Autism is not inherited exclusively from the mother or the father. Both parents contribute genetic risk, but they tend to do so through different biological channels. Fathers pass along more spontaneous new mutations, especially as they age, while mothers appear to transmit a greater share of the common inherited variants that collectively raise autism risk. The picture is further complicated by mitochondrial DNA, epigenetic marks, and the prenatal environment, each with its own parent-of-origin bias. Understanding these distinct pathways explains why the “mother or father” framing, though natural, oversimplifies a genuinely complex genetic story.
Autism Is Highly Heritable, Regardless of Which Parent
Twin studies consistently show that genes account for the majority of autism risk. A meta-analysis of twin research placed heritability estimates between 64% and 91%, depending on diagnostic criteria and how the studies were pooled.1PubMed Central. Heritability of autism spectrum disorders: a meta-analysis of twin studies A large population-based study across several countries narrowed the estimate to about 83%, with the remaining 17% attributable to unique environmental influences and essentially no role for the shared family environment.2JAMA. The Heritability of Autism Spectrum Disorder A UK twin sample found a similar range of 56% to 95% depending on how autism traits were measured.3JAMA Psychiatry. Heritability of Autism Spectrum Disorder in a UK Population-Based Twin Sample A Swedish registry study estimated additive genetic heritability at about 85%, with little or no contribution from maternal effects.4PubMed Central. Heritable Variation, With Little or No Maternal Effect, Accounts for Recurrence Risk to Autism Spectrum Disorder in Sweden
That last finding is worth pausing on. “Maternal effects” in genetics refers to influences from the mother’s own biology on the developing child, things beyond the DNA she passes on, like her uterine environment or her egg cytoplasm. The Swedish data suggest those non-genetic maternal influences are modest at best when it comes to overall autism liability. The bulk of the risk is genetic, and it flows through both parents’ chromosomes.
What Fathers Contribute: New Mutations and Aging
Sperm cells divide continuously throughout a man’s life, and each division introduces a small chance of copying errors. The result is that older fathers pass along more brand-new mutations, called de novo mutations, that were not present in either grandparent. One landmark Icelandic study found that the number of new mutations in a child’s genome increases by roughly two per year of the father’s age, with paternal mutations doubling approximately every 16.5 years.5PubMed Central. Rate of de novo mutations and the importance of father’s age to disease risk
These new mutations are relevant because some of them land in genes important for brain development. However, the actual increase in autism risk tied to paternal age is modest in absolute terms. A study modeling risk across five disorders found that offspring of a 45-year-old father had only about a 9% higher risk of autism compared with offspring of a 25-year-old father.6Nature Communications. Paternal-age-related de novo mutations and risk for five disorders That sounds small, and it is, because the baseline rate of autism is itself low. A 9% increase on a small number is still a small number. The paternal-age effect is real and well-documented, but it does not mean older fathers are likely to have autistic children. It means the per-child odds tick upward slightly with each decade.
What Mothers Contribute: Polygenic Risk and the Female Protective Effect
While fathers dominate the de novo mutation story, mothers appear to play a larger role in transmitting the common inherited variants that each contribute a tiny nudge toward autism risk. This finding ties into one of the most interesting patterns in autism genetics: the so-called female protective effect. Boys are diagnosed with autism roughly three to four times more often than girls, and the genetic data suggest this is not just a diagnostic bias. Girls seem to require a higher load of autism-associated genetic variants before they show clinical features.
A study of over 7,400 mothers and nearly 6,000 fathers of children with autism found that mothers carried significantly more common polygenic risk for autism than fathers did. The increase in autism-related genetic risk among mothers of autistic children was about 50% greater, relative to the general population, than the corresponding increase among fathers.7Cell Genomics. Common inherited genetic risk and a female protective effect in autism spectrum disorder In other words, these mothers harbored substantial autism-associated genetic variation themselves but were buffered from a clinical diagnosis, likely because of the female protective effect. They then passed that genetic load to their children, including their sons, who lacked the same buffer.
This pattern was replicated across two independent cohorts and held up when looking at complete mother-father-child trios.8PubMed Central. The female protective effect against autism spectrum disorder The practical upshot is striking: a mother can carry a high genetic predisposition to autism, never be diagnosed herself, and still transmit that predisposition to her children. Fathers contribute common variants too, but because they lack the same protective buffer, men with high autism polygenic risk are more likely to already be on the spectrum themselves rather than passing it along silently.
Parental Age Effects From Both Sides
The paternal age story gets most of the attention, but maternal age matters too. A meta-analysis found that mothers aged 35 and older had roughly 1.5 times the risk of having an autistic child compared with mothers aged 25 to 29, and this association persisted even after controlling for paternal age.9Journal of the American Academy of Child & Adolescent Psychiatry. Advancing Maternal Age Is Associated With Increasing Risk for Autism: A Review and Meta-Analysis The risk increased in a dose-response fashion: the older the mother, the higher the odds.
