There's a story we tell ourselves about exercise. It goes like this: we used to be hunter-gatherers walking twenty kilometers a day, and now we sit at desks, so we go to the gym for forty-five minutes to make up the difference.
The story is half right. We did walk a lot, and most of us don't anymore, and the gym is better than nothing.
But it's missing something. The kind of movement our ancestors did wasn't just more — it was different. And a growing body of neuroscience research suggests that the cognitive component of that movement might be doing something to the brain that the modern treadmill version doesn't quite replicate.
I want to walk through what's actually established here, because there are some real findings, and there are also some claims floating around that go further than the evidence supports.
The discovery that broke the rules
For most of the twentieth century, the consensus in neuroscience was that adult mammals — including humans — couldn't grow new neurons. You were born with what you had, lost some over time, and that was the inventory.
That consensus came apart in the 1990s. A series of studies, building on earlier work by Joseph Altman in the 1960s and Fernando Nottebohm in songbirds, established that adult brains do produce new neurons in at least one region: the dentate gyrus of the hippocampus. The hippocampus is the brain area most directly involved in forming new memories.
The breakthrough study in mice, by Henriette van Praag and Fred Gage, showed that running specifically — not just any cage activity — dramatically increased the production of new neurons in this region. Mice with running wheels grew new hippocampal cells at roughly twice the rate of sedentary mice.
That finding was striking enough on its own. But the next decade of research kept extending it. The mechanism turned out to involve a protein called brain-derived neurotrophic factor — BDNF — which acts something like fertilizer for neurons. Exercise raises BDNF. Higher BDNF supports the survival of new neurons and the strengthening of existing connections.
In humans, you can't easily count new neurons, but you can measure proxies. Aerobic exercise reliably raises circulating BDNF. It's associated with larger hippocampal volume. And in the most-cited human trial in this area — Erickson and colleagues, 2011 — older adults who walked three times a week for a year showed about a 2% increase in hippocampal volume, while a control group showed the typical age-related shrinkage.
That study was real, and well-controlled, and it changed how a lot of clinicians talk about exercise and aging. It's also worth noting that 2% is meaningful but not enormous. The effect is real; the magnitude is modest.
Where the evolutionary story comes in
The puzzle that bothered neuroscientists was: why? Why would exercise specifically — as opposed to, say, sleep or food intake — affect the brain at all? Most organs care about exercise. The brain isn't a muscle. Why does it respond?
A hypothesis put forward by anthropologist Daniel Lieberman, biologist David Raichlen, and others is that this is a residue of how human cognition and human movement evolved together.
For most of our evolutionary history, the things our ancestors did with their bodies were also cognitively demanding. Walking long distances meant tracking landmarks, remembering routes, monitoring the environment for threat and opportunity. Hunting meant pattern recognition, prediction, sustained attention, complex coordination. Foraging meant spatial memory and decision-making under uncertainty.
The hypothesis is that the brain's coupling between movement and growth was shaped under conditions where movement and cognition were always linked. The neurons that get supported by the BDNF response are largely in regions involved in spatial navigation, memory, and learning. That's not a coincidence; that's what the system was built around.
Some research has tried to test this directly. Studies comparing complex movement — dance, martial arts, navigation through varied terrain — to simple cardio find that the cognitively demanding versions produce somewhat larger benefits to memory and executive function. The effect sizes are modest, the studies are not all consistent, and the field is still working out the details. But the general pattern holds.
A useful framing: cardio raises BDNF. Cognitive challenge during movement raises BDNF and provides the kind of activity that BDNF most usefully supports.
The treadmill problem
Here's the part that's worth taking seriously without overstating.
A lot of modern exercise has been engineered to be cognitively easy. The treadmill is the cleanest example — fixed surface, controlled speed, often paired with a screen. The Peloton is similar. Most weight machines isolate one movement and remove the balance and coordination demands that the same exercise would have if you were doing it outside or with free weights.
This isn't bad. Treadmill running still raises BDNF. Cycling on a stationary bike still improves cardiovascular health. The point isn't that the gym is broken. The point is that movement that's free of cognitive demand may be leaving some brain benefit on the table.
Studies on this aren't unanimous, and the differences between, say, trail running and treadmill running aren't huge in any single study. But the general direction is suggestive enough that it's worth taking the practical lesson on board.
What this looks like in practice
You don't have to overhaul your routine. A few targeted shifts seem to capture most of the available benefit.
**Move outside sometimes.** Uneven ground, navigation, weather, and the constantly shifting visual environment all add cognitive load. A trail walk has more brain content than the same distance on a treadmill.
**Vary the movement.** Strength work that involves balance and coordination — kettlebells, single-leg variations, complex compound lifts — engages more of the system than fixed-machine work. Same with sports. Pickleball, tennis, climbing, dance, martial arts — anything that requires real-time decisions while you're moving.
**Don't zone out.** The Netflix-on-treadmill version is the most cognitively passive form of exercise possible. If you want the brain effect alongside the body effect, leave the screen off some of the time and let the movement be the thing.
**Add navigation.** If you walk regularly, occasionally walk a route you don't know well, without GPS. Spatial memory is a hippocampus function. Use it.
**Keep the cardiovascular base.** None of the cognitive-load work substitutes for sustained aerobic activity. The two stack.
What I wouldn't claim
I want to be careful with a couple of things that get oversold in this area.
The phrase "your brain is starving" makes for good headlines, but it's not literally true. Sedentary brains aren't starving. They're not getting an input that supports growth and plasticity, which is a meaningful health issue, but the framing matters.
I also wouldn't claim that exercise alone prevents cognitive decline. The evidence base on exercise and dementia risk is real and consistent — physically active people have lower dementia rates — but the effect operates alongside many other factors, including cardiovascular health, sleep, social connection, and education. Exercise is one of the strongest single levers we know about for brain aging. It isn't the only one.
And the specific claim that we evolved to do cognitively challenging exercise — while consistent with the data — is still a hypothesis. It's plausible. The mechanistic story (BDNF, hippocampus, neurogenesis) is well-supported. The evolutionary explanation for why that mechanism exists is reasonable but harder to prove directly.
What this changes if you let it
The point of all this isn't that your treadmill workout is wasted. It's that movement is one of the most powerful inputs your brain receives, and the version of movement that humans actually evolved with — varied, navigated, cognitively engaged, often outside — may carry a little more signal than the version that fits cleanly into a forty-five-minute window between meetings.
This pairs with most of the other things we know about brain aging. The brain wants novelty. It wants the body to do real work in a real environment. It wants problems to solve. The decline curve we see in the population isn't an inevitability; it's largely the curve of a system that stops getting the inputs it was built to receive.
Move your body. When you can, move it somewhere that requires you to think while you do it.
That's most of the playbook.
---
*Pairs well with: "What People Over 90 Actually Do" and "Cognitive Reserve Is Compound Interest for Your Brain."*