The Cognitive Environment: How Place and Education Shape the Habit of Reasoning

Socioeconomic environments shape how cognitive ability is expressed
Stress, sleep, and opportunity influence developing brain function
Rigorous education can cultivate causal and analytical reasoning

Intelligence is often discussed as though it were an isolated property of the individual. Educational achievement, test performance, and professional success are consequently interpreted as direct reflections of talent, motivation, or discipline.

A major study published in Science presents a more complicated picture. By comparing hundreds of biological, psychological, social, and environmental variables with brain-imaging data from children, researchers found that socioeconomic conditions were more strongly associated with brain structure and function than any other category they examined.

The study does not suggest that a child’s location mechanically determines intelligence or destiny. Rather, it shows how strongly the environment surrounding a child may be reflected in the developing brain—and how easily differences in stress, sleep, security, and opportunity can be mistaken for differences in underlying cognitive ability.

The socioeconomic signal in the developing brain

The researchers analysed data from 11,878 children aged nine to ten who were participating in the Adolescent Brain Cognitive Development Study, a long-term research programme examining child health and brain development in the United States. The analysis considered 649 variables across 12 broad categories, including socioeconomic conditions, cognitive performance, screen time, physical and mental health, parenting, personality, demographics, social adjustment, and medical history.

Socioeconomic variables showed the strongest and most consistent relationships with both brain structure and brain function. Of the 40 variables most closely associated with brain function, 37 were socioeconomic. Of the 40 most closely associated with brain structure, 35 were socioeconomic. Taken together, socioeconomic factors accounted for approximately 16 per cent of the observed variation in measures of brain function.

These variables included household income, homeownership, neighbourhood poverty, transportation access, and the social and economic opportunities available in the surrounding community. The study also incorporated measures related to the Child Opportunity Index, which assesses neighbourhood resources such as housing conditions, access to food, environmental quality, and educational opportunities.

This does not mean that every child living in a disadvantaged area develops the same brain characteristics. The results describe statistical associations across a large population, not deterministic outcomes for particular individuals.

Nevertheless, the strength of the socioeconomic signal was striking. It substantially exceeded the associations found for parenting style, medical history, mental health, and measured cognitive ability.

A tired and stressed brain is not a less intelligent brain

One of the study’s most consequential findings concerns the location of the observed brain associations.

Socioeconomic conditions were particularly associated with functional features in motor and sensory systems—areas that are sensitive to daily changes in sleep, fatigue, and stress. Brain regions more directly associated with abstract cognition and problem-solving were comparatively less affected.

The researchers therefore cautioned against interpreting the observed patterns as evidence that children from disadvantaged environments possess lower intellectual capacity. Instead, the brains of children exposed to lower socioeconomic opportunity appeared more similar to those of children experiencing sleep deprivation and chronic stress.

This distinction is essential.

A child who performs poorly while tired, anxious, or under persistent environmental pressure may be displaying the effects of those conditions rather than a fundamental limitation in cognitive ability. Educational systems that treat observed performance as a transparent measure of innate intelligence may therefore reproduce environmental inequalities while appearing to measure merit.

The study also questioned some previously reported associations between IQ scores and brain characteristics. After the researchers statistically adjusted for socioeconomic conditions, approximately 70 per cent of the observed associations between IQ and brain measures were no longer statistically significant. Among children from relatively advantaged socioeconomic backgrounds, IQ scores were not significantly correlated with the brain measures examined in the study.

These findings do not settle the broader scientific debate over intelligence, genetics, or cognitive measurement. They do, however, show that socioeconomic context can confound relationships that might otherwise be interpreted as biological signatures of intelligence.

Environment is not destiny

The findings should not be interpreted as evidence that childhood circumstances permanently determine the brain.

Many of the associations were found in brain function rather than fixed structural characteristics. The researchers consequently suggested that at least some of the observed differences may be responsive to changes in sleep, stress, and the immediate environment.

That possibility is encouraging, but it remains a hypothesis rather than a demonstrated intervention effect. The study was observational and cannot by itself establish that socioeconomic disadvantage caused the brain patterns it identified. It included measurements from only two points in the children’s development, and it could not determine when the relevant environmental effects began or whether they would continue through adolescence and adulthood.

The analysis also did not fully incorporate individual genetic differences. Although genetic ancestry was considered, the study did not include comprehensive polygenic measures of predispositions related to health, behaviour, or educational attainment.

The appropriate conclusion is therefore neither that biology is irrelevant nor that environment explains every individual difference. It is that environment is too important to be treated as background noise.

Schools as intellectual environments

A child’s environment is not limited to household income or neighbourhood infrastructure. Schools also form a substantial part of the environment in which attention, confidence, curiosity, and reasoning develop.

The Science study did not directly test particular teaching methods, nor did it demonstrate that one curriculum changes the brain more effectively than another. Its findings nevertheless reinforce a broader principle: cognitive performance emerges within an environment.

