BDNF — the brain growth factor that links exercise to cognition

BDNF had been identified in the early 1980s, isolated from pig brain tissue, characterized as a member of a family of proteins that supported neuronal survival. The neurotrophin family — NGF, BDNF, NT-3, NT-4 — had been understood primarily in the context of development: these were the proteins that determined which neurons survived during the embryonic pruning process and which didn't. What the exercise researchers found was that BDNF was doing something else in the adult brain, something ongoing and activity-dependent and far more dynamic than its developmental role had suggested. It wasn't just determining survival. It was regulating the brain's capacity to change.

The basic biology runs like this. BDNF is synthesized primarily in neurons — specifically in the cell body, with expression driven by neural activity, and it's released at synapses where it acts locally on both the neuron that released it and neighboring cells. Its primary receptor is TrkB, tropomyosin receptor kinase B, a tyrosine kinase receptor. When BDNF binds TrkB, it activates two major downstream signaling cascades: the PI3K/Akt pathway, which drives cell survival and growth, and the MAPK/ERK pathway, which drives gene expression changes, protein synthesis, and synaptic modification. These aren't subtle effects. TrkB activation drives the actual physical remodeling of synaptic connections — dendritic spine growth, synapse formation, the structural changes that encode learning at the cellular level.

The synaptic connection to learning is direct. Long-term potentiation — LTP — is the cellular mechanism underlying memory formation: repeated activity across a synapse strengthens the synapse, making the connection more efficient and enduring. BDNF is required for the late phase of LTP, the phase that converts short-term synaptic change into lasting structural modification. Without adequate BDNF, synaptic potentiation occurs but doesn't consolidate into durable memory traces. In animal studies, blocking BDNF with specific antibodies impairs LTP induction; infusing BDNF into the hippocampus enhances it. The relationship is causal and has been demonstrated in many experimental contexts.

What this means for the aging brain is significant, because BDNF levels are not constant across the lifespan.

They peak during development and adolescence, when the brain's plasticity is highest, and decline gradually through adult aging. Lower BDNF is consistently associated, in human observational studies, with worse episodic memory performance, slower processing speed, and higher risk of depression. Post-mortem studies have found lower BDNF protein concentrations in the hippocampi of individuals who died with Alzheimer's disease compared to age-matched controls. The causality in humans is harder to establish than in animals — lower BDNF may be a consequence of neurodegeneration rather than a driver of it — but the association is robust across populations and consistent with the mechanistic picture.

The genetic layer adds nuance that matters practically. The BDNF gene has a common single-nucleotide polymorphism called Val66Met, present in roughly thirty percent of people of European ancestry and higher proportions in some East Asian populations. In the Met version of this variant, the activity-dependent secretion of BDNF is impaired: neurons with Val66Met release less BDNF in response to neural activity than neurons with the more common Val/Val genotype. This matters for exercise response, for learning, and for depression risk in ways that have been studied extensively. People with the Met allele may have a blunted BDNF response to exercise and a different profile of response to pharmacological interventions that work partly through BDNF. This is not a deterministic disadvantage — Val66Met is a common variant, not a disease mutation — but it's an example of how individual variation in BDNF biology is real and worth knowing about.

Now, the interventions.

Exercise is the most evidence-supported way to raise BDNF, and the evidence here is unusually robust for a behavioral intervention: human trials, animal studies, mechanistic work, and epidemiological associations all point in the same direction. Aerobic exercise raises BDNF acutely — a single session elevates circulating BDNF within minutes to hours — and chronic exercise training produces sustained elevations in hippocampal BDNF and measurable increases in hippocampal volume. Both aerobic and resistance training have demonstrated effects, though the aerobic exercise evidence is larger. The mechanism involves multiple pathways: lactate produced during exercise crosses the blood-brain barrier and stimulates BDNF transcription directly; irisin, a muscle-released hormone (myokine) induced by exercise, may cross the blood-brain barrier and upregulate BDNF; reduced chronic inflammation from exercise removes a brake on BDNF expression. Exercise is not a supplement. It's a pharmacological intervention with a pharmacological effect on a specific molecular target, and the BDNF data is one of the reasons the effect on cognition isn't just correlation.

Caloric restriction and intermittent fasting also raise BDNF, by mechanisms that partially overlap with exercise. Fasting reduces mTOR activity and activates AMPK, which in turn influences BDNF expression; reduced circulating glucose and insulin appear to be permissive for BDNF upregulation. Sleep, as noted elsewhere, is a major BDNF consolidation window — slow-wave sleep in particular appears to facilitate hippocampal BDNF signaling. Social engagement, environmental novelty, and active learning all drive BDNF through the activity-dependent secretion pathway. The common thread is neural activity: BDNF follows the brain's work. When the brain is challenged — by exercise, learning, social complexity, novel environments — BDNF rises to support the cellular changes that encode the experience.

