BDNF is the molecule people name when they say exercise makes you smarter. It is an important protein: it keeps neurons alive, helps them form and strengthen connections, and supports the birth of new cells in the hippocampus, the brain's memory hub. Exercise raises the BDNF measurable in your blood, and BDNF is the leading candidate for how movement protects the brain. The weak point is the last step.
The chain from a rise in blood BDNF to sharper thinking in a person is inferred from animals and correlations, not demonstrated, partly because a blood reading is only a rough proxy for the brain. This page explains the machinery and marks the line between what is solid and what is still inferred. The practices that act on it are linked at the bottom.
Findings & Outcomes
Walking, VO2 max intervals and lifting are all called good for the brain, and the molecule usually named to explain why is BDNF. Several practices in this section point back to it, so this page explains it once: what BDNF is, why movement plausibly reaches the brain through it, and how much of the popular account holds up.
This page recommends nothing. It describes the machinery, and the pages that tell you what to do are linked at the bottom.
What BDNF Is
BDNF, brain-derived neurotrophic factor, is a protein the nervous system makes to support its own cells. It keeps neurons alive, it helps them grow and strengthen the connections between them, the property called synaptic plasticity that underlies learning, and it supports neurogenesis, the birth of new neurons, in the hippocampus, the brain's memory region.
That cellular role is well established. The popular account then jumps from it to a promise: that you can do one thing, raise BDNF, and reliably get a sharper mind. The rest of the page tests that jump.
The strongest evidence that BDNF is part of the causal path, and not a molecule that rises alongside the benefit, comes from rodents. When rats exercise they learn better and their hippocampal plasticity signals climb. When researchers block the action of BDNF in the hippocampus during that exercise, the learning gain disappears. Removing the molecule removes the benefit, which is stronger evidence than a correlation. It is also an experiment that can only be done in an animal, which is why the same step stays inferred in people.
Exercise Raises It
Exercise raises the BDNF you can measure in blood, and this link is well replicated. Pooling the studies, a single bout of exercise produces a moderate rise (Hedges' g about 0.46), and training over weeks increases the size of that post-session response (g about 0.58). The reliable, repeatable phenomenon is the acute spike after a session.
Two qualifications sit next to it. A training program's effect on resting BDNF, the everyday baseline measured away from a workout, is smaller (g about 0.28) and inconsistent across studies, so a lasting shift in day-to-day BDNF stands on weaker ground than the acute rise. And every one of these numbers comes from blood, which is where the difficulty begins.
From a Blood Marker to Actual Thinking
What we can sample easily is BDNF in blood. What matters is BDNF doing its work in the brain. Those are two different readings.
Blood and brain BDNF move together enough that blood is used as a proxy: measured across rats, pigs and mice, the two correlate. But the correlation is partial, not one-to-one, and blood BDNF has its own biology. A large fraction of it is stored in platelets rather than freshly released from neurons, and serum and plasma give different values. A rise in blood BDNF carries information about the brain while staying an indirect, noisy signal.
So the sentence "exercise raises BDNF, therefore exercise makes you smarter" is weaker than it sounds. Each link holds: exercise raises blood BDNF, blood BDNF partly tracks brain BDNF, brain BDNF supports plasticity and memory. A chain of partial links is itself only partial, and the final step, from a measured BDNF change to a measured improvement in a person's thinking, is inferred from the mechanism, not shown directly.
Exercise raises the BDNF in your blood. That the rise reaches the brain and sharpens thinking is inferred from the mechanism, not measured in people. BDNF is the best candidate for the link between movement and the brain; it is not a proven mediator in humans.
The Hippocampus Result, and What Replication Did to It
The most cited human evidence is a one-year trial in older adults. A moderate walking program increased the volume of the anterior hippocampus by about 2%, against a decline in the control group, which reverses one to two years of normal age-related shrinkage, and it improved spatial memory. Within the exercise group, the people whose hippocampus grew most tended to be the ones whose serum BDNF rose most. That single trial made the BDNF story famous, and it connects the animal mechanism to a measurable human outcome.
Later trials tried to repeat it. Pooling them, across 737 participants, aerobic exercise did not reliably increase total hippocampal volume. What survived was narrower: a possible benefit to the left hippocampus, and a plausible role in slowing the age-related decline of the structure. Preserving tissue and enlarging it are two different claims, and only the first holds up across studies. This is the correction the popular version skips. The supported claim is that exercise may help an aging brain hold on to what it has, not that exercise grows the brain.
