What the research actually says
A plain summary of where the evidence is strong, where it's genuinely uncertain, and where claims are being made that the science doesn't support.
Settled enough to rely on
Autism is substantially genetic
Twin and family studies put heritability estimates high — commonly in the 60–90% range. Hundreds of genes have been implicated, most with small individual effects, alongside rarer single-gene conditions like Fragile X. There is no single "autism gene" and no useful prenatal test.
Vaccines do not cause autism
This has been tested repeatedly in very large populations, including a Danish study of over 650,000 children, and the link consistently fails to appear. The original paper claiming it was fraudulent and retracted. Brain differences associated with autism are measurable before the age at which the vaccines in question are given.
Parenting does not cause autism
The refrigerator-mother theory has no empirical support and was abandoned by researchers decades ago.
Early support helps
Children who receive developmental support early tend to make greater gains in communication and daily living skills. The effect is real but often overstated in marketing — early intervention improves trajectories, it does not make a child non-autistic.
AAC doesn't suppress speech
The long-standing fear that giving a child a device or picture system will stop them from talking is not supported. Reviews find the opposite association.
Autistic people are at higher risk of harm
Higher rates of bullying, exclusion, unemployment, mental health conditions and suicidality — much of which appears to be driven by how autistic people are treated rather than by autism itself.
Genuinely open questions
How much of the rise is real?
Most researchers attribute the great majority of the increase to broader diagnostic criteria, better awareness, diagnostic substitution and improved identification in girls and in Black, Hispanic and Asian children. Whether some genuine increase sits underneath that is still debated, and the honest answer is that current data can't fully separate the two.
Environmental contributors
Advanced parental age, extreme prematurity, certain prenatal exposures and pregnancy complications show consistent but small associations. These are risk factors at a population level, not explanations for any individual child, and no single environmental cause has been established.
The gut–brain link
Differences in gut bacteria have been observed, but cause and effect are unresolved — restricted diets could just as easily explain the differences. Microbiome treatments remain experimental.
Which intervention, for which child
Comparative research is weaker than the marketing suggests. Many studies are small, short, unblinded, or measure adult-rated behaviour rather than the child's wellbeing. Almost no research follows outcomes into adulthood.
Claims to treat with real caution
If someone promises recovery or a cure, walk away. Families are a target market. Some of what's sold is merely expensive and useless; some of it is dangerous. Always run a proposed treatment past your child's doctor first.
Chelation
Drugs to strip "heavy metals." No benefit demonstrated; children have died.
MMS / "miracle mineral"
Industrial bleach, sometimes given as an enema. Causes serious injury. Subject of federal warnings and prosecutions.
Hyperbaric oxygen
Marketed widely for autism; evidence does not support it.
Stem cell "treatments"
Expensive unproven clinics, often overseas. Not an approved autism treatment anywhere.
Facilitated communication
Hand-over-hand support that research shows produces the facilitator's words. Independent typing and AAC are different and legitimate.
Restrictive diets and supplement stacks
Gluten- and casein-free diets show little effect in controlled trials and carry nutrition risks. Some supplements interact with medication.
Medication has a legitimate role — for anxiety, ADHD, sleep, seizures, severe irritability or self-injury — prescribed and monitored by a doctor. It treats those specific conditions, not autism itself.