For most of the history of neuroscience, the female brain appears in the literature only in contrast to the male brain — measured when both are present, defined by the difference. What happens inside the brain during the reproductive transitions unique to women — puberty, pregnancy, and menopause — has rarely been studied at all, and almost never studied together.

A team of researchers at Amsterdam University Medical Center set out to change that. Led by cognitive neuroscientist Sophie R. van ’t Hof and senior researcher Elseline Hoekzema, they assembled three separate cohorts — 142 girls moving through puberty, 110 women during first or second pregnancy, and 843 women around menopause — and applied the same analytical framework to each. The question across all three groups was identical: how does cortical gray matter volume change, where does it change, and are the patterns shared or distinct?

What they found, published September 19, 2026 in Nature Communications, challenges what textbooks say about the aging female brain. One of the transitions looked almost nothing like the others. And the difference was not what anyone expected.

Three Cohorts, One Framework

The study enrolled 1,095 participants in total, recruited and measured at different points within each transition. For the puberty cohort, 49 girls had not yet reached menarche, 34 were in the transition period, and 59 had crossed it. For pregnancy, the researchers studied 40 women during a first pregnancy, 30 during a second, and 40 nulliparous women as a comparison group. The menopause cohort included 120 women in the perimenopausal transition, 49 premenopausal, and 674 postmenopausal. (van ’t Hof et al., Nature Communications, 2026)

Each participant’s brain was imaged by MRI, and the researchers quantified cortical gray matter volume — the outer layers of the brain where cell bodies are concentrated — across these groups. By comparing volume across transition stages within each cohort, they estimated rates of change and mapped which regions were changing.

The analytical design was built around one central question: which cortical regions behave consistently across reproductive transitions, and which ones diverge? The brain is not a single tissue. It is dozens of functionally distinct regions, and they do not necessarily change together. The study was structured to find out whether, and where, the same pattern shows up in different transitions separated by years and driven by different hormonal contexts.

Puberty: Widespread Reorganization

In the puberty cohort, the analysis found widespread cortical gray matter reduction concentrated in frontal, parietal, and temporal regions — areas that mature late in development and support higher-order cognition, social processing, and executive function. The rate of total gray matter decline across this cohort was approximately 0.13% per month. Cortical gray matter specifically declined at roughly 0.16% per month. (van ’t Hof et al., Nature Communications, 2026)

Those figures are population averages. The range of individual trajectories within the cohort was wide, and a monthly rate calculated from group data cannot predict any individual’s trajectory. What the average describes is the direction and rough pace of change at a group level, not a rule for any particular brain.

This kind of cortical volume reduction during adolescence has been documented in prior research, though rarely in cohorts of this size. It is understood to reflect, at least partly, a process called synaptic pruning — the selective elimination of redundant neural connections in favor of more efficient pathways. The cortex that emerges from puberty is not smaller in the way an atrophied muscle is smaller. It has been reorganized. The molecular and cellular mechanisms underlying the changes remain an active area of investigation, and no single mechanism has been established as the primary driver.

Pregnancy: A Second Remodeling with a Familiar Shape

The pregnancy findings carried a similar overall profile — widespread cortical volume reductions concentrated in frontal, parietal, and temporal areas — but with differences in magnitude and distribution. The monthly rate of cortical loss during pregnancy was approximately 0.12%, somewhat lower than the puberty cohort. (van ’t Hof et al., Nature Communications, 2026)

The spatial overlap between the two transitions was more striking than the rate difference. The researchers identified 34 cortical regions — largely in higher-order association areas — that showed statistically similar patterns of change in both puberty and pregnancy. These regions were involved in both transitions despite the years separating them and the different hormonal and developmental contexts driving each.

A second group of 33 regions showed a stepwise gradient across all three transitions: largest changes in puberty, moderate changes in pregnancy, and minimal changes in menopause. These tended to cluster in sensory and motor areas rather than the association cortex.