Individual Variation
People carry different versions of the BDNF gene, and the difference is measurable. A common variant called Val66Met reduces the activity-dependent release of BDNF, the on-demand secretion that happens when neurons are busy, and carrying it is associated with slightly poorer episodic memory and altered hippocampal engagement on brain imaging. This describes variation between people. There is no established reason to genotype yourself for it, and it points to nothing you can act on. It matters for one reason: it is part of why the same amount of exercise does not produce the same response in everyone, and why group averages hide a wide spread.
Mood, and the Muscle-to-Brain Signal
Two further threads, both promising and both early.
Low BDNF is a recurring finding in depression, and exercise is an effective treatment for it, which makes BDNF an obvious candidate for how movement lifts mood. Trials have pooled resting BDNF before and after exercise programs in people with major depression to test exactly this. The same gap applies as for cognition: even where mood and BDNF both move in the right direction, that does not show the BDNF change caused the mood change, because sleep, activation and treatment response all shift together. The mediation is a reasonable hypothesis under active study, not a settled fact.
The second thread is a specific route from working muscle to the brain. In mice, running raises a muscle-secreted protein called cathepsin B that crosses into the brain, and it is required for exercise-induced hippocampal neurogenesis and memory and raises BDNF in neurons. In a small human sample, fitness and cathepsin B rose together. It is a concrete candidate for the muscle-to-brain link, established in mice and only correlational in the few people studied so far. The wider idea that muscle signals to the brain is covered on muscle as an organ.
What Is Solid and What Is Inferred
The chain's strength is uneven along its length.
Solid:
- BDNF supports neuron survival, synaptic plasticity and hippocampal neurogenesis.
- Exercise raises blood BDNF acutely.
- In rodents, BDNF is necessary for exercise to improve learning.
Inferred, and marked as such in the findings below:
- that the blood BDNF we measure faithfully reflects the brain,
- that a training program durably lifts resting BDNF,
- above all, that raising BDNF is what turns exercise into better thinking or mood in a person.
The mechanism is established and important. The instruction "do X to boost BDNF and get smarter" runs ahead of the human evidence for that final step.
Evidence
The Research & Studies
Everything here is based on the research we have collected and checked, sorted into groups and ordered with the strongest evidence first. Click any claim to open the studies behind it.
How it works
A single workout raises blood BDNF, and training makes the rise bigger
A single workout reliably bumps up the BDNF measurable in your blood, and training over time makes that bump bigger. The rise after one session is the clearest part; the change in resting level is smaller.
This meta-analysis pooled studies measuring peripheral brain-derived neurotrophic factor around exercise in people. A single bout of exercise produced a moderate increase in circulating BDNF. Regular training increased the magnitude of the BDNF response to a subsequent session, and had a smaller effect on resting (baseline) concentrations. The analysis is of a blood marker, not brain tissue, and heterogeneity across the included studies was substantial, so the pooled estimate describes a direction and rough size rather than a precise number that transfers to any one protocol.
The study · 1
Szuhany, Bugatti and Otto, a meta-analytic review of the effects of exercise on brain-derived neurotrophic factor · J Psychiatr Res 2015
In rats, blocking hippocampal BDNF erased the learning gains from exercise
When researchers ran rats and then chemically blocked BDNF in the memory part of the brain, the exercise no longer improved learning. That is the clearest sign BDNF is actually doing the work, not just rising alongside it.
Rats that exercised showed better spatial learning and higher levels of plasticity-related proteins in the hippocampus. When the action of BDNF was blocked in the hippocampus during the exercise period, using a molecule that intercepts it before it reaches its receptor, those exercise-induced improvements in learning and in the downstream plasticity signalling were abolished. This is a loss-of-function design: it shows BDNF is necessary for the effect in this animal model, which is stronger causal evidence than a correlation, but it is a rat study and the intervention is one a human experiment cannot replicate.
The study · 1
Vaynman, Ying and Gomez-Pinilla, hippocampal BDNF mediates the efficacy of exercise on synaptic plasticity and cognition · Eur J Neurosci 2004
Blood BDNF only partly reflects brain BDNF, so it is a rough proxy
Blood BDNF is used as a stand-in for what is happening in the brain because the two track each other across animals. It is only a rough stand-in: much of the BDNF in blood comes from platelets, so a blood reading is not a clean window into the brain.
