BDNF appears in nootropic marketing and cognitive-aging discourse more than almost any other endogenous protein. The framing collapses several distinct claims: that BDNF is a master regulator of plasticity (mostly true), that interventions reliably raise brain BDNF (partially true, with a serum-vs-brain measurement gap), and that raising BDNF translates predictably to cognitive gain (partially true, dose- and population-dependent). This page lays out what BDNF actually does, which interventions move it on what timescale, and the methodological limits of the human evidence.
What BDNF does
BDNF is a 27 kDa secreted protein, member of the neurotrophin family alongside NGF, NT-3, and NT-4/5. It binds two distinct receptors: TrkB (high-affinity, drives plasticity-promoting signaling) and p75NTR (low-affinity, can promote apoptosis depending on context). The TrkB pathway activates PI3K/Akt, MAPK/ERK, and PLC-gamma cascades, which together drive dendritic spine growth, AMPA receptor trafficking, long-term potentiation (LTP), and adult neurogenesis in the dentate gyrus.
In the adult human brain, BDNF is most highly expressed in the hippocampus, prefrontal cortex, amygdala, and cerebellum. Hippocampal BDNF levels correlate with spatial memory and contextual learning across rodent studies and emerging human imaging work. Genetic knockout of Bdnf in mice produces severe deficits in fear conditioning, spatial learning, and adult neurogenesis. The BDNF Val66Met polymorphism (rs6265, a common human variant) reduces activity-dependent BDNF secretion by 18 to 30% in heterozygotes and is associated with smaller hippocampal volume, lower episodic memory performance, and accelerated cognitive decline in some elderly cohorts.
The plasticity case is strong enough that BDNF is sometimes framed as "Miracle-Gro for the brain". The framing oversells: BDNF is necessary for plasticity but not the only relevant factor (IGF-1, VEGF, and FGF-2 all contribute), and acute BDNF elevation does not always translate to measurable cognitive change. The gap between mechanism and outcome is wide enough that careful trial design matters.
The exercise-BDNF axis
The cleanest intervention signal for BDNF is aerobic exercise. Szuhany 2015 meta-analyzed 32 studies of acute exercise effects on serum BDNF and found a robust pooled effect ( Szuhany, Bugatti & Otto 2015 ). Acute moderate-to-vigorous aerobic exercise (typically 30 to 40 minutes at 65 to 85% max heart rate) raised serum BDNF roughly 2 to 3 fold within 30 minutes post-session, with normalization to baseline by 60 to 90 minutes. Resistance training raised serum BDNF less reliably and at smaller magnitudes; the effect size for acute resistance was about one-third that of acute aerobic.
Chronic training raised resting (pre-exercise) BDNF more modestly: roughly 10 to 20% above baseline after 8 to 12 weeks of regular aerobic training, with the effect dependent on intensity and consistency. The acute spike does not appear to fully attenuate with training; trained athletes still show 2 to 3 fold acute increases, suggesting the system does not desensitize.
Erickson 2011 connected the molecular signal to a structural endpoint ( Erickson et al. 2011, n=120 ). 120 adults aged 55 to 80 were randomized to 12 months of moderate aerobic training (40 minutes 3 times per week at 60 to 75% max heart rate) versus stretching/toning control. The training group gained about 2% hippocampal volume; the control group lost about 1.4%. The effective volume difference was roughly 3.4%, comparable to reversing 1 to 2 years of typical age-related hippocampal atrophy. Serum BDNF rose 6 to 8% in the training group and partially mediated the volume change in path analysis.
The Erickson trial remains one of the strongest pieces of evidence that exercise-induced BDNF translates into structural CNS change. The dose was modest (around 120 minutes per week of moderate intensity), the duration was a year, and the effect appeared on MRI rather than on a serum proxy alone.
Ketosis and the BHB-HDAC-BDNF axis
The mechanism by which exercise raises BDNF transcription was incompletely understood until the Sleiman 2016 eLife paper ( Sleiman et al. 2016 ). Sleiman showed that beta-hydroxybutyrate (BHB), the dominant ketone body produced during exercise and fasting, inhibits histone deacetylases 2 and 3 (HDAC2/3) in mouse hippocampus. HDAC inhibition derepresses the Bdnf gene, allowing transcription factors at the promoter to drive Bdnf mRNA up about 2 to 3 fold. The effect was dose-dependent on circulating BHB and was reproduced by direct hippocampal infusion of BHB.
This mechanism connects three superficially distinct interventions to the same end:
- Aerobic exercise raises BHB transiently during and after the session.
- Fasting raises BHB once hepatic glycogen depletes (around 12 to 18 hours, see Fasting topic page).
- Ketogenic diets raise BHB chronically to 0.5 to 3.0 mmol/L.
Each of these conditions, in mouse data, raises hippocampal Bdnf transcription via the same HDAC-derepression mechanism. The human translation is plausible but not directly demonstrated in CNS biopsy or PET imaging; the inference rests on serum BDNF rising during fasting and ketogenic states and on the molecular mechanism being conserved across mammals.
The implication for protocol design is that BHB is the proximate mediator, not exercise per se. Exercise raises BDNF partly because it raises BHB. This explains some otherwise puzzling findings: that ketogenic diets can raise serum BDNF in sedentary subjects, that exogenous ketone esters acutely raise BDNF in some small trials, and that fasting raises BDNF on a similar timescale to exercise.
Cold and BDNF: smaller signal
Cold exposure raises serum BDNF in some small trials. Mechanism candidates include the catecholamine surge (norepinephrine binds beta-adrenergic receptors that can drive Bdnf transcription via cAMP-CREB) and direct cold-induced metabolic stress. The magnitude is smaller than exercise (roughly 30 to 50% acute serum elevation versus 200 to 300% for exercise), and the chronic effect is less well-characterized.
