Movement &
the brain.
Can moving the body change how the brain works — and does that mean exercise can protect memory or prevent dementia? The honest answer is more useful than the headline.
Before asking what exercise does to the brain, we need a working picture of what the brain is.
The brain is not one uniform organ with a single “fitness” score. It is a living network of specialised cells, blood vessels and support systems. Different regions contribute to perception, movement, emotion, attention and memory, while constantly exchanging signals with the rest of the body.
That matters because an exercise session changes many conditions at once: blood flow, energy demand, temperature, hormones, signalling molecules, sleep pressure and mood. When cognition changes after exercise, one mechanism rarely explains the entire result.
The brain is less like a muscle and more like a changing network.
A stronger biceps can be measured mainly by what the muscle produces. Brain health has many layers: cell survival, connectivity, blood supply, behaviour and performance on specific cognitive tasks. “Good for the brain” is therefore a starting point, not a complete scientific claim.
Neurons
Cells specialised for receiving, processing and sending signals. Electrical activity travels along the cell; chemical messengers often carry the signal across a synapse.
Glial cells
Several cell types that support, nourish, insulate and defend neural tissue. They are active partners in brain function, not simply packing material.
Synapses
Junctions where one neuron influences another cell. Their strength and organisation can change with experience — part of what scientists call plasticity.
Grey matter
Tissue rich in neuronal cell bodies, dendrites and synapses. “Grey” describes its appearance, not its importance relative to white matter.
White matter
Bundles of long nerve fibres, many wrapped in myelin. It helps distant brain regions communicate efficiently.
Networks
Distributed regions that coordinate for a task. Memory, attention and decision-making emerge from interacting systems rather than a single isolated centre.
How does one nerve cell
pass a message on?
A neural message is neither pure electricity nor pure chemistry. It is an electrochemical relay: voltage changes carry information along a cell; molecules usually carry it across the narrow gap to the next cell.
A resting neuron keeps different concentrations of charged particles — ions — inside and outside its membrane. Pumps and channels maintain that separation. When incoming signals move the membrane voltage past a threshold, voltage-gated sodium channels open and an action potential begins. Sodium enters, the voltage rapidly rises, potassium channels then help restore it, and a brief refractory period prevents the impulse from immediately travelling backwards.
The action potential is often described as “all-or-none”: once threshold is reached, one spike is not half-sized because the intention is weak. Information can instead be represented by which neurons are active, their timing and how frequently they fire. Myelin allows impulses to travel efficiently between gaps called nodes of Ranvier, but conduction speed varies with axon diameter and myelination.
Cross a synapse.
Select each stage. The drawing enlarges a gap measured in nanometres and simplifies thousands of proteins into a readable sequence.
The spike reaches the terminal.
The action potential depolarises the end of the axon. It has carried the message along one cell without neurotransmitter travelling the full distance.
Electrical here: charged ions cross the neuronal membrane in a precisely timed sequence.
Voltage opens calcium channels.
Depolarisation opens voltage-gated calcium channels. Calcium enters the terminal down its electrochemical gradient and acts as the immediate trigger for vesicle fusion.
Why calcium matters: it couples an arriving electrical event to chemical release.
Vesicles release transmitter.
Membrane-bound vesicles fuse with the presynaptic membrane and release neurotransmitter into the cleft. The identity of the transmitter and receptor helps determine the effect.
At a motor end plate: the transmitter is acetylcholine. In the brain, glutamate and GABA are common, but far from the only examples.
Receptors change the next cell.
Transmitter binds to receptors on the receiving membrane. Ionotropic receptors directly open an ion channel; metabotropic receptors start slower signalling cascades. The result may increase or decrease the probability that the next cell fires.
No single vote decides: a neuron integrates many excitatory and inhibitory inputs across space and time.
The signal is cleared.
Transporters can take transmitter back up, enzymes can break it down, and molecules can diffuse away. Calcium is removed from the terminal and vesicle components are recycled.
Resetting is part of signalling: without termination, the next message would be difficult to distinguish.
A rapid, regenerating change in membrane potential that travels along an excitable cell.
The microscopic extracellular space between the sending and receiving membranes.
A chemical released by one cell that binds receptors and changes another cell.
A lasting alteration in signalling strength, structure or network organisation; not every brief signal becomes a lasting change.
A synapse does not simply “send electricity across a gap.”
Voltage changes travel along the neuron. At most human synapses, an arriving voltage change controls chemical release; receptors then convert that chemical information into a new electrical or biochemical response in the next cell. A single movement depends on enormous populations of these relays working in parallel.
You decide to lift.
What happens next?
Imagine a dumbbell resting on a table. You reach, grip and lift it. The movement feels like one act, but it is a continuously corrected conversation between brain, spinal cord, peripheral nerves, muscle and sensory receptors.
The sequence below is a teaching model, not a literal queue in which one part waits for the previous part to finish. Planning, descending command, spinal processing and sensory feedback overlap within milliseconds. Both sides of the brain contribute to planning and posture; many corticospinal fibres influence interneurons rather than contacting a motor neuron directly.
