Are Cognitive Changes in Perimenopause Temporary?

Say you have already decided the fog is real. You lost a word in a meeting last month, or you reread the same email three times before it sank in, and you stopped arguing with yourself about whether it counts. The question that comes next is quieter and, I think, harder: will this pass?

I wrote about whether the experience shows up on tests in a separate piece. This one asks only about the clock. The only studies that can speak to "temporary" are the ones that tested the same women again and again over many years, so those are the ones I read.

What the long-running studies found

The best known is SWAN, the Study of Women's Health Across the Nation, a U.S. cohort that tested women across the menopause transition. A 2009 paper in Neurology analyzed 2,362 of its participants, women aged 42 to 52 when they joined, who were tested repeatedly over four years. The pattern was subtle. On a processing speed test, women who were premenopausal, early perimenopausal, or postmenopausal got better with repeated practice, which is what usually happens on a test you take more than once. Women in late perimenopause did not. Delayed word recall followed a similar shape: gains before and after the transition, not during it. Immediate recall and working memory did not change by stage. The authors wrote that these difficulties "may be time-limited," because performance recovered after menopause.

It helps to be exact about what that says. Nobody's scores fell off a cliff. What went missing was the small improvement people normally get from practice. The authors also named limits themselves: a short battery of tests, no adjustment for running many comparisons, and a ceiling problem: many women were already scoring near the top of the memory test.

A later SWAN summary, a 2019 progress report in Menopause written by members of the study team, drawing on the same cohort (about 3,300 women aged 42 to 52 at enrollment), adds that hot flashes and poor sleep did not explain the late perimenopause dip, and that adjusting for depressive symptoms and anxiety left it unchanged. That is a narrative review of the group's own work, not a new trial, so I treat it as a useful summary, not independent confirmation.

Where the story gets less tidy

A second long study did not tell the same tale. The Penn Ovarian Aging Study followed 403 women who were 35 to 47 at enrollment, with yearly testing for 14 years (a 2013 paper in the Journal of Clinical Endocrinology and Metabolism). After adjusting for age, both immediate and delayed verbal recall declined across the transition. Delayed recall dropped early in the transition. For immediate recall the authors reported a decline but no rebound afterward. Delayed recall, though, was not significantly lower after menopause than before it, and the authors said they could not tell whether the decline levels off or whether there were simply too few postmenopausal women and test sessions to see it.

Then there is the question of what "recovery" would even be measured against. A 2017 SWAN analysis in PLoS One skipped the first two testing visits, on purpose, to reduce practice effects, and followed 2,124 women (average age 54 at its start) for a median of 6.5 years. Processing speed declined by an estimated 4.9 percent over 10 years and delayed recall by about 2 percent. Immediate recall and working memory showed no significant decline. The authors read this as cognitive aging in midlife. The 2019 SWAN progress report goes a step further and says the declines track chronological age, not menopause itself. Both can be true at once: a dip around the transition that eases, and a slow drift with age that was never about the transition at all.

A 2023 analysis of UK Biobank data, with 15,486 women (average age about 52) followed for about eight years, found something different again. Slowing in reaction time was less pronounced in postmenopausal women than in premenopausal ones. Perimenopausal and postmenopausal women scored lower than premenopausal women on memory and fluid intelligence, but their rates of change were similar. Menopausal status there was self-reported, which the authors flag as a source of misclassification.

And a 2013 meta-analysis in the Journal of Steroid Biochemistry and Molecular Biology pooled four cross-sectional studies, 1,993 women combined, in midlife and grouped by menopausal stage. It found no significant difference in cognition between perimenopausal and premenopausal women, but postmenopausal women did worse on delayed verbal memory than premenopausal women. Four studies is a small pile, and a snapshot of different women at different stages cannot show whether the same woman recovers.

What I can and cannot say from this

If I lay these side by side, here is what I can say. The SWAN 2009 data suggest that a small dip in learning from repeated testing around late perimenopause was temporary in that cohort. The Penn study found some verbal recall decline that did not clearly rebound, with no way to tell whether it stabilized. The later aging analyses suggest that a gradual drift continues with age regardless of stage. These studies used different tests, different recruitment, different ways of defining menopause stages, and different ages at the start, so I would not stack their numbers on top of each other.

Most of what is known also comes from tests lasting minutes, not from the everyday experience of searching for a word. Almost all of it is observational, so it can describe patterns and cannot say why. And the cohorts differ in who was included (the Penn study found race was associated with most of its outcomes), so a woman reading this should hold her own situation loosely. I did not find a study that followed women long enough, with enough postmenopausal testing, to settle the question of full recovery. I cannot rule out that one exists.

What this can look like in practice

As far as I can tell, the answer to "will it pass" is probably some of it, in the sense that the specific dip researchers saw around late perimenopause looked time-limited in the SWAN cohort, and no study I read described a collapse. Beyond that, nobody can promise a particular woman a particular outcome.

What I can offer is a small habit I find useful when I am trying to read my own days honestly. Once a day I jot down a line or two: how my thinking felt, how I slept. A few weeks later I look back. Nothing scientific about it. It just gives me something steadier than my mood on a bad Thursday, and something concrete to bring to a doctor if I ever decide to.

If you would like a ready-made place to do that, I made the Daily Check-In Journal for Women Over 40, a printable PDF built around one short page a day, not a productivity system. It is a notebook, nothing more. I cannot say it will change how your memory or thinking works, and an ordinary notebook you already own does the same job.

What not to do

I would not read one slow day, or even a slow stretch of weeks, as a sign of something permanent. Nor would I take the studies above as a promise that it will all pass. A change that keeps getting worse, or gets in the way of work or daily life, is different. That one is a question for a doctor, who can see your whole picture. A research roundup cannot.