One set of regions stood apart: areas associated with emotion processing and bodily awareness that showed substantial changes during puberty but much smaller effects in pregnancy and almost none during menopause. The data suggested that puberty may reorganize certain interoceptive and emotional circuits in ways the subsequent transitions do not replicate — at least not to the same degree or in the same locations.

Thirty-Four Regions That Respond the Same Way Twice

The 34 higher-order regions showing shared patterns across puberty and pregnancy are worth pausing on. These are areas of the cortex associated with cognitive integration, attention, social cognition, and other functions that are not tied to a single sensory input or motor output. They sit in what is sometimes called the association cortex — regions that bring information together across sources.

The fact that these same areas change during two transitions separated by roughly a decade, driven by different hormones and different developmental demands, suggests something about reproductive transitions themselves rather than any specific stage. The brain appears to draw on a consistent set of higher-order regions each time a major hormonal shift occurs.

Whether that consistency reflects a common biological mechanism — such as a shared sensitivity of these regions to estrogen or progesterone — is not established by the current data. The finding is pattern-level: the same regions show up in both transitions. The explanation for why they do so requires future work that would include harmonized hormonal measurements taken at the same time as the brain scans. This study did not include that, and the researchers named it directly as a limitation.

Menopause: What Stays Constant, and Why That Matters

The most counterintuitive finding in the study belongs to the menopause cohort.

For the 843 women in that group — the largest of the three, and the one with the most statistical power — the analysis found no significant cortical gray matter reduction during the menopausal transition. Not a slowed rate. Not a gradual plateau. The typical age-related decline in cortical volume that accumulates across adulthood appeared to pause during the transition itself.

Van ’t Hof described the finding directly: “Menopause is defined not by what changes, but by what does not.” (Nature Communications, 2026)

What makes this striking is the baseline expectation. The menopause cohort spanned exactly the age range when age-related cortical atrophy would normally be visible in neuroimaging data. Instead, the transition period appeared to interrupt it. Whether this reflects a protective hormonal effect, a transient influence on tissue composition such as water content or myelin, or a reorganizational process that produces the pause as a secondary consequence is not something the current data can answer.

The finding is descriptive. The MRI shows the absence of the expected decline. The mechanism behind that absence remains unresolved, and the researchers were explicit that they could not draw causal conclusions from data that did not include harmonized hormonal measurements.

What This Study Cannot Establish

The cohort sizes are substantial for neuroimaging research, where participant recruitment is demanding and imaging time is costly. Applying the same analytical framework across all three transitions, rather than treating them as three unrelated studies, makes the comparisons meaningful. These are genuine strengths.

But the study has constraints that belong beside any reading of the findings.

First, menarche and menopause onset were determined by single self-reported items. Reproductive timing varies considerably across individuals, and errors in self-reported onset will introduce noise into the cohort groupings. The extent to which that noise affects the measured patterns is not quantified in the published analysis.

Second, the three cohorts were scanned on different MRI machines. The researchers applied statistical harmonization to account for scanner differences, but some residual variation from equipment cannot be ruled out. This limits the confidence with which very small between-cohort differences can be interpreted.

Third, and the researchers named this explicitly, the study had no harmonized measure of hormonal levels across the three cohorts. The transitions are understood to be hormonally driven, but the paper does not link specific hormone profiles to specific cortical patterns. The relationship between what the hormones are doing and what the cortex is showing is inferred rather than measured.

Fourth, monthly rate calculations can obscure moments of rapid change. A transition that involves a burst of cortical plasticity followed by stability would produce the same average monthly rate as a continuous gradual change. The two processes would appear identical in this kind of analysis, and the study cannot distinguish between them.

These are not reasons to set aside the findings. They are reasons to read them for what they are: a large, carefully designed comparative study that documents cortical patterns across three reproductive transitions, establishes meaningful overlap between puberty and pregnancy, and surfaces a menopause finding that will need mechanistic follow-up. Population-level data of this kind generates hypotheses. Establishing mechanisms, and determining whether the patterns have behavioral or clinical consequences, is the work of future studies.