The Messy Evolution of Diagnosing a Moving Target
Autism was never a monolith, though we spent the better part of the 20th century pretending it was. Go back to 1943, when Leo Kanner first described "infantile autism" in Baltimore, or look at Hans Asperger’s work in Vienna; they were looking at vastly different expressions of what we now throw into one massive clinical bucket. The thing is, our definition keeps stretching. It is a biological puzzle wrapped in an administrative riddle. When the American Psychiatric Association published the DSM-5 in 2013, they collapsed separate diagnoses like Asperger’s Disorder and Pervasive Developmental Disorder-Not Otherwise Specified (PDD-NOS) into a single spectrum. That changes everything because you cannot hunt for a single cause when the diagnostic criteria themselves are fluid.
The Problem With Clinical Labels
Honestly, it’s unclear where the boundaries truly lie. Because two children with the exact same ASD diagnosis can share zero overlapping symptoms—one might be entirely non-verbal with severe intellectual disabilities, while the other is a hyper-lexic tech prodigy with minor social anxiety—looking for a uniform etiology is like trying to find the single cause of "coughing." Experts disagree constantly on where the behavioral ends and the biological begins. We are dealing with an incredibly diverse set of neurodivergent traits, which explains why the hunt for what are the three main causes of autism has forced researchers to stop looking for a single switch and start analyzing entire networks of synaptic pruning dysfunction.
The Heavyweight Champion: Inherited Genetic Architecture
When you look at the raw data, genetics does not just tip the scales; it practically breaks them. Twin studies—the gold standard for parsing nature from nurture—have consistently shown that if one identical twin has autism, the co-twin has an estimated concordance rate of 60% to 90%. That is an astronomical number in behavioral genetics. Yet, people don't think about this enough: this does not mean there is an "autism gene" waiting to be snipped out via CRISPR. Instead, we are looking at a vast, swirling sea of common genetic variants. These are tiny variations in DNA, known as single nucleotide polymorphisms, that are present throughout the general population but happen to cluster together in specific, high-density patterns in autistic individuals.
Polygenic Risk and the Common Variant Myth
It is all about the cumulative load. A single common variant does next to nothing on its own, but when an individual inherits hundreds of these tiny genetic nudges from both parents, the threshold is crossed, altering cortical histogenesis during embryonic development. Think of it like a downpour; one drop of rain is harmless, but a cloudburst floods the basement. A massive 2019 study published in Cell analyzed over 37,000 genetic samples and revealed that these subtle, inherited variations actually account for the vast majority—around 50% of the total liability for autism. It is a slow, additive process that quietly shapes the architecture of the brain long before a child takes their first breath.
The Synaptic Connection
Where it gets tricky is figuring out what these genes actually do. Many of them regulate NLGN3 and NLGN4X, which are cell adhesion molecules located at the synapses where neurons communicate. Because these specific genes are altered, the brain struggles to prune away unnecessary neural connections during early childhood. The result? A hyper-connected brain that gets easily overwhelmed by sensory data, leading to the intense sensory overloads so common in autistic individuals. But wait, if inherited genetics explains so much, why do we see families where autism seemingly appears out of nowhere? That brings us to our second major pillar.
Spontaneous Glitches: De Novo Mutations and Copy Number Variants
Not every genetic cause is passed down through generations. Sometimes, the blueprint gets smudged during the copy-paste process of conception itself. These are called de novo mutations—genetic alterations that appear for the very first time in a family member because of a glitch in the father’s sperm or the mother’s egg. And this is where the conventional wisdom about heredity falls apart completely. You can have two parents with absolutely zero familial history of neurodevelopmental conditions who give birth to a child with profound autism, simply because a specific chunk of DNA was deleted or duplicated during meiosis. These structural changes are known as Copy Number Variations (CNVs).
The Advanced Paternal Age Factor
And here is a piece of data that makes a lot of people uncomfortable: the age of the father matters significantly. A landmark Swedish study tracking over two million people discovered that children born to fathers over the age of 45 were 3.5 times more likely to be diagnosed with autism compared to those born to fathers in their early twenties. Why? Because while a biological female is born with all her eggs, a male’s germline cells divide continuously throughout his life—roughly 23 times a year. By the time a man reaches his late 40s, his sperm-producing cells have replicated hundreds of times, vastly increasing the statistical probability of a catastrophic copying error in the genetic code. It is simple replication mechanics, yet it is a variable we ignored for a generation while unfairly scrutinizing mothers.
High-Impact Genes and Syndromic Autism
These spontaneous mutations often hit heavy-duty genes. We are talking about genes like CHD8, which acts as a master regulator for chromatin remodeling. When a de novo mutation wrecks a gene like CHD8, it triggers a domino effect that disrupts hundreds of downstream genes, frequently resulting in what clinicians call syndromic autism. This type of autism often co-occurs with distinct physical traits, macrocephaly, or severe gastrointestinal issues. It is a completely different biological pathway than the polygenic inheritance we discussed earlier, proving that what are the three main causes of autism cannot be reduced to a single narrative about family trees.
The Womb Environment: Prenatal Triggers and Maternal Immune Activation
Genetics loads the gun, but the prenatal environment often pulls the trigger. This is not about bad parenting or cold, distant mothers—let us permanently bury that "refrigerator mother" myth from the 1950s. Instead, we are looking at the literal environment inside the uterus during critical gestational windows. The fetal brain is incredibly sensitive to external disruptions, and a growing body of epidemiological data shows that maternal immune activation (MIA) plays a massive role in tilting development toward an autistic phenotype.
When the Immune System Misfires
If a pregnant individual suffers from a severe viral or bacterial infection that requires hospitalization during the late first or second trimester, the risk profile shifts. It is not the virus itself crossing the placenta and attacking the fetus; rather, it is the mother’s own inflammatory response that does the damage. High levels of pro-inflammatory cytokines, specifically interleukin-6 (IL-6) and interleukin-17, flood the amniotic fluid. These signaling molecules cross the fetal blood-brain barrier and interfere with the migration of neural progenitor cells, which explains why the architectural layers of the autistic cortex often appear disorganized in post-mortem studies. A 2016 cohort study in Denmark confirmed this, linking severe maternal infections to a significant increase in ASD risk in offspring.
Metabolic and Chemical Stressors
But the gestational environment faces other pressures too. Maternal metabolic conditions, such as poorly managed gestational diabetes or severe obesity, alter the nutrient and hormone profiles available to the developing fetus, creating a state of chronic low-grade inflammation. Furthermore, prenatal exposure to specific medications—most notably valproic acid, an anticonvulsant used to treat epilepsy—has a direct, proven link to autism. In fact, children exposed to valproate in utero face an estimated 10% risk of developing ASD. This is a massive jump from the general population baseline. It is clear that the womb is not a sterile sensory deprivation tank; it is a dynamic, reactive environment where chemical signals can alter gene expression without changing the underlying DNA sequence through epigenetic mechanisms.
The Epigenetic Landscape: Nature Versus Nurture is a Dead Debate
To truly grasp what are the three main causes of autism, you have to realize that separating genetics from environment is a fool's errand. They are locked in a constant, intricate dance. This is the realm of epigenetics—chemical tags like methyl groups that attach to DNA and turn genes on or off based on external signals. You could inherit a high genetic vulnerability for autism, but if your gestation is entirely free of inflammatory or chemical stressors, those genes might remain silenced. Conversely, a moderate genetic risk combined with a severe mid-pregnancy influenza infection might cause those same genes to express aggressively. The issue remains that we are trying to use binary thinking on a multi-dimensional biological reality. In short, autism is the ultimate product of gene-environment interactions, where neither side can claim total dominance.