Transcription
In 2022, a research team at the Broad Institute in Cambridge, Massachusetts, walked into an annual genetics conference in Boston and presented findings from a study that had taken 3 years to complete and had consumed more computational resources than almost any previous genetics project in history. The lead researcher stood at the podium and began describing what her team had found after mapping brain tissue DNA from over 15,000 individuals drawn from every major population group on Earth. Europeans, Africans, East Asians, South Asians, Indigenous Americans, Pacific Islanders, all represented in carefully balanced sample sizes designed to eliminate the statistical bias that had plagued previous studies.
The room was full of scientists who had spent careers working on human genetics, people who thought they had a reasonably clear picture of what the data would show. As the findings appeared on the presentation screen, the room went silent. Not the polite silence of an audience following a technical presentation, the silence of people confronting something that contradicted assumptions so deeply embedded in the scientific literature that most researchers had stopped questioning them. What the Broad Institute team had found did not confirm what centuries of scientific thinking had predicted. It demolished it. And the implications of what those 15,000 brain tissue samples revealed are still reverberating through genetics, anthropology, neuroscience, and every field that touches the question of what actually differs between human populations at the biological level.
To understand why the silence in that conference room was significant, you need to understand what the researchers had expected to find and why that expectation was so deeply rooted in scientific tradition. The human brain contains roughly 86 billion neurons connected by approximately 100 trillion synaptic junctions organized into regional architectures that handle everything from basic motor control to abstract reasoning to emotional regulation. The genes that govern how those neurons form during fetal development, how they migrate to their correct positions, how they establish their connections, and how they maintain their function across a lifetime represent some of the most complex regulatory networks in the entire human genome.
Scientists going into the Broad Institute study operated under an assumption that had been embedded in biological thinking since the 19th century. If human populations had been separated for tens of thousands of years, evolving under different environmental pressures and different climates, facing different survival challenges, then surely the genes controlling brain development and cognitive function would have diverged along with everything else. The assumption was not fringe thinking. It was the logical extension of standard evolutionary reasoning, and it had been repeated, refined, and built upon across more than a century of scientific literature.
The historical weight behind that assumption was substantial and ugly. In the 1800s, researchers measured skull sizes with calipers and published tables purporting to demonstrate that certain populations had larger brain volumes than others. The measurements were selectively reported, methodologically compromised, and shaped at every stage by the racial hierarchies the researchers assumed they were confirming rather than testing. In the early 20th century, IQ tests were administered to immigrant populations arriving at Ellis Island under conditions that guaranteed the tests were measuring English language fluency and familiarity with American cultural references rather than cognitive capacity. And the low scores were published as evidence of the intellectual inferiority of Southern and Eastern European populations. As late as 1994, a book called The Bell Curve argued at length that intelligence differences between racial groups had a genetic basis, and that these differences were substantial enough to have policy implications. The science underlying that argument was criticized extensively by geneticists and psychologists, but the argument continued to circulate because no study had yet been conducted at a scale and methodological rigor sufficient to address it with the kind of definitive authority that the question deserved.
The Broad Institute study was designed from the beginning to be that study. The team sequenced brain tissue samples from deceased donors who had agreed to contribute to genetic research, focusing specifically on tissue from the prefrontal cortex, the region responsible for executive function, decision-making, planning, and abstract reasoning. This is the most distinctively human region of the brain, the structure most associated with the cognitive capacities that separate human intelligence from that of other primates. If population differences in cognitive genetics existed anywhere, the prefrontal cortex was where they would appear most clearly.
The technical methodology used was single-cell RNA sequencing, a technique that examined gene expression in individual neurons rather than averaging signals across entire tissue samples. This granular resolution meant the researchers could see not just which genes were present in a tissue sample, but which genes were actively being transcribed into proteins in specific cell types at the time of death, providing a functional picture of gene activity rather than simply a catalog of genetic sequences. The team also analyzed the hippocampus, the brain region critical for memory formation, spatial navigation, and the consolidation of new information from short-term into long-term storage. They analyzed the amygdala, which processes emotional responses, fear conditioning, and threat evaluation. They analyzed the cerebellum, which coordinates movement and motor learning, and has more recently been implicated in cognitive processing as well. Each region was examined separately to allow the researchers to detect population differences that might be specific to particular cognitive domains rather than distributed uniformly across brain architecture. The study controlled for sex, for age at death, and for every confounding variable the team could identify using a sample size large enough to detect effect sizes considerably smaller than those that previous research had claimed to find.
