Exercise & the Brain: BDNF and Beyond
Exercise is the best-supported behavioral lever for brain health that currently exists — and much of what it is advertised to do sits on weaker evidence than the sales pitch suggests. This page runs the brain claims through a ledger: the trial that grew a hippocampus ~2%, the depression numbers that deflate under honest controls, the hour of acute cognition that is real, and the fixes that aren't.
What the evidence supports
- A one-year RCT found aerobic training increased hippocampal volume ~2% in older adults while stretching controls declined (Erickson et al., PNAS, 2011).
- Exercise performed comparably to sertraline for mild-to-moderate depression in the SMILE trial, and helps at fair-control effect sizes.
- A single session modestly improves attention and processing speed for roughly an hour after moderate exercise (Chang et al., 2012).
What remains uncertain
- Whether exercise prevents dementia — association-grade; no exercise trial has delayed dementia diagnosis.
- Whether blood BDNF says anything about brain BDNF — much of the measured rise comes from platelets.
- Whether adult humans grow meaningful new hippocampal neurons — neurogenesis in people remains contested.
Evidence last reviewed: September 17, 2026. Conclusions may change as new research is published.
the brain page the pillar owed
Product picks are generic categories, not brands. We may earn a commission on Amazon or iHerb purchases at no cost to you — this never changes our evidence conclusions. Full disclosure
Running shoes
The intervention behind every trial on this page — the cheapest brain-health equipment that exists.
⚠️ Fit matters more than price; start below your ambition to protect the habit.
Check price on Amazon →Why the Story Is Credible in the First Place
Before the ledger, the mechanisms — because the exercise–brain story is not invented. Three plausible channels connect moving more to a brain that ages better:
- 🩸 Blood flow: the brain takes roughly a fifth of cardiac output at rest; aerobic training improves vascular function and cerebral perfusion — the same machinery that protects the heart (see VO2 max).
- 🧪 BDNF: brain-derived neurotrophic factor supports neuronal survival and synaptic plasticity, and exercise reliably raises the blood version — though what that rise certifies is a subtler question, handled below.
- 😴 Sleep mediation: part of the benefit likely arrives indirectly, through deeper sleep — the nightly repair machinery does consolidation and clearance work no workout replaces.
Plausibility is not proof — sensible ideas have died in trials before. What earns this page its keep is that some claims survived real randomized testing, and those are where the ledger begins.
The Ledger: Six Brain Claims, Priced Honestly
This table is the page. Every claim the exercise–brain world makes, sorted by what the human evidence showed and the honest ceiling on it — because on this site no benefit is ever presented without its watch-items.
| Claim | Human evidence | The honest ceiling |
|---|---|---|
| 🧠 Hippocampal volume responds to training | Strongest structural result in the field: one-year RCT, 120 adults 60–80, walking grew anterior hippocampus ~2% while stretching controls lost ~1.4% (Erickson et al., PNAS, 2011); a 14-trial meta-analysis finds preserved volume | One landmark trial plus supportive-but-mixed replications; gains concentrate in anterior hippocampus and largely offset control-group decline, not unlimited growth |
| 📉 Depression adjunct | Large pooled effects (SMD ~0.9–1.1) deflate to ~0.4–0.6 against active controls and near-zero in strict head-to-heads (Cooney et al., 2013); SMILE found supervised exercise comparable to sertraline — 45% vs 47% remission (Blumenthal et al., 2007) | An adjunct with a moderate effect, not a treatment replacement; wait-list comparisons inflate the headline |
| ⚡ Acute cognition boost | Meta-analysis of 79 studies: small overall effect (g ≈ 0.10) for attention and executive function after moderate sessions (Chang et al., 2012) | Modest, lab-measured, transient — opens ~10 minutes after stopping, mostly spent within the hour; the first minutes can be slightly negative |
| 🧪 BDNF as the mediator | Acute sessions reliably raise blood BDNF (Huang et al., 2014); higher fitness tracks better cognition in cohorts | Blood BDNF is largely platelet release, not a brain readout; the mediation chain is association-level — plausible mechanism, not documented fact |
| 🛡️ Dementia prevention | Inactivity appears on the Lancet Commission's modifiable-risk list; fitter people show less cognitive decline in cohorts | Association-grade — no exercise trial has delayed dementia diagnosis; population-attributable framing is policy language, not a personal promise |
| 🌱 Neurogenesis in humans | Animal work shows running-driven new hippocampal neurons; some human post-mortem studies find neurogenesis into late life | Contested in humans — other labs find little new-neuron production after childhood; no study has counted new neurons in living people |
How to read this table: the ratings describe the evidence, not your result. Notice the pattern — the claims with randomized support are real but modest, and the biggest promises sit on the weakest rungs. That gradient is the honest story of exercise and the brain.
The Hippocampus Trials: The +2% Story
The hippocampus — the memory-filing structure deep in each temporal lobe — was long treated as a one-way street: it shrinks with age, by roughly 1–2% a year in late life, and nothing you do changes that. In 2011, Kirk Erickson's group tested the assumption directly: 120 sedentary adults aged 60–80, randomized for a year to walking built up to ~40 minutes, three times a week, or to stretching-and-toning control.
The aerobic group's anterior hippocampus grew by about 2% — enough, the authors calculated, to offset roughly one to two years of typical age-related loss — while the stretching controls lost about 1.4%. Larger gains accompanied rising fitness, higher serum BDNF, and better spatial memory. Later trials and a meta-analysis of 14 RCTs are supportive but mixed, with much of the "growth" being held ground — controls declined, and training prevented the decline. A real win against assumed inevitable shrinkage; not a brain getting younger. The hippocampus trials walks through the replication record study by study.