This study measured BDNF in blood and in brain tissue across rats, pigs and mice and found that blood BDNF concentrations correlate with brain-tissue BDNF, which is the main reason peripheral BDNF is treated as a usable proxy. The correlation is partial rather than one-to-one, and peripheral BDNF has its own biology: a large fraction circulates stored in platelets, measured values differ between serum and plasma, and assay methods vary. So a blood BDNF number carries real information about the brain but is a noisy indirect signal, which is the measurement weak point under the whole exercise-to-cognition story.
The study · 1
Klein et al., blood BDNF concentrations reflect brain-tissue BDNF levels across species · Int J Neuropsychopharmacol 2011
Aerobic training raised resting BDNF a small amount; resistance training did not
A single workout clearly raises BDNF for a while. Whether weeks of training move your everyday resting level much is far less clear, and the pooled evidence for a lasting change is weak.
This meta-analysis of exercise training studies assessed resting peripheral BDNF, the baseline level away from an acute session. Aerobic training raised resting BDNF (SMD about 0.66) and the overall pooled effect was about 0.39; resistance training alone did not significantly change it. The effect is smaller and more variable than the reliable acute post-exercise rise, but this review did find a durable aerobic increase rather than nothing.
The study · 1
Dinoff et al., the effect of exercise training on resting concentrations of peripheral BDNF: a meta-analysis · PLoS One 2016
In mice, muscle-made cathepsin B is needed for exercise to grow new memory cells and raise BDNF
Running makes muscle release a protein, cathepsin B, that reaches the brain, and in mice this protein is needed for exercise to grow new memory-cells and to raise BDNF. A small human study saw fitness and this protein rise together.
Exercise increased secretion of the protein cathepsin B from muscle. In mice, cathepsin B was necessary for running-induced adult hippocampal neurogenesis and for the associated improvement in spatial memory, and applying cathepsin B to neurons raised BDNF expression. In a small human cohort, treadmill fitness and plasma cathepsin B rose together with exercise, and cathepsin B changes correlated with a memory measure. This is a specific candidate for the muscle-to-brain link behind BDNF, but the causal chain is established in mice, and the human evidence is a small correlation rather than a controlled outcome.
The study · 1
Moon et al., running-induced systemic cathepsin B secretion is associated with memory function · Cell Metab 2016
Cognition
A year of walking grew the hippocampus about 2% and improved memory in older adults
Older adults who walked regularly for a year grew the memory-related part of the brain by about 2%, reversing a year or two of normal shrinkage, and remembered better. The people whose BDNF rose most tended to be the ones whose hippocampus grew.
This randomized controlled trial assigned older adults to a moderate aerobic walking program or a stretching-and-toning control for one year. The aerobic group increased anterior hippocampal volume by roughly 2%, against a decline in the control group, effectively reversing the age-related loss expected over one to two years, and improved on a spatial memory task. Increased hippocampal volume was associated with increased serum BDNF within the exercise group, which is the human observation that connects the animal mechanism to a measurable brain-structure and memory outcome. It is a single trial, the BDNF link is a within-group correlation rather than a demonstrated mediator, and the memory gains were specific rather than global.
The study · 1
Erickson et al., exercise training increases size of hippocampus and improves memory · PNAS 2011
The common Val66Met variant lowers BDNF release and tracks with slightly weaker memory
People carry different versions of the BDNF gene. One common version, called Val66Met, releases less BDNF when neurons are active and is linked to slightly weaker episodic memory. It is one reason a given amount of exercise does not do the same thing for everyone.
This work characterized the BDNF Val66Met polymorphism, a single common variant in the BDNF gene. In cell studies, the Met form showed impaired activity-dependent secretion and abnormal intracellular trafficking of BDNF. In people, carrying the Met allele was associated with poorer performance on episodic memory tasks and with altered hippocampal engagement on functional imaging. It is a source of individual variation rather than a benefit or harm of any practice: it helps explain why the BDNF response, and its downstream effects, differ between people, and why group averages hide real individual spread.
The study · 1
Pooled trials in 737 people found no gain in total hippocampal volume, only the left side preserved
When later trials were combined, exercise did not clearly grow the hippocampus overall the way the first famous study suggested. What held up was smaller: exercise may help preserve part of it rather than enlarge it.
This systematic review and meta-analysis pooled controlled trials measuring hippocampal volume after aerobic exercise. Across the pooled data, spanning 737 participants, aerobic exercise did not significantly increase total hippocampal volume. A more specific effect on left hippocampal volume was reported, and the authors framed the likely benefit as attenuating the age-related decline in hippocampal volume rather than producing net growth. This is the correction to the single-trial result: the direction of benefit is plausible and partly preserved, but the strong claim that exercise grows the hippocampus does not survive replication as stated.