The researchers focused their comparative analysis on two categories of genetic variants. Coding variants are changes to the DNA sequence that directly alter the structure of the proteins a gene produces, and these are the variants most likely to have functional consequences for how a gene works. Regulatory variants are changes in the DNA regions that control when a gene is activated and how strongly it expresses, and these can produce significant differences in biological outcomes, even when the gene sequence itself is identical between individuals. A gene present at the same frequency in two populations might express twice as strongly in one population if regulatory variants differ. And that difference in expression level could in principle have functional consequences. The team looked at both categories across all the genes associated with brain development, neural architecture, synaptic function, and cognitive performance that the scientific literature had identified as relevant.
When the analysis came back, the researchers ran it again. Then they ran it a third time using different statistical approaches to verify the result was not a methodological artifact. What they were seeing was not what the assumptions embedded in a century of scientific literature had predicted. The genetic differences between population groups in the genes controlling brain development and cognitive function were not the substantial and patterned differences that the evolutionary reasoning behind the old assumption had predicted. The variation within each population group was dramatically larger than the variation between groups. An individual from Nigeria and an individual from Sweden might show more similar brain gene expression profiles than two individuals who both identified as having Nigerian ancestry. The genetic lottery that governed how any individual's brain developed, how their neurons connected during fetal development, how efficiently their synaptic signaling operated, that lottery showed no alignment with the continental ancestry categories the researchers had used to group their samples.
The team identified over 200 genetic variants associated with cognitive traits, including working memory capacity, processing speed, verbal ability, and spatial reasoning. These variants existed in every population studied. Some individuals carried combinations associated with higher performance in specific cognitive domains. Others carried fewer of these variants, but when the researchers looked for patterns in which populations carried more or fewer of the high-performance variants, they found nothing. The distribution across populations was effectively random. There was no cluster of cognitive performance variants that appeared at systematically higher frequencies in any population group compared to any other.
The lead researcher published a statement in her paper that cut through the technical language of the field with unusual directness. "The genetic architecture of the brain," she wrote, "shows no evidence of population-specific adaptations related to cognition." This was not the first study to reach this conclusion. Earlier work had pointed in the same direction, but the Broad Institute study was the largest, the most technically sophisticated, and the most methodologically rigorous examination of the question ever conducted. It used sequencing technology that had not existed when earlier studies were done. It included populations that had been chronically underrepresented in genetic research, particularly indigenous communities and populations from sub-Saharan Africa and Oceania, whose exclusion from earlier studies had left gaps in the evidence base that critics could point to as limiting the conclusions. And it produced its finding not as a marginal result at the edge of statistical significance, but as a robust pattern that held up across every analytical approach the team applied.
The evolutionary reasoning that explained the finding was as important as the finding itself. Modern Homo sapiens evolved in Africa approximately 300,000 years ago. Small groups began leaving the African continent around 70,000 years ago, eventually reaching Europe, Asia, Australia, and eventually the Americas. In the context of evolutionary time, 70,000 years is extraordinarily brief. Natural selection requires many generations operating under consistent directional pressure to produce significant population-level changes in complex traits, and the changes it produces in that time frame are most apparent in simple traits controlled by small numbers of genes under intense and direct environmental pressure. Skin pigmentation changed rapidly because the relationship between UV radiation exposure and vitamin D synthesis created direct survival pressure that operated on every individual in every generation. The genetic changes producing lighter skin in high-latitude populations involved relatively few genes. The selective pressure was intense and consistent, and the trait changed measurably within just a few thousand years.
Brain development is the precise opposite of a simple trait. It involves thousands of genes working in coordinated regulatory networks of extraordinary complexity. Mutations that significantly alter the developmental trajectory of the brain are overwhelmingly harmful rather than beneficial because the system has been refined across hundreds of millions of years of vertebrate evolution into a precisely tuned architecture where almost any significant deviation from the existing plan produces catastrophic developmental failure. Evolution conserves these critical systems rather than experimenting with them. The FOXP2 gene, sometimes described as the language gene because mutations in it cause severe speech and language disorders, illustrates this conservation perfectly. The version of FOXP2 that humans carry differs from the chimpanzee version by just two amino acid changes that occurred somewhere in the last 6 million years of human evolution. The version in a person from Japan is identical to the version in a person from Kenya or Norway or Brazil because any change that disrupted FOXP2 function would be eliminated from the population by natural selection within a generation. The same pattern of conservation across all human populations holds for hundreds of other genes involved in neural development.
The cognitive challenges that shaped human intelligence across the period since the out-of-Africa dispersal were not unique to specific environments in ways that would have favored the evolution of population-specific cognitive profiles. They were universal. Arctic populations tracking migrating animals across featureless ice fields needed sophisticated spatial memory, weather prediction capabilities, and the engineering intelligence to create insulating clothing from animal skins. Desert populations navigating across vast territories between scattered water sources needed spatial reasoning, stellar navigation, and encyclopedic environmental knowledge. Rainforest populations needed to identify hundreds of plant species as edible or toxic, track prey through dense vegetation using subtle sensory cues, and maintain complex kinship networks across dispersed communities. Each of these challenges was different in its specific content, but identical in its cognitive requirements. All of them demanded working memory, executive function, social intelligence, pattern recognition, and language simultaneously. There was no environment in which intelligence was a disadvantage, and no environment in which a specialized and narrowed form of intelligence would have outperformed the general cognitive toolkit that all human populations carried.
The social brain hypothesis, developed by evolutionary anthropologists studying the relationship between group size and brain development, provided a complementary explanation for why cognitive genetics had not diverged between populations. The primary driver of human cognitive evolution was not environmental problem solving, but social complexity. As human group sizes grew from small family bands to larger tribal structures to the complex hierarchical societies that formed during the agricultural transition, the cognitive demands of tracking social relationships, anticipating others' behavior, navigating status hierarchies, detecting deception, and coordinating cooperative activities with non-kin expanded dramatically. The individuals who could handle this social complexity survived and reproduced at higher rates than those who could not, driving the expansion of the prefrontal cortex and the cognitive networks it supports across the entire human species simultaneously. These social pressures were identical in every human population regardless of geography. A Pleistocene hunter-gatherer in the Congo Basin and a Pleistocene hunter-gatherer on the Mongolian steppe faced different physical environments, but identical social cognitive demands, and natural selection responded to those demands in the same way in both populations.
The finding that racial categories predicted almost nothing about individual brain gene profiles dismantled one of the most persistent myths in the popular understanding of cognitive differences between populations. The stereotypes were specific and familiar. East Asians are better at mathematics. Europeans are better at abstract reasoning. Africans are naturally gifted at tasks requiring rapid physical reflexes. These ideas had been repeated so frequently across so many generations that they had acquired the cultural weight of established facts for many people. The genetic data contained no support for any of them. The genes associated with mathematical reasoning, spatial processing, verbal fluency, and working memory were distributed without pattern across all population groups.
Cultural and educational factors produced real differences in measured performance on specific cognitive tasks. Because a child raised in an educational environment that emphasized mathematical training would perform differently on mathematics assessments than a child who was not. But those differences in measured performance reflected differences in experience and opportunity, rather than differences in the underlying genetic architecture of cognition.
The epigenetic component of the Broad Institute findings added a dimension to the story that reached beyond genetics into the lived experience of poverty, chronic stress, and educational deprivation. The researchers included a secondary analysis comparing brain gene expression patterns in individuals from different socioeconomic backgrounds within the same ancestry group. And what they found illuminated the mechanism through which environmental disadvantage produced measurable differences in cognitive outcomes without involving any difference in the genetic code itself. Epigenetics is the study of how chemical modifications to DNA and to the histone proteins that DNA wraps around can alter which genes are expressed without changing the underlying sequence. A gene tightly wrapped and tagged with suppressive chemical markers might never be read by the cellular machinery that produces proteins. A gene loosely wrapped and marked for expression might be transcribed continuously. The DNA sequence is identical in both cases. The functional outcome is entirely different.