A large international study examining over five million children confirmed that both advancing paternal and maternal age independently raised the risk. Fathers over 50 compared with fathers in their twenties had about 1.66 times the risk, while mothers aged 40 to 49 compared with mothers in their twenties had about 1.15 times the risk. Interestingly, very young mothers, those under 20, also showed a raised risk.10Molecular Psychiatry. Autism risk associated with parental age and with increasing difference in age between the parents The mechanisms behind the maternal age effect are less clear than the paternal one. They likely involve a mix of egg-cell aging, changes in the uterine environment, and possibly accumulated epigenetic shifts, rather than the straightforward mutation-accumulation story seen in sperm.
One wrinkle worth noting: in families that already have an elevated genetic background for autism, the parental-age pattern can look different. A study of high-risk families, where an older sibling was already diagnosed, found that younger fathers (under 30) actually had higher odds of having another autistic child, and advanced parental age did not show the expected association with autism diagnosis.11PubMed Central. The Association Between Parental Age and Autism-Related Outcomes in Children at High Familial Risk for Autism This suggests the age effect seen in general-population studies may be more about introducing new risk in low-risk families than amplifying existing risk in high-risk ones.
Mitochondrial DNA: A Uniquely Maternal Channel
Mitochondria, the energy-producing structures inside cells, carry their own small genome. Unlike the DNA in the cell’s nucleus, mitochondrial DNA is inherited almost entirely from the mother. This creates one channel of genetic risk that is genuinely mother-specific.
Research has found that children with autism carry more potentially harmful mitochondrial DNA variants than their unaffected siblings. One study found roughly a 1.5-fold enrichment of protein-altering mitochondrial mutations and about a 2.2-fold enrichment of predicted pathogenic mutations in autistic children compared with their non-autistic siblings. The risk was most pronounced in children with lower IQ or greater social impairment.12PLoS Genetics. Genetic Evidence for Elevated Pathogenicity of Mitochondrial DNA Heteroplasmy in Autism Spectrum Disorder A follow-up study confirmed that both inherited and newly arising mitochondrial variants were elevated in autistic children, and that harmful variants inherited from the mother appeared to be transmitted preferentially to the affected child rather than to unaffected siblings.13Nature Communications. Association of mitochondrial DNA content, heteroplasmies and inter-generational transmission with autism
This does not mean mitochondrial mutations are a major driver of autism on their own. The brain is energy-hungry, and mitochondrial dysfunction can affect neurodevelopment, but the effect sizes are modest and the vast majority of autism cases involve nuclear DNA variants from both parents.14PubMed Central. Mitochondrial dysfunction: A hidden trigger of autism? Still, mitochondrial DNA is one concrete example of a genetic pathway that runs exclusively through the maternal line.
Epigenetic Marks in Sperm
Beyond the DNA sequence itself, chemical tags on DNA can influence which genes get turned on or off. These epigenetic marks can differ between individuals and, in some cases, be passed from parent to child. Research on fathers of autistic children has turned up intriguing patterns in sperm.
One study identified 193 regions of altered DNA methylation in the sperm of fathers whose children later showed early signs of autism risk. Many of these regions clustered near genes involved in brain development, including genes within a chromosomal region associated with Prader-Willi syndrome, a neurodevelopmental condition caused by a known imprinting defect. Some of the same methylation differences appeared in the cerebellums of autistic individuals.15PubMed Central. Paternal sperm DNA methylation associated with early signs of autism risk in an autism-enriched cohort A separate study identified 805 differentially methylated regions in the sperm of fathers with autistic children compared to fathers without, and these markers predicted paternal autism susceptibility with roughly 90% accuracy in a blinded validation test.16PubMed Central. Sperm DNA methylation epimutation biomarker for paternal offspring autism susceptibility
These are small, early studies, and the field is far from using sperm methylation as a clinical biomarker. But they point to a father-specific epigenetic channel that sits alongside the de novo mutation pathway. What a father’s sperm carries is not just a DNA sequence but also a set of chemical annotations that may shape how genes behave in the developing brain.
Non-Coding DNA and Parent-of-Origin Patterns
Most of the genome does not code for proteins. Instead, large stretches of DNA act as regulatory switches that control when and where genes turn on. Research has found that structural changes in these regulatory regions show a parent-of-origin bias in autism. In a study of over 2,600 families, rare structural variants in regulatory elements were preferentially transmitted from fathers to affected children, but not to their unaffected siblings.17PubMed Central. Paternally inherited cis-regulatory structural variants are associated with autism
In at least one documented case, a paternally inherited variant in a regulatory RNA altered the binding of several autism risk genes and appeared to compound a separate maternally inherited coding mutation in the same gene family, with both parents’ contributions layering to raise risk in a single child.18Molecular Psychiatry. An integrative analysis of non-coding regulatory DNA variations associated with autism spectrum disorder This kind of cross-parental interaction, where a regulatory hit from one parent amplifies a coding hit from the other, makes it essentially impossible to assign autism to one parent’s lineage.