Follow a lift from thought to force.
Select a stage. The highlighted route shows the dominant focus while the rest of the loop continues to operate.
“Pick it up” becomes a motor problem.
Visual and somatosensory systems help estimate the dumbbell’s location, shape and likely weight. Parietal networks combine information about the object with a continuously updated map of the body. Prefrontal and action-selection circuits contribute the goal and context: lift it, do not push it, and use the right hand.
The basal ganglia participate in selecting and scaling actions; they are not a single start button. The cerebellum contributes predictions about timing, load and the sensory consequences expected from the movement.
The nervous system organises a solution.
Premotor and supplementary motor networks help assemble the reach, grip, posture and order of muscle activity. The plan must account for the shoulder and trunk, not only the elbow. Anticipatory postural adjustments can begin before the obvious arm motion so the lift does not pull the whole body off balance.
The plan is shaped by experience and context. A familiar 5 kg dumbbell and an unknown black suitcase invite different initial force estimates.
Populations of neurons specify changing commands.
Primary motor cortex contributes strongly to voluntary force and fractionated movement, especially of the hand. Individual cortical neurons do not map neatly onto one muscle or one instruction. Activity across populations relates to direction, force, timing, joints and task context.
Descending output is accompanied by an internal copy of the command — often called an efference copy or corollary discharge — that helps predictive systems compare what was intended with what returns from the body.
The command travels through the corticospinal system.
Axons descend through deep cerebral white matter, the brainstem and, for many fibres, cross in the lower medulla before continuing in the spinal cord. The corticospinal tract is crucial for skilled voluntary control, but it is only one descending system. Brainstem pathways also contribute to posture, tone and whole-body coordination.
Many corticospinal signals reach spinal interneurons first. Direct cortex-to-motor-neuron connections are especially developed for dexterous hand and finger actions.
The spinal cord is an active controller.
Interneurons combine descending commands with sensory input and local circuitry. Alpha motor neurons in the ventral spinal cord form the final common pathway to skeletal muscle. One motor neuron and every muscle fibre it innervates form a motor unit.
Force rises by recruiting additional motor units and by increasing the firing rate of active units. Lower-threshold units are commonly recruited before larger, higher-threshold units, although dynamic tasks, mechanics and sensory feedback make natural recruitment more flexible than a single rule suggests.
The nerve speaks acetylcholine to muscle.
When the motor-axon action potential reaches its terminal, voltage-gated calcium channels open. Calcium triggers vesicles to release acetylcholine. Acetylcholine binds nicotinic receptors on the muscle end plate, opening ion channels and producing an end-plate potential. If threshold is reached, a muscle-fibre action potential spreads along the sarcolemma.
Acetylcholinesterase rapidly breaks down acetylcholine, helping each neural command remain discrete. Disorders of this junction can cause weakness even when intention and muscle tissue are otherwise present.
Voltage releases calcium; calcium permits force.
The muscle action potential travels down T-tubules. Voltage-sensing CaV1.1 channels — historically called DHPRs — are mechanically coupled to RyR1 calcium-release channels in the sarcoplasmic reticulum. RyR1 opens and calcium floods the muscle-cell interior.
Calcium binds troponin C, moves tropomyosin away from actin’s binding sites and permits myosin cross-bridge cycling. ATP is required for cross-bridge cycling and for ion pumps. The force at the hand reflects motor-unit recruitment, firing rate, muscle length and velocity, tendon mechanics, joint leverage and coordination across agonists, synergists and antagonists.
The lift is controlled by what comes back.
Muscle spindles report aspects of muscle length and change; Golgi tendon organs contribute information related to muscle force; skin and joint signals add contact and position information. Vision and the vestibular system help stabilise the object and body. Spinal, cortical and cerebellar circuits use this feedback alongside prediction to correct grip and trajectory.
To stop, descending drive and motor-neuron firing fall. Acetylcholine release ceases, acetylcholine is cleared, and SERCA pumps calcium back into the sarcoplasmic reticulum. Troponin loses calcium, tropomyosin again limits actin–myosin interaction, and active force declines. Lowering the weight still requires controlled muscle activity; “relaxation” is not simply switching the body off.
Prediction gets the lift started.
The nervous system uses prior experience to estimate grip force, timing and posture before feedback from the new movement can arrive.
Error correction keeps it accurate.
If the weight is heavier than expected, sensory evidence updates motor output. Prediction and feedback cooperate; neither works alone.
One alpha motor neuron plus all skeletal-muscle fibres supplied by its axon.
A major descending pathway from several cortical areas to the spinal cord.
Information contributing to the sense and control of body position, movement, effort and force.
The repeating contractile unit where actin and myosin generate active tension.
Memory is a process,
not a storage box.
To remember an event, information must be encoded, stabilised and later retrieved. These steps recruit overlapping, not identical, circuits.
The hippocampus is a curved structure deep in each temporal lobe. It is strongly involved in forming and organising new episodic and spatial memories: what happened, where it happened and how one experience relates to another. It does not hold every memory forever, and it does not work alone.