A productive educational environment provides more than information. It establishes routines, expectations, intellectual models, and forms of feedback. It influences whether students perceive difficult questions as threats to be avoided or problems that can be investigated systematically.

Rigour is important within such an environment, but rigour should not be confused with excessive pressure.

An intellectually rigorous school asks students to explain why an answer follows from the available evidence. It exposes them to problems for which the method is not given in advance. It requires them to distinguish correlation from causation, test assumptions, identify uncertainty, and revise conclusions when evidence contradicts an initial belief.

A merely stressful school may produce long hours, high test scores, and intense competition without cultivating these habits. Chronic stress can narrow attention toward immediate performance and error avoidance. Intellectual rigour, by contrast, should expand the learner’s capacity to formulate questions, tolerate ambiguity, and construct defensible arguments.

The difference lies not simply in how difficult the curriculum is, but in what type of thinking the environment repeatedly rewards.

From answering questions to discovering causes

The founding motto of the Swiss Institute of Artificial Intelligence is Rerum Cognoscere Causas—to know the causes of things.

This principle describes a particular intellectual orientation. A problem should not be approached only by searching for the fastest available technique or by selecting an answer from a predefined set of alternatives. It should be examined by identifying the underlying mechanism, separating relevant evidence from noise, and constructing a sequence of reasoning that can be challenged and verified.

SIAI’s MSc in Artificial Intelligence and Data Science and its MBA programmes in Artificial Intelligence are organised around this transition.

Students are expected not merely to reproduce statistical or machine-learning procedures, but to determine what problem is actually being investigated. They must identify assumptions, select appropriate evidence, construct models, evaluate competing explanations, and explain why their conclusions should be trusted.

Institutional observations suggest that this process can produce a noticeable change in how students approach unfamiliar problems. Students frequently report that, after sustained exposure to the programmes, they begin to search more deliberately for hidden causes, divide complex questions into sequential components, and use statistical and artificial-intelligence tools as instruments of verification rather than as automated answer-generating devices.

These reports should not be presented as controlled scientific evidence that SIAI education produces neurological changes. They are educational observations rather than results from a formal longitudinal experiment.

They nevertheless illustrate how an intellectual environment can alter the habits that students bring to a problem. The most significant change may not be the acquisition of a particular programming language or modelling technique. It may be the movement from asking, “Which method should I apply?” to asking, “What mechanism could have generated what I observe, and how could that explanation be tested?”

Artificial intelligence as a tool of inquiry

This distinction is becoming increasingly important as generative artificial intelligence makes the production of plausible answers inexpensive.

When software can generate summaries, computer code, forecasts, and polished arguments within seconds, the ability to produce an answer is no longer sufficient evidence of expertise. The more valuable capability is the ability to determine whether the answer is logically coherent, empirically grounded, and appropriate to the underlying problem.

Statistical and AI tools should therefore be taught as components of scientific reasoning.

A model can reveal patterns, but the researcher must determine whether those patterns are stable. An algorithm can make predictions, but the analyst must examine selection effects, measurement error, confounding variables, and distributional changes. A language model can construct an argument, but the reader must determine whether the premises are true and whether the conclusion follows from them.

The central educational task is not to train students to compete with machines in producing rapid responses. It is to develop the causal and evaluative reasoning required to direct, test, and correct those machines.

The scientific and institutional lesson

The child-development study offers a scientific warning against interpreting performance without considering context. A tired and stressed child may be judged as less capable when the observed difference is partly a product of the environment in which the child is expected to perform.

The institutional lesson is broader. Intellectual capability is not expressed in a vacuum. It is supported—or constrained—by the environments that organise attention, expectations, feedback, and opportunity.

For children, this means that adequate sleep, reduced chronic stress, safe communities, family support, and well-resourced schools should be regarded as components of cognitive development rather than peripheral social benefits. The authors of an accompanying Science commentary consequently argued that the findings strengthen the case for early, society-level support for families and children.

For universities and professional schools, the challenge is to create an intellectual environment that does more than transfer technical knowledge. It must repeatedly train students to ask better questions, locate hidden assumptions, distinguish evidence from assertion, and search for the causes behind observable outcomes.

A postcode should not determine a child’s intellectual future. Neither should a student’s earlier educational environment permanently determine the limits of their reasoning.

Education cannot eliminate every social disadvantage. It can, however, provide a new environment—one in which intellectual curiosity is expected, causal reasoning is practised, and analytical tools are used not merely to obtain answers, but to understand why those answers should be believed.

That is the educational meaning of Rerum Cognoscere Causas.

Primary reference

Marek, S. A., et al. “Patterns of Brain-Wide Associations Reflect Socioeconomics.” Science, 11 June 2026.

Related commentary

Sisk, L. M., and Satterthwaite, T. D. “Childhood Environments Shape the Brain.” Science, 2026.