Depression: The Effect Sizes, Deflated Honestly
Here is the claim that sells most: exercise treats depression as well as medication. Here is what the trials show, and how much the answer depends on who the comparison group was. The big meta-analyses report pooled effects around SMD 0.9–1.1 — "large" by convention. But a standardized mean difference is a distance between two groups, and much of that distance is manufactured by the control condition.
The Cochrane review led by Cooney (2013) pooled 35 trials and found a moderate SMD of −0.62 overall — then applied a filter: restricting to trials with adequate allocation concealment and blinded outcome assessment shrank the effect further. Against wait-list controls, who know they received nothing, exercise posts SMD ≈ 0.9–1.1. Against usual care or a placebo pill, ~0.4–0.6. In the few head-to-head trials, exercise was statistically indistinguishable from psychological therapy (SMD −0.03) and close to medication (−0.11) — meaning "comparable," not "better."
The trial behind the headline: SMILE (Blumenthal et al., 2007) randomized 202 adults with major depression to supervised exercise, home exercise, sertraline, or a placebo pill. At four months, remission was 45% supervised exercise, 40% home exercise, 47% sertraline, 31% placebo — exercise and medication statistically indistinguishable. Read the placebo row carefully: nearly a third remitted on expectations, contact, and time alone. Exercise & depression: the effect sizes, honestly carries the full audit.
⚠️ Adjunct, never a substitution
"Comparable to sertraline in one trial" is a statement about averages in mild-to-moderate depression, not a plan to stop your medication — medication changes belong with the prescriber. And if depression has flattened the ability to exercise at all, treat that as data for a clinician, not a discipline failure. If danger feels immediate, contact your local crisis line or emergency services.
The Acute Window: The Hour After
The same-day effect is real and oversold in opposite directions. The 2012 meta-analysis by Chang and colleagues synthesized 79 studies of single sessions followed by cognitive testing: overall effect positive and small (g ≈ 0.10), with a timing moderator that is the practical takeaway. Tests within 0–10 minutes of finishing average slightly negative — you are warm, wired, and briefly worse. The window opens around 11–20 minutes after stopping, peaks for moderate efforts, and is mostly spent within the hour.
Translation for your calendar: a moderate session before deep work is a reasonable experiment, not a pharmacology. Schedule the demanding task 15–45 minutes after cooling down, expect a small edge on attention, and remember the lab-to-real-work transfer is largely untested. The acute session prices the whole window, including its failure modes.
What Exercise Can't Fix
Boundaries are what keep a strong claim strong. Exercise cannot promise to prevent Alzheimer's — the Lancet Commission lists inactivity among modifiable risk factors, legitimate at the population level, but no exercise trial has delayed a dementia diagnosis, and fitter cohorts are not randomly assigned to their fitness. It cannot substitute for psychiatric care. Its BDNF story is a plausible mechanism with a measurement problem — BDNF: what it is, what exercise does to it explains why the blood number doesn't certify the brain number. And neurogenesis — new neurons from running — remains contested in humans.
For the boundaries in full, see What exercise can't fix. The short version: exercise is the best-supported behavioral lever for brain health we have — which is exactly why its claims deserve the same skepticism as any other intervention's.
The Bottom Line
- The structural result is real: a year of moderate aerobic training grew the hippocampus ~2% in older adults — roughly 1–2 years of typical shrink offset — in the field's strongest RCT.
- Depression effects deflate honestly: large pooled numbers (~0.9–1.1) shrink to moderate (~0.4–0.6) against fair controls and to equivalence against medication — an adjunct worth adding, not a treatment to replace.
- The acute boost is a modest hour: small gains on attention, opening ~10 minutes after a moderate session and mostly spent within the hour.
- Respect the boundaries: dementia prevention is association-grade, blood BDNF is a platelet-flavored proxy, and neurogenesis in humans is contested — none of which is a reason to skip the workout.
Go Deeper: Exercise & the Brain
Five subtopics take each piece of this ledger to full depth.
- 🔗 BDNF: what it is, what exercise does to it — the molecule, the acute rise, and the platelet problem. Read it →
- 🔗 The hippocampus trials — the Erickson RCT study by study, and what "held ground" really means. Read it →
- 🔗 Exercise & depression: the effect sizes, honestly — the meta-analyses ranked by strictness, and the SMILE sertraline comparison. Read it →
- 🔗 The acute session: cognition in the hours after — the window's size, peak, and scheduling. Read it →
- 🔗 What exercise can't fix — the boundary list, each limit with its citations. Read it →
Related Topics
- Erickson KI et al., "Exercise training increases size of hippocampus and improves memory," Proceedings of the National Academy of Sciences (2011)
- Cooney GM, Dwan K, Greig CA, et al., "Exercise for depression," Cochrane Database of Systematic Reviews (2013)
- Blumenthal JA et al., "Exercise and pharmacotherapy in the treatment of major depressive disorder," Psychosomatic Medicine (2007)
- Chang YK, Labban JD, Gapin JI, Etnier JL, "The effects of acute exercise on cognitive performance: a meta-analysis," Brain Research (2012)
- Huang T, Larsen KT, Ried-Larsen M, Møller NC, Andersen LB, "The effects of physical activity and exercise on brain-derived neurotrophic factor in healthy humans: a review," Scandinavian Journal of Medicine & Science in Sports (2014)
- Livingston G et al., "Dementia prevention, intervention, and care: 2020 report of the Lancet Commission," The Lancet (2020)