The pharmacological interventions are more complicated and the evidence more varied.

SSRIs raise BDNF over weeks, which is thought to be part of the mechanism underlying their antidepressant effects — not the immediate serotonin reuptake inhibition, but the downstream BDNF upregulation that follows sustained serotonergic activity. This is one reason SSRIs take weeks to produce antidepressant effects even though their primary pharmacological action is immediate. Lithium, which is used both in psychiatric treatment and has been studied in aging research, raises BDNF through a distinct pathway involving glycogen synthase kinase 3 beta inhibition. Both are examples of well-characterized pharmacological BDNF effects in humans.

Ketamine and esketamine produced a different understanding. The rapid antidepressant effect of ketamine — improving mood within hours in treatment-resistant depression, not weeks — was unexpected and didn't fit the BDNF-requires-weeks model. Research, including a key 2010 study from the Bhagya lab and subsequent work, found that ketamine produces a rapid burst of BDNF release, and that blocking TrkB in animal models eliminates the antidepressant effect. The mechanism involves glutamate — ketamine blocks NMDA receptors, which disinhibits AMPA receptor activity, which triggers rapid BDNF release and TrkB signaling. Esketamine, the S-enantiomer of ketamine, is FDA-approved for treatment-resistant depression under the brand name Spravato. This is a case where BDNF-adjacent pharmacology has reached regulatory approval, though the compound's primary label is for depression rather than cognition.

The peptide territory is worth understanding with appropriate precision about the evidence.

Semax, the synthetic ACTH-derived neuropeptide developed in Russia, has been researched primarily in post-ischemic contexts and has demonstrated BDNF upregulation in animal studies and in some Russian human trials focused on stroke recovery and cognitive impairment. It is not FDA-approved and is not available as a conventional pharmaceutical in the United States; it's accessed as a compounded peptide. The BDNF-upregulating effect is one of the proposed mechanisms underlying cognitive effects reported by users and observed in the research, but the evidence base comes primarily from Russian research institutions, involves limited sample sizes, and has not been replicated in large-scale Western clinical trials. It's a plausible mechanism with a narrow but real evidence base. PE-22-28 is a synthetic peptide fragment studied primarily in preclinical models for antidepressant effects, with proposed mechanisms involving TrkB; the evidence in humans is very early. These peptides belong to a research category where the biology is interesting and the clinical translation is genuinely preliminary.

The honest framing for BDNF interventions is a hierarchy. At the top: exercise, sleep, and social and cognitive engagement — these have the strongest human evidence and the clearest mechanistic rationale, and anyone looking for BDNF support who isn't optimizing these first is looking for a workaround to a problem they haven't solved. In the middle: dietary patterns consistent with lower chronic inflammation (Mediterranean diet has some association data with BDNF), and for people with clinical indications, medications that raise BDNF as part of their mechanism. At the periphery: peptide research compounds where the mechanism is coherent and the human evidence is preliminary, always in consultation with a prescribing provider who understands the full picture.

The reason BDNF occupies so much space in brain health research is not that it's a single magic molecule. There are hundreds of proteins in the brain that affect cognition. BDNF is particularly well-studied because its role sits at the intersection of plasticity, survival, and activity-dependence — three concepts that together describe what the brain needs to do throughout life. It's the molecular correlate of the observation that the brain responds to what you ask of it.

That observation has implications that are broader than any pharmacology. The exercise-BDNF-hippocampus connection says something structural about how the brain maintains itself: it is not a fixed hardware system that degrades passively. It is a responsive system that upregulates its own maintenance machinery in response to demand. Use creates the conditions for continued function. Sedentary, unstimulating, socially isolated conditions allow those maintenance signals to fall. This isn't a motivational claim. It's a molecular mechanism. BDNF is one of the reasons the brain is an argument for challenge — not because challenge is virtuous, but because challenge is, quite literally, what the machinery runs on.

What BDNF biology ultimately teaches is that the brain's capacity for change doesn't turn off. It narrows with age, in specific ways, in specific regions. The Val66Met variant modifies it. Poor sleep and chronic stress suppress it. But the underlying machinery — the activity-dependent release of a protein that physically reshapes synaptic connections — is present and operating in the hippocampus of a seventy-five-year-old who exercises regularly, sleeps well, and keeps learning. The brain's capacity for change is not a youthful property that runs out. It's a system that responds to signals. The question is what signals you're sending it.