Children raised in environments characterized by chronic stress, material deprivation, and limited cognitive stimulation showed distinct patterns of gene expression in brain regions associated with memory formation and emotional regulation. Specifically, genes involved in cortisol receptor production showed altered expression profiles consistent with a biological adaptation to chronic stress exposure. When the stress hormone cortisol floods the brain repeatedly and continuously, as it does in children living with the chronic unpredictability of poverty, the brain responds by downregulating the sensitivity of the cortisol system, attaching epigenetic tags that reduce the expression of cortisol receptor genes. This adaptation reduces the overwhelming effect of chronic stress signals in the short term, but it impairs the normal stress response dynamics that are necessary for consolidating memories and regulating emotional responses over the long term. The adaptation is a solution to an immediate problem that creates a different and longer-lasting problem in its place.
The BDNF gene, which produces brain-derived neurotrophic factor, a protein that acts essentially as fertilizer for neural connections and is critical for the synaptic plasticity underlying learning and memory, showed reduced expression in brain tissue from individuals who had experienced childhood deprivation. Less BDNF meant slower synapse formation, weaker memory consolidation, and reduced capacity for the neural reorganization that underlies learning new skills. Critically, these were not permanent, irreversible changes to the genetic code. They were epigenetic modifications that could in principle be reversed when environmental conditions improved. Research tracking individuals who had experienced early deprivation but later gained access to stable and stimulating environments showed partial normalization of gene expression patterns with BDNF levels recovering toward baseline as the epigenetic tags were modified by the improved environmental context. The brain's plasticity, the very capacity that had been suppressed by chronic stress, allowed it to partially recover when the stressor was removed.
This epigenetic finding provided the biological mechanism explaining something that educational research had documented empirically for decades. Early intervention programs targeting children from disadvantaged backgrounds produced cognitive benefits that persisted into adulthood and could not be explained by the content of the educational program alone. Programs like Head Start in the United States produced measurable improvements in cognitive outcomes that follow-up research showed were still detectable decades later. The genetic data provided an explanation for why this worked. These programs were not compensating for genetic deficits that did not exist. They were removing the epigenetic effects of chronic stress and deprivation that had been suppressing the expression of cognitive genes that were present and functional in those children's genomes from birth. They were giving the genome the environmental conditions it needed to express its full potential, and the genome responded.
The researchers also examined ancient DNA recovered from Neanderthal and Denisovan remains to explore whether the small amounts of archaic hominin DNA carried by modern populations outside Africa had contributed to any cognitive differences between groups. Modern humans of European and Asian descent carry between 1 and 4% Neanderthal DNA, the genetic legacy of interbreeding that occurred after the out-of-Africa founding population encountered Neanderthals in the Near East and Europe. Some populations in Oceania carry Denisovan DNA at similar frequencies. The analysis found that the archaic DNA variants that had survived in modern human populations were concentrated in genomic regions associated with immune function, skin physiology, and metabolic traits. Very few were located in genes associated with brain development, and the ones that were showed no functional pattern suggesting enhancement or diminishment of cognitive capacity in any direction. The Neanderthal and Denisovan contributions to modern human genomes, whatever advantages they may have provided in immune defense against Eurasian pathogens, had not created cognitive differences between the populations carrying them and those that did not.
The persistence of cognitive racial hierarchy myths, despite the absence of supporting genetic evidence, demanded its own explanation, and the Broad Institute researchers addressed it directly in the discussion section of their paper. The answer was not simply ignorance or misunderstanding. It was that the myths served specific social and political functions that made them resistant to empirical correction. For centuries, pseudo-scientific theories of racial intellectual hierarchy provided the ideological justification for slavery, colonial exploitation, and systematic exclusion of specific groups from education, economic participation, and political power. If certain populations were biologically predisposed to intellectual inferiority, then the economic and social structures built on their exploitation were not moral failures, but natural arrangements reflecting biological reality. The science was always constructed to confirm this conclusion, rather than to test it, because the political and economic interests of those conducting and funding the research were aligned with a specific outcome. The methodology was adjusted, the data was selectively reported, and alternative explanations were dismissed without engagement, because the conclusion was determined before the analysis began.
