Three months of online exercise classes lifted mood in midlife but left the memory test unchanged

The claim that exercise grows new brain cells has become one of those things everyone half knows. It rests largely on rodent work, because you cannot look inside a living person’s hippocampus to count new neurons. A trial of 52 middle-aged adults in the UK, published today in npj Aging, tried to get at the question a different way, and its headline result is a negative one.

Three months of online exercise classes did not improve the memory ability the researchers had set out to test. Mood did improve, and something in the participants’ blood changed how human brain stem cells behaved in a dish.

Why test exercise in midlife specifically?

Middle age is when many people first notice memory slips and when the risk factors for later dementia start to accumulate. The authors argue it represents a critical window for mental and cognitive change, yet remains relatively understudied as a target for preventative lifestyle programs.

Most exercise-and-brain trials have focused on older adults, often after problems appear. Testing earlier is harder, because healthy 50-year-olds have less room to improve, and any benefit takes longer to show up.

What is pattern separation, and why test it in midlife?

Pattern separation is the ability to tell similar things apart: which of two near-identical images you saw before, or where you parked today as opposed to yesterday. It is one of the few human abilities thought to depend specifically on new neurons in the hippocampus, the brain’s memory hub.

That makes it the natural choice if you want a behavioral readout of neurogenesis in people. It was the trial’s primary outcome, the measure chosen in advance as the test of success.

What happened in the three months of classes?

The team recruited 52 healthy middle-aged adults and randomly assigned them to a three-month exercise program or a control condition. The program was structured, instructor-led and delivered live online, which the authors report was feasible for this age group.

On the primary outcome, pattern separation, there was no effect. The paper reports that no significant intervention-by-time effect was observed for the primary outcome, which in plain terms means the exercise group did not improve more than the control group.

Two secondary measures did shift. Depressive symptoms, measured with the PHQ-9 questionnaire, fell in the exercise group. And in the lab, when participants’ blood serum was applied to human hippocampal progenitor cells, the share of cells that were actively dividing changed.

What does blood applied to brain cells actually show?

This is the clever part of the design and also the part most easily oversold. Because neurogenesis cannot be measured in a living person, the researchers used what they call an in vitro parabiosis assay: they took serum from participants before and after the program and grew human hippocampal progenitor cells in it, watching how the cells responded.

A change in those cells means something in the blood changed in a way that these brain stem cells can detect. The authors read it as exercise remodeling the circulating environment in ways that influence hippocampal progenitor biology. What it does not show is that new neurons appeared in anyone’s brain, or that the direction of the change is good: the proportion of dividing cells went down, not up, and how that maps onto neurogenesis in a living brain is unresolved.

How much weight can a 52-person exercise trial carry?

Not much on its own, and the authors frame the trial as showing feasibility as much as effect. With this many participants, a real but modest cognitive benefit could easily be missed, and secondary findings like the mood result are the kind that need confirming in a larger trial designed to test them.

There is also the publisher’s page, which was not accessible for this article, so the full limitations section was not reviewed. The mood measure is a self-report questionnaire in people who knew which group they were in, which is a known weakness of exercise trials: you cannot blind someone to whether they are exercising.

So is exercise still worth it for your brain?

The honest reading is that this trial supports exercise for mood in midlife and does not yet support it for memory, at least not after three months. The authors suggest cognitive benefits may require longer exposure to emerge, which is plausible and untested here.

None of that argues against exercising. MedlinePlus calls regular exercise one of the best things you can do for your health, and notes that cardio exercises keep your heart, lungs and circulatory system healthy while strength training builds stronger muscles. If you have been told exercise will grow you new brain cells, this trial is a reminder that the human evidence is thinner than the slogan.