Epigenetics and IQ Malleability

The heritability of intelligence sits at roughly 70–80% in adulthood, but genome-wide association studies aggregating thousands of common variants currently account for only about 10–15% of variance in cognitive ability. The gap — the so-called “missing heritability” — is one of the most discussed problems in modern intelligence research. Epigenetics is one of the leading candidates for what fills it: chemical modifications that change how genes are expressed without altering the DNA sequence itself, sitting in the causal pathway between environment and phenotype. The picture that has emerged across a decade of work is not “intelligence is epigenetic” but something more interesting — epigenetic marks are part of the mechanism by which environment shapes the readout of an otherwise stable genome.

What epigenetics actually means

The term covers heritable changes in gene expression that do not involve changes in the underlying DNA. The most-studied form is DNA methylation: the addition of a methyl group to a cytosine base, typically at CpG dinucleotides, which generally suppresses transcription of the affected gene. Histone modifications and non-coding RNA effects matter too, but DNA methylation is the most accessible and most heavily measured in human cognitive studies because it can be assayed at scale from blood samples.

Two features make epigenetic marks interesting for intelligence research. First, they are responsive to environment: prenatal nutrition, early-life stress, exposure to pollutants, and even neighbourhood characteristics have been linked to specific methylation patterns. Second, they are partially heritable: some marks are passed from parent to offspring, blurring the simple separation between “genetic” and “environmental” causes. Together those two properties give epigenetics the structural role of a translator between an environmental input and a downstream cognitive output.

The missing heritability problem

Twin and adoption studies converge on adult IQ heritability of 50–80%. Modern genome-wide association studies, summarised in Plomin and von Stumm’s (2018) review of “the new genetics of intelligence,” now identify hundreds of loci associated with cognitive ability, but polygenic scores aggregating these variants predict only about 10–15% of cognitive variance. The remaining gap is partly attributable to rare variants and gene-gene interactions, but a substantial portion is plausibly carried by mechanisms that the standard GWAS framework does not capture — including epigenetic regulation. Heritability is a population statistic about variance partition, not a mechanism for any individual; epigenetics is one of the molecular candidates for what is doing the work underneath.

The DRD2 finding: epigenetic variance and IQ

The clearest case study to date is Kaminski et al.’s (2018) analysis of the IMAGEN sample — 1,475 healthy adolescents with combined genetic, epigenetic, structural-imaging, and functional-imaging data. The authors focused on the dopamine D2 receptor gene (DRD2), a well-established node in the neural circuitry of reward processing and working memory.

The headline finding was that methylation of DRD2 was significantly associated with general intelligence, with higher methylation tracking lower IQ scores. The methylation signal also tracked grey matter density in the striatum and striatal activation during a reward-anticipation task — tying the peripheral epigenetic measure to a plausible central nervous system substrate. Because DRD2 methylation is partly environmentally responsive, the result implicates a specific molecular pathway through which environmental input could modulate cognitive performance: methylation silences DRD2, fewer dopamine receptors are expressed, signalling efficiency drops, and downstream cognitive performance is reduced.

Two caveats are important. The methylation was measured in blood, and the assumption that blood methylation reflects brain methylation is reasonable for some loci but not universal. The IMAGEN sample is also adolescent and European; replication in older and more diverse samples is still needed. The finding nonetheless gave the field its first concrete worked example of an epigenetic-cognitive link with neurobiological correlates.

Beyond DRD2: epigenome-wide evidence

A single-gene candidate study like Kaminski et al. (2018) is informative but limited; the modern complement is the epigenome-wide association study (EWAS), which scans hundreds of thousands of CpG sites simultaneously and asks which are associated with cognitive performance. Marioni et al. (2018) conducted the first large meta-analysis of EWAS for cognitive abilities, pooling six adult cohorts. They identified replicable associations between DNA methylation at specific loci and cognitive performance independent of measured genetic variants — evidence that methylation captures a signal that polygenic scores miss.

Complementary work on the upstream side examined the genes that lay down methylation marks. Haggarty et al. (2010) tested polymorphisms in DNA methyltransferase genes (DNMT1 and DNMT3) for association with intelligence in older adults from the Aberdeen Birth Cohort and the Lothian Birth Cohort 1936. They found small but statistically significant associations between specific DNMT polymorphisms and adult cognitive ability, supporting the broader picture that the machinery laying down methylation marks is itself relevant to individual differences in intelligence.