The persistence of these ideas after their scientific basis had been repeatedly challenged, reflected the degree to which they had become embedded in cultural assumptions that operated below the level of explicit reasoning. People observing gaps in educational achievement, economic attainment, and performance on standardized tests sought explanations that were simple, stable, and absolved existing social structures of responsibility. Genetic determinism provided exactly this kind of explanation. If the gaps were biological, they were inevitable and immutable, requiring no examination of the historical processes that produced them, and no action to address the ongoing systems that maintain them. When those gaps were instead explained by the documented effects of unequal access to education, nutrition, health care, and stable environments across multiple generations, the explanation required confronting the active role of social policy in producing and maintaining cognitive disadvantage, a far more uncomfortable and politically complex conclusion than the genetic alternative.
The Broad Institute data, combined with the epigenetic findings from the secondary analysis, reframed the entire question of cognitive differences between population groups in a way that pointed not toward biology, but toward policy. The differences in measured cognitive cognitive performance between groups were real. Nobody conducting the research disputed that measured differences existed. What the genetic data established was that those differences had no foundation in the DNA sequence of the populations involved, that the brain genetic architecture was essentially identical across all human groups, and that the environmental factors known to alter gene expression in ways that impaired cognitive development were distributed unequally across populations in patterns that directly reflected historical and ongoing inequities in resource access. The gaps were the product of choices, systems, and histories, not of genomes.
The statistic that most directly communicated the magnitude of within-group versus between-group genetic variation in cognitive genes was one that the research team presented as a thought experiment to make the technical findings accessible. Take two people from the same ethnic group, any group, and compare their complete genomes. You will find hundreds of thousands of genetic differences between them. Now, take two people from the most genetically distant populations on Earth and compare their genomes. You will find perhaps 10 to 20,000 additional differences beyond what you found within the first group. The between-group difference represents a small fraction of the within-group difference. And of those additional between-group differences, virtually none touch the genes involved in cognitive development. Two siblings with identical ancestry and identical parents can have dramatically different cognitive profiles because the random recombination of genetic variants that occurs during reproduction produces enormous individual variation within families. That individual variation dwarfs anything that separates population groups, and it makes the concept of predicting cognitive potential from ancestry not just scientifically unsupported, but statistically absurd.
What the 15,000 brain tissue samples ultimately revealed when the Broad Institute team finished their analysis was something that should have been predictable from first principles of evolutionary biology, but that required the scale of evidence to establish with the authority the question demanded. The human brain, the structure that makes the human species what it is, was too important and too complex for evolution to experiment with across the brief period since populations began diverging from their common African origin. The genes that built it were locked in place by the weight of millions of years of vertebrate neural evolution, conserved across all human populations because any significant deviation from the existing plan was immediately lethal rather than adaptive. What varied between populations were the traits that had faced specific and direct environmental pressure in the recent evolutionary past. The melanin content of skin, the efficiency of vitamin D synthesis, the immune system's catalog of pathogen recognition tools, the metabolic machinery for processing specific dietary staples. The machinery of the mind itself remained common property of the entire human species, distributed without pattern across every population, available in equal potential to every individual regardless of where their ancestors had lived for the last 70,000 years.
The results shocked everyone in that conference room in Boston, not because they were scientifically implausible. In retrospect, every principle of evolutionary biology pointed toward exactly this outcome. They shocked everyone because the weight of a century of a pseudo-scientific tradition had made the alternative seem like the more plausible prediction. What the data showed was that the alternative had never been supportable on scientific grounds. It had been supported by the social and political interests of the people who constructed it. The genome, read carefully enough and at sufficient scale, told a different story. One species, one cognitive inheritance. And every gap in outcome a product of the world humans had built around themselves rather than anything written in the DNA they had inherited from their common ancestors in Africa.