The practical read is that cold is a complementary BDNF stimulus rather than a primary one. An adult exercising 4 sessions per week at moderate intensity is already running the BDNF axis hard. Adding cold exposure layers a smaller signal on top. The reverse (cold-without-exercise) produces a measurable but smaller BDNF effect.
Intermittent fasting and BDNF
The fasting-BDNF link follows directly from the BHB-HDAC mechanism. As BHB rises into the 0.5 to 3.0 mmol/L range during fasts of 16 to 24 hours or longer, the conditions for hippocampal HDAC inhibition are met. Mouse data shows clear Bdnf induction during fasting; human serum BDNF shows smaller and more variable changes across studies.
Time-restricted eating windows of 16 to 18 hours typically do not reach the BHB thresholds at which the mechanism activates strongly, unless combined with carbohydrate restriction. Multi-day fasts (48 to 72 hours) reach BHB ranges of 2 to 5 mmol/L and would predict robust BDNF effects, but human trials at these durations are sparse and methodologically limited.
The honest read on fasting and BDNF is that the mechanism is plausible, the rodent evidence is strong, and the human evidence is suggestive but not conclusive. Adults considering fasting protocols should probably not anchor on BDNF as the primary outcome; the metabolic switching effects are better characterized.
The serum BDNF problem
A central methodological caveat: most BDNF measurements in humans are serum or plasma, not brain. About 99% of circulating BDNF is platelet-stored; activated platelets release it during clot formation. This means serum BDNF is heavily influenced by platelet count, platelet activation state, time of day, and processing artifacts (centrifugation speed, anticoagulant choice). The correlation between serum BDNF and brain BDNF is weak in studies that have measured both: a 2017 meta-analysis found pooled correlations of 0.3 to 0.5, suggesting serum captures roughly 10 to 25% of brain BDNF variance.
This means that interventions which clearly raise serum BDNF may or may not raise brain BDNF, and interventions that fail to raise serum BDNF may still raise brain BDNF. Cerebrospinal fluid (CSF) BDNF is a better proxy but requires lumbar puncture and is rarely measured in intervention trials. PET imaging of BDNF in living humans remains technically challenging.
The practical implication: when a supplement or intervention claim cites "raises BDNF" with a serum measurement, the brain effect is uncertain. The interventions with the strongest case for genuine brain BDNF effect are the ones with structural or cognitive endpoints (Erickson 2011 hippocampal volume) rather than serum measurements alone.
Compounds and BDNF
A few compounds have plausible BDNF effects:
- Lion's mane (Hericium erinaceus) contains hericenones and erinacines that cross the blood-brain barrier in mouse data and induce NGF and BDNF expression. The Mori 2008 RCT in mild cognitive impairment (n=30) showed improved cognitive scores on lion's mane that regressed after washout ( Mori K et al. 2008, n=30 ). The trial is small and underpowered for definitive conclusions but is the strongest single human signal for the supplement.
- SSRIs (fluoxetine, sertraline) chronically raise BDNF expression in animal models and in some human serum studies. The effect is part of the mechanistic case for the delayed onset of antidepressant action: BDNF-driven neuroplasticity takes 2 to 4 weeks to manifest, matching the clinical onset window.
- Lithium at therapeutic doses raises BDNF and is one of the strongest pharmacological BDNF-elevating agents. The therapeutic window is narrow and the application is bipolar disorder, not general cognitive enhancement.
- Psilocybin and LSD acutely raise BDNF and induce dendritic spine growth in cortex (mouse data, n>3 trials). The clinical translation in human depression and PTSD trials is consistent with a BDNF-mediated plasticity window.
For a general adult pursuing cognitive longevity, the highest-yield BDNF-relevant intervention is exercise. Lion's mane is plausible but underpowered. Pharmacological agents are reserved for clinical indications.
What BDNF does not do
A few common claims about BDNF deserve scrutiny:
- "BDNF reverses Alzheimer's." It does not. BDNF supports plasticity in healthy and mildly impaired tissue; it cannot rescue established neurodegeneration. Trials of intracerebral BDNF infusion in Alzheimer's have shown small or null effects.
- "BDNF is the master longevity protein." It is one of several growth factors important for CNS healthspan. IGF-1 and VEGF have comparable importance for vascular and structural maintenance.
- "Supplements can reliably raise BDNF." Most supplement claims rest on serum measurements with weak brain-translation. The few interventions with structural endpoints (exercise primarily) are the more credible BDNF tools.
Practical synthesis
For BDNF support in a longevity-oriented adult:
- Aerobic exercise. 150 to 300 minutes per week at moderate intensity (60 to 75% max HR), with at least 1 to 2 sessions per week reaching higher intensity (80 to 90% max HR) for the acute BDNF spike. This is the dominant driver, accounting for most of the modifiable BDNF variance in adults under 70.
- Resistance training. 2 to 3 sessions per week. The acute BDNF effect is smaller than aerobic but the systemic insulin-sensitivity and muscle-IGF-1 effects support the plasticity machinery downstream.
- Fasting or carbohydrate restriction. Not required if exercising adequately. If pursued, target the BHB ranges (0.5 mmol/L+) where the HDAC-derepression mechanism is plausible.
- Sleep. 7 to 9 hours, with attention to slow-wave sleep. SWS is when hippocampal memory consolidation depends most heavily on the plasticity machinery BDNF supports.
- Cold and supplements. Smaller-magnitude signals; treat as optional add-ons rather than primary tools.