Meanwhile, when researchers looked at rare copy-number variants (deletions or duplications of DNA chunks) more broadly, the transmission was remarkably even. In one large genome-wide analysis, 233 ultra-rare events came from fathers and 223 came from mothers, showing no gender bias.19Neuron. Genome-Wide Rare Copy-Number Variations Associated with Autism Spectrum Disorders Through Genomic Analysis of Simplex Family Kinds The parent-of-origin effects in autism appear to be concentrated in specific types of genetic variation, not spread uniformly across the genome.
Genomic Imprinting and Why Parent of Origin Matters at All
The reason any of this is parent-specific comes down to a phenomenon called genomic imprinting. For most genes, both copies, one from each parent, are active. But for a small set of genes, only the copy from one parent is switched on, while the other is chemically silenced. If the active copy happens to carry a harmful variant, the silent copy from the other parent cannot compensate.
A genome-wide scan of autism families found the strongest evidence for parent-of-origin effects on chromosomes 4, 15, and 20, all sites where imprinted genes potentially related to autism may reside.20PLoS ONE. Parent-Of-Origin Effects in Autism Identified through Genome-Wide Linkage Analysis of 16,000 SNPs The chromosome 15 region is particularly well-studied because duplications of 15q11-13, when inherited from the mother, are one of the more common single genetic causes of autism, while the same duplication from the father typically has little effect. A comprehensive genomic analysis of multiplex families confirmed this maternal over-transmission at the 15q11-13 locus.21American Journal of Human Genetics. Comprehensive Genomic Analysis of Multiplex Families Identifies Rare and Common Risk Loci for Autism Spectrum Disorder
Imprinting means that for certain specific genetic events, it genuinely matters which parent a variant came from. But these are exceptions to the general rule. For the vast majority of autism-related genetic variation, both parents’ contributions are active and roughly equal in their potential influence. A broad narrative review of parent-of-origin effects in autism noted that while imprinting, mitochondrial DNA, sex chromosomes, and mutational transmission biases all create legitimate parent-specific channels, no single one of these mechanisms dominates autism risk overall.22PubMed Central. Evidence for parent-of-origin effects in autism spectrum disorder: a narrative review
Family History Through Each Lineage
If one parent is the predominant source of autism risk, you would expect family history on that side to be a much stronger predictor. A Swedish registry study tested this directly. Having a maternal aunt or uncle with autism raised a child’s risk by about threefold compared to the general population, while having a paternal aunt or uncle raised it about twofold.23PubMed Central. Inherited Risk for Autism through Maternal and Paternal Lineage The maternal side carried slightly more risk, consistent with the female protective effect discussed earlier: women on the maternal side may harbor more autism-related variants without being diagnosed. But the paternal side was not far behind, and the difference was modest. Family history from either parent’s side is informative.
The Maternal Immune Environment
One route of maternal influence that has nothing to do with DNA is the mother’s immune system during pregnancy. Some mothers produce antibodies that react against proteins in the developing fetal brain. These maternal autoantibodies can cross the placenta and, in animal models, cause behavioral changes in offspring. Researchers have identified seven specific brain proteins targeted by these antibodies, and the presence of the antibodies is associated with a diagnosis of what has been termed “maternal autoantibody-related” autism.24PubMed Central. Maternal Anti-Fetal Brain IgG Autoantibodies and Autism Spectrum Disorder: Current Knowledge and its Implications for Potential Therapeutics This subtype appears to account for a minority of autism cases, but it represents a genuinely mother-specific, non-genetic risk factor that has no equivalent on the paternal side.
Assortative Mating and Why Both Parents Often Carry Risk
There is another reason the “mother or father” question can be misleading: parents of autistic children tend to resemble each other on autism-related traits more than random pairs of adults. A study of parental traits found significant positive correlations between spouses on measures of social responsiveness and broader autism-related personality features.25American Journal of Human Genetics. Assortative mating and parental-relatedness biases shape the genetic architecture of neurodevelopmental disorders People with subtle autism-related traits appear to gravitate toward partners with similar traits. When both parents carry subclinical genetic predisposition, the child inherits a double dose from both sides, making it impossible to single out one parent as “the source.”
Risk Echoes Across Generations
The age-related mutation accumulation story extends beyond a single generation. Research has found that a grandfather’s age at the time he fathered the parent also influences a grandchild’s autism risk. One study showed a monotonic relationship: the older the grandfather was when the parent was born, the higher the grandchild’s risk of autism.26PubMed Central. Autism risk across generations: a population-based study of advancing grandpaternal and paternal age A larger follow-up found U-shaped associations, with both very young and older grandparents linked to elevated risk in grandchildren. The pattern held for maternal grandmothers, maternal grandfathers, and paternal grandmothers alike.27PubMed Central. Association of Grandparental and Parental Age at Childbirth With Autism Spectrum Disorder in Children
These multigenerational effects underscore a humbling reality: the genetic variants that contribute to a child’s autism risk may have originated not in either parent, but in a grandparent, carried silently through one generation before combining with other risk factors in the next. Autism genetics is not a snapshot; it is a slow-moving process that accumulates across family lines from both sides.