing social and emotional signals that require coordination between widely separated brain areas. Disrupted connectivity between the prefrontal cortex and the amygdala (a structure central to processing emotions and social cues) has been linked to sensory hypersensitivity, difficulties with social interaction, and differences in emotional regulation.
Prenatal Environment Plays a Role
Genetics load the gun, but the prenatal environment can pull the trigger, or at least nudge the trajectory. Several maternal factors during pregnancy are associated with increased autism risk: advanced parental age, obesity, gestational diabetes, high blood pressure, and infections during pregnancy. Certain medications taken during pregnancy, particularly the anti-seizure drug valproate, have also been linked to higher risk.
The biological mechanisms connecting these factors to brain development appear to involve inflammation, disruptions to cellular energy production, and oxidative stress. Maternal immune activation is one of the more studied pathways. When the mother’s immune system is significantly activated during pregnancy, whether from infection, chronic asthma, or other causes, it can influence fetal brain development. Animal studies have shown that maternal immune activation produces autism-like behaviors in offspring, and these effects can be traced to changes in how genes are expressed in the developing brain.
Epigenetics: When the Environment Changes Gene Behavior
Epigenetics is the study of how gene activity can be turned up or down without altering the DNA sequence itself. Think of DNA as a piano: epigenetics determines which keys get played and which stay silent. Several mechanisms do this, including chemical tags that attach to DNA or to the protein spools that DNA wraps around, making genes more or less accessible.
Maternal lifestyle factors, including smoking, alcohol use, obesity, and poor nutrition, can influence these epigenetic settings during fetal development. Folate metabolism is one specific pathway of interest. A common genetic variation in the enzyme that processes folate is associated with increased autism risk, because this enzyme plays a central role in the chemical tagging of DNA during neural development. Animal studies have found that protein-restricted or folate-supplemented diets in mothers can produce autism-like traits in offspring through epigenetic changes.
Valproate exposure in rodents produces autism-like behaviors that persist and can be epigenetically transmitted to at least the third generation, meaning the original environmental exposure creates heritable changes in gene expression without touching the DNA code. Small regulatory molecules called microRNAs, which fine-tune gene activity after the fact, have also been implicated in autism. These molecules can silence or destabilize the messenger molecules that carry genetic instructions, altering how much of a given protein gets made.
The Gut-Brain Connection
Gastrointestinal problems are common in autistic individuals, and a growing body of research connects gut bacteria to brain function through what’s known as the gut-brain axis. Gut microbes produce short-chain fatty acids, vitamins, and compounds that act as neurotransmitters, including serotonin and GABA. These chemicals can communicate directly with the brain through the vagus nerve.
Some autistic individuals show differences in the composition of their gut bacteria compared to neurotypical peers. A compromised intestinal barrier, sometimes called “leaky gut,” may allow bacterial metabolites and dietary products to cross into the bloodstream, triggering immune responses that affect the central nervous system. In animal models of autism, introducing specific beneficial bacteria (like Lactobacillus reuteri) reversed social behavior deficits and restored oxytocin levels in the brain. Another probiotic bacterium, Bacteroides fragilis, reversed some autism-like behaviors in mice, improved gut barrier function, and normalized blood levels of bacterial metabolites.
In humans, microbiota transfer therapy has shown sustained improvements in both GI symptoms and autism-related behaviors lasting at least two years after a 10-week treatment period. This research is still relatively young, but it suggests the gut microbiome is a meaningful piece of the puzzle, not just a bystander.
How All of This Comes Together
Autism doesn’t happen because of one thing going wrong. It emerges from a cascade that typically begins with genetic susceptibility, gets shaped by prenatal conditions and epigenetic changes, and plays out through differences in how the brain builds and prunes its connections. The process starts in the womb, often before any outward signs are visible, and continues through the early years of life as the brain refines its networks.
This is why autism looks so different from one person to the next. The specific combination of genetic variants, the timing of prenatal disruptions, the degree of pruning differences, and even the composition of gut bacteria all vary. Two children can both meet the diagnostic criteria, which require persistent differences in social communication plus restricted or repetitive behaviors present from early development, and yet have arrived there through substantially different biological pathways. The wide spectrum of autism isn’t just a clinical observation. It reflects genuinely different underlying biology.