The study · 1
Firth et al., effect of aerobic exercise on hippocampal volume in humans: a systematic review and meta-analysis · NeuroImage 2018
Mood & stress
Whether raising BDNF is how exercise lifts depression is proposed, not shown
Exercise helps depression, and low BDNF is part of the leading biological story for why depression happens, so raising BDNF is an appealing explanation. Whether the BDNF change is actually what lifts mood has not been shown, only inferred.
Low BDNF is a recurring finding in major depression, and exercise is an effective treatment for depression, which makes BDNF an attractive candidate mediator. This meta-analysis examined resting peripheral BDNF before and after exercise interventions in people with major depressive disorder. Even where BDNF and mood both move in the right direction, that co-movement does not demonstrate that the BDNF change caused the mood change: antidepressant response, sleep, and general activation all shift together, and blood BDNF is an indirect signal. The mediation hypothesis is reasonable and actively studied; it is not settled.
The study · 1
Dinoff, Herrmann and Lanctot, the effect of exercise on resting concentrations of peripheral BDNF in major depressive disorder: a meta-analysis · J Psychiatr Res 2018
Two notes on reading these findings. Several are graded mechanistic or animal, not as a trial,
because you cannot block or measure BDNF inside a living human brain on demand. And the grades are
deliberately uneven: the exercise-raises-BDNF and the animal-mechanism findings rest on firmer
ground, while the blood-to-cognition, resting-BDNF and mood findings are emerging and are marked as
such.
Go Deeper
This page describes the machinery. The practices that act on it:
- Walking, the most accessible aerobic stimulus and the one behind the older-adult hippocampus trial.
- VO2 max intervals, higher-intensity aerobic work that produces a larger acute BDNF response.
- Resistance training, which also engages the muscle-to-brain signaling around exercise.
- Muscle as an organ, where the cathepsin B route from working muscle to the brain fits the wider picture, and mitochondria and the stress axis, the other biology behind why movement changes how you feel.
Common Questions
What is BDNF?
BDNF, brain-derived neurotrophic factor, is a protein the nervous system makes to support its own cells. It keeps neurons alive, helps them form and strengthen connections (synaptic plasticity), and supports the birth of new neurons in the hippocampus, the brain's memory hub. It is the molecule most often named when people explain why exercise is good for the brain.
Does exercise really raise BDNF?
Yes, this part is well replicated. A single session of exercise produces a moderate rise in the BDNF measured in blood, and training over weeks increases the size of that response. The weaker, less consistent claim is that a training program durably raises your resting, day-to-day BDNF level, which the pooled evidence does not clearly support.
So does raising BDNF make you smarter?
That is the step where the story runs ahead of the evidence. Each link holds: exercise raises blood BDNF, blood BDNF partly tracks brain BDNF, and brain BDNF supports memory. But the final link, from a measured BDNF change to sharper thinking in a person, is inferred from the mechanism, not shown directly. It does not help that blood BDNF is only a rough proxy for what is happening in the brain, since much of it is stored in platelets. BDNF is the best candidate for the link; it is not a proven one in humans.
Didn't a study show exercise grows the hippocampus?
One well-known one-year trial in older adults did: a walking program increased hippocampal volume by about 2% and improved memory, reversing a year or two of age-related shrinkage. When later trials were pooled, exercise did not reliably grow total hippocampal volume. What held up was gentler, a likely role in slowing the age-related decline of the structure rather than enlarging it. That still matters; it is just a smaller claim than the famous result.
Should I get tested for the BDNF Val66Met gene?
There is no established reason to. Val66Met is a common variant that lowers the on-demand release of BDNF and is linked to slightly weaker episodic memory, but it describes variation between people, not something you can act on. It is useful mainly for understanding why the same amount of exercise affects different people differently, not as a test to seek out.
The Chinese Medicine Reading
Explore Related
Other pages this one connects to, by the evidence they share, the outcomes they touch, and the ground they cover.
All 9 sources on this page independently checked and cross-referenced.
Thomas Dehli, Founder & Editor, Sacred Lotus
Sacred Lotus has published Chinese medicine reference material since 2001. Integrative pages are held to the same standard as the herb and formula library: cite the source, grade the claim at its real strength, and say where the research has not looked. This page is educational and it is not medical advice. Last reviewed and updated August 10, 2026.
Evidence strength
How confidently the research supports a claim. Strength describes the evidence, not our endorsement.