Memory consolidation describes the processes through which a memory trace becomes more stable after initial learning. Sleep, repeated retrieval and time all matter. Plasticity means that neural connections and networks can alter with experience. Plasticity is not automatically beneficial; the brain also learns unhelpful habits and fear responses.
Explore four parts of the system.
Select a label to see where it sits and what it contributes. The drawing is intentionally simplified and is not a diagnostic image.
Hippocampus
Deep in the medial temporal lobe, it helps organise new episodic and spatial memories.
Precision: MRI volume is a broad structural measure. It does not directly count new neurons.
Frontal networks
Distributed frontal circuits contribute to planning, working memory, inhibition and flexible behaviour.
Precision: executive function is a family of abilities, not one brain switch.
Cerebral circulation
Arteries deliver oxygen and fuel while vascular regulation keeps brain blood flow within a workable range.
Precision: harder exercise does not simply mean proportionally more brain blood flow.
White matter
Myelinated fibres connect distant regions, allowing networks to exchange information efficiently.
Precision: network health depends on connectivity as well as local grey matter.
How can a moving body signal to the brain?
Exercise begins in contracting muscle, but the response travels through circulation, metabolism and the nervous system.
Follow the signal
Tap each step. The sequence is simplified; several pathways run in parallel and influence one another.
Working tissue needs energy.
Muscle contraction raises demand for ATP, the cell’s immediately usable energy currency. Breathing, heart rate and blood distribution adjust to help meet that demand.
Delivery changes.
Cardiac output — heart rate multiplied by the blood pumped per beat — rises. Cerebral blood flow is regulated rather than simply “maximised”; intensity, carbon dioxide and fitness all influence the response.
Chemical information changes too.
Lactate, catecholamines, growth factors and muscle-derived signals are among the candidates linking exercise to neural effects. Finding a change in blood does not prove that the same change occurred inside a specific brain region.
Different systems respond on different timescales.
Arousal and attention can change within minutes. Vascular, synaptic and structural adaptations — if they occur — require repeated exposure and recovery over longer periods.
The outcome depends on the person and the task.
Age, baseline fitness, sleep, disease, medication, exercise dose and genetics can all alter the response. A laboratory memory test and daily cognitive independence are related questions, not identical outcomes.
BDNF is a family of biological signals, not “fertiliser for the brain.”
Brain-Derived Neurotrophic Factor is a protein in the neurotrophin family. Neurons and some glial cells can produce it, and neural activity can influence its synthesis and release. “Neurotrophic” refers to support for the development, maintenance or function of nervous-system cells. That broad definition does not mean more BDNF is universally better, nor that one blood value measures “brain growth.”
BDNF is first synthesised as a precursor. ProBDNF can be cleaved into mature BDNF. The two forms can favour different receptor systems and effects: mature BDNF commonly activates the TrkB receptor, while proBDNF can signal through p75 NTR with co-receptors. Real tissue biology depends on location, timing, receptor availability, activity and disease state. Reducing this to “BDNF good, proBDNF bad” would also be inaccurate.
Support and maintenance
During development and in selected adult circuits, BDNF–TrkB signalling supports neuronal survival, differentiation and maintenance. It is one part of a larger trophic environment, not a universal rescue switch.
Transmission
BDNF can alter transmitter release, receptor trafficking and excitability. Its effect varies by cell type and circuit; it can modify both excitatory and inhibitory transmission.
Long-term potentiation
BDNF–TrkB contributes to forms of long-lasting synaptic strengthening, including protein synthesis and structural changes associated with late-phase LTP.
Dendrites and spines
Experimental work links BDNF to dendritic growth, spine remodelling and cytoskeletal change. A microscopic structural effect cannot be inferred from a routine blood test.
Motor plasticity
BDNF is implicated in cortical plasticity and motor learning. Genetic variation, baseline state and task design can modify associations, so it is not a direct score of skill acquisition.
Adult neurogenesis
Exercise, BDNF-related signalling and hippocampal neurogenesis are strongly connected in animal models. The scale, function and measurement of adult human neurogenesis remain debated; blood BDNF cannot establish it.
What can we actually conclude?
Move from molecular mechanism to a person’s health outcome. Certainty does not automatically travel upward with the same strength.
BDNF binding to TrkB starts signalling cascades.
TrkB activation can recruit pathways involving MAPK/ERK, PI3K–Akt and PLCγ. These influence gene expression, protein synthesis, cytoskeletal organisation, survival and synaptic function. The exact response depends on the cell, circuit and timing.
Exercise changes BDNF-related biology in animals.
Rodent studies can sample brain tissue and manipulate pathways directly. They support links among exercise, hippocampal plasticity, learning and neurogenesis. Translation is informative, not automatic: species, dose and experimental conditions differ from ordinary human training.
Serum or plasma BDNF is not a window into one brain region.
Platelets store large amounts of BDNF and release it during clotting, so serum and plasma values answer different questions. Sampling time, exercise intensity, hydration, platelet handling, assay method, medication, age and sex can all affect the result. A 2022 human systematic review found that peripheral BDNF did not reