Childhood Adversity and Cognition: What a PNAS Study Found

Does childhood adversity change the brain circuits that support learning and emotion? A study published in PNAS on 10 September 2026 says the difference shows up on brain scans, but the numbers behind the headline are small. The team, from the University of Melbourne and Deakin University, compared people from birth to age 21 and found that more stressful life events went with a more "differentiated" link between the frontal cortex and the amygdala, while higher cognitive ability went with a less differentiated link between the frontal cortex and the hippocampus. The first association was weak (r = 0.122). The second vanished in a second sample (r = 0.015).

What did the PNAS study actually measure?

The paper is "Differentiation of frontolimbic functional connectivity from birth to early adulthood links with adversity exposure and cognition," by Wonyoung Kim, Sarah Whittle, Andrew Zalesky and Ye Ella Tian (PNAS 123(37), doi 10.1073/pnas.2600520123). It is cross-sectional: it compares different people at different ages rather than following the same people over time. The main sample pooled 822 people aged 0 to 21, from the Baby Connectome Project (195 people aged 0 to 5) and the Human Connectome Project in Development (627 people aged 5 to 21). A separate Philadelphia sample of 898 people aged 8 to 23 was used to check whether the results repeat.

Three terms carry the headline. Frontolimbic circuits link the frontal cortex, which handles planning and self-control, with the amygdala (threat and emotion) and the hippocampus (memory and learning). Connectivity here is functional: it describes how closely activity in two regions rises and falls together on a resting-state fMRI scan, which is a statistical relationship and not a count of physical wiring. Differentiation is the authors' term for two circuits developing connectivity patterns that become distinct from each other.

How do these brain circuits develop from birth to adulthood?

The developmental result is the firmest part of the paper. Fronto-hippocampus connectivity develops rapidly in early childhood and then stabilizes. Fronto-amygdala connectivity keeps separating from it, with the biggest change in adolescence. As that happens, the hippocampus becomes preferentially tied to association regions, the areas involved in higher-order thinking, and the amygdala to sensorimotor regions. It is a map of typical development across a wide age range, which is why the title runs "from birth to early adulthood."

Does childhood adversity change the brain? What r = 0.122 means

Adversity was measured as the total number of adverse life events a young person reported over the past year. That is not the same as childhood abuse, neglect or long-term poverty. More events went with a more differentiated fronto-amygdala circuit than expected for that person's age (r = 0.122, p = 0.003). Squared, a correlation of 0.122 means adversity accounts for about 1.5% of the variation in that brain measure. The authors report that the link did not differ by sex, age or family socioeconomic status, and that it held after controlling for socioeconomic status, stress around birth and current stress.

In the replication sample the adversity measure was different, lifetime exposure to traumatic events, and the association was smaller (r = 0.071, p = 0.034). The direction held, but the share of variation explained fell from about 1.5% to about 0.5%.

Is a higher IQ linked to a different brain circuit?

Not in this paper, because it did not measure IQ. Cognitive ability was the NIH Toolbox composite, an age-adjusted score built from tasks covering memory, cognitive flexibility, inhibition, language, processing speed and working memory. It is a research battery, not an IQ test. Higher scores went with a less differentiated fronto-hippocampus circuit (r = -0.138, p = 0.003), about 1.9% of the variation. In the Philadelphia sample the same association was effectively zero (r = 0.015, p = 0.664).

Why do brain-scan headlines about adversity run ahead of the data?

Four features of this design limit what can be said. First, the data are cross-sectional, so a difference at a given age cannot show whether adversity came before the brain pattern or the pattern came first; the authors themselves note that individual differences could exist before adversity or learning. Second, the brain measure is a deviation from the typical age curve, so "more differentiated" means different from same-age peers, not damaged. The abstract describes a difference and does not label it a deficit. Third, the adversity score counts events in the past year, a very different exposure from years of deprivation. Fourth, effects of this size are hard to see in one child: a correlation of 0.12 cannot tell you which child in a classroom has a different circuit.

How does it compare with bigger studies of adversity and the brain?

A larger 2025 PNAS paper by Sofia Carozza, Isaiah Kletenik, Duncan Astle, Lee Schwamm and Amar Dhand used a different brain measure, white-matter structure, in the Adolescent Brain Cognitive Development study of 9,082 children. Children aged 9 and 10 who had been exposed to prenatal risk factors, interpersonal adversity, household economic deprivation or neighborhood adversity had lower fractional anisotropy (a measure of how organized white-matter fibers are) and lower streamline counts across the brain. Lower fractional anisotropy went with later difficulty in mental arithmetic and receptive language, and it statistically accounted for part of the link between adversity and cognition in later adolescence.

Two studies, two brain measures, one consistent message: adversity goes with measurable but modest differences. Neither measured IQ, and neither can prove cause. Our report on cognitive reserve and brain wiring found the same shape in a study of older adults: small coefficients under a large headline word.

Does childhood adversity lower IQ?

Nothing in the frontolimbic paper answers that. The stronger evidence on severe early deprivation and measured IQ comes from designs that can support cause, such as the randomized placement trial of children in Romanian institutions covered in our toxic stress and IQ explainer, and from large observational work on socioeconomic status and IQ. A hard year at home or at school sits at a very different point on that scale from years of institutional care. Childhood scores do carry information: the Scottish 1932 survey, covered in childhood IQ scores and later health, is a famous example. But a brain-scan correlation of 0.12 should never be turned into an IQ estimate for one child.

What should a parent or teacher do with a test score after a hard year?

Start with measurement, not brain circuits. Every full-scale IQ score carries a standard error, commonly around 2 to 3 points on a scale with a mean of 100 and a standard deviation of 15, which puts a 95% interval at roughly 5 to 6 points either side (see our explainer on the margin of error). Sleep, anxiety, practice and test conditions also move a result, so a change of a few points after a stressful period is inside the noise. To see what any score means on the population scale, the IQ percentile calculator turns it into a percentile, and if you want a baseline of your own you can take the IQ Metrics IQ test.

The bottom line: the paper is a good map of how frontal-amygdala and frontal-hippocampal circuits separate between infancy and early adulthood. Its links to adversity and cognition are small, come from cross-sectional scans, and in the cognition case did not repeat. Treat "adversity rewires the brain" as a hypothesis with modest early support, and wait for the longitudinal follow-up the authors say is needed.

Common questions