From environment to epigenetic mark to cognitive outcome

The most striking recent evidence comes from studies that trace a full environmental-to-phenotype pathway through methylation. Lee et al.’s (2021) Korean prospective cohort followed children with measured residential greenness exposure at age 2 and assessed both blood DNA methylation and IQ at age 6. They identified 25 CpG sites whose methylation was significantly associated with greenness exposure, including a site in the body of SLC6A3, the dopamine transporter gene. Methylation at that specific site at age 2 was in turn associated with performance IQ at age 6, completing the chain: environmental input → epigenetic mark → cognitive outcome.

This is the kind of evidence that earlier candidate-gene work could only gesture at, and it gives epigenetics a credible mechanistic role in the broader literature on how environment shapes cognitive development — including the well-documented evidence that years of formal education raise IQ and that childhood environmental factors leave durable cognitive signatures. Cogn-IQ’s overview of epigenetics in cognitive abilities covers the broader landscape accessibly.

The reversibility angle and what it implies for malleability

Unlike DNA mutations, methylation marks are not permanent. They can be added and removed throughout life by enzymatic machinery that is itself responsive to nutrition, stress, exercise, and pharmacological intervention. This is the hook that makes epigenetics interesting to people thinking about whether IQ can be raised: a mechanism that is sensitive to environment and is biochemically reversible is exactly the kind of substrate that could in principle support cognitive intervention.

The honest reading of the current evidence is more cautious than the popular framing suggests. Reversibility at the level of individual CpG sites is well demonstrated; reversibility of complex behavioural phenotypes through targeted methylation manipulation is not. No clinical intervention is currently available that can reliably shift cognitive performance by manipulating methylation in healthy individuals, and the established environmental interventions (education, nutrition, exercise) plausibly act partly through epigenetic mechanisms but are administered without targeting methylation specifically. Epigenetics provides a molecular framework for understanding why intelligence responds to environment despite high heritability; it does not yet provide a clinical handle on the trait.

Limitations and open questions

Several caveats apply across the literature. Most studies measure methylation in peripheral tissue (blood, saliva) rather than brain, and the cross-tissue correlation varies by locus. Effect sizes for individual CpG sites on cognitive outcomes are typically small, in line with the polygenic-architecture pattern seen in genetics. Causal direction is often ambiguous: a methylation difference could cause a cognitive difference, reflect a downstream consequence of one, or share a common upstream cause. Longitudinal designs (like Lee et al., 2021) help with this, but most published evidence is still cross-sectional.

The relationship between epigenetic marks and the heritability gap is itself empirical: how much of the missing heritability is genuinely epigenetic, versus rare-variant-driven, versus reflecting non-additive effects, is an active research question. Treating epigenetics as the obvious explanation overshoots the data. Treating it as a non-factor undershoots them.

Frequently asked questions

Is intelligence determined by epigenetics?

No. Intelligence is influenced by inherited DNA sequence variation (heritability 50–80%), environmental factors, and the epigenetic marks that mediate between them. Epigenetics is one mechanism among several, not the cause of intelligence.

Can epigenetic changes raise or lower IQ?

The evidence is suggestive rather than conclusive. Specific methylation patterns in genes like DRD2 and SLC6A3 are statistically associated with cognitive performance (Kaminski et al., 2018; Lee et al., 2021), and the marks are environmentally responsive. Whether deliberately changing methylation in those genes would change IQ in any meaningful way has not been demonstrated.

What is the “missing heritability” problem?

Twin studies estimate intelligence heritability at 50–80% in adulthood, but polygenic scores from genome-wide association studies currently capture only about 10–15% of cognitive variance. The gap between these numbers is the missing heritability. Epigenetic variation is one plausible contributor to closing it.

Can stress or pollution change my IQ through epigenetics?

Environmental factors including chronic stress, exposure to pollutants, prenatal nutrition, and even residential greenness have been shown to leave specific DNA methylation patterns that are statistically associated with cognitive outcomes. The effect sizes for any single environmental input are small, but the cumulative contribution of environment-driven epigenetic regulation across development is consistent with the literature on environmental modifiability of intelligence.

Are epigenetic changes inherited?

Some are. Transgenerational epigenetic inheritance is well documented in animal models and increasingly supported in humans, though the magnitude of human transgenerational effects remains debated. Most epigenetic marks are reset each generation; a minority appear to escape this reset and contribute to parent-to-child transmission of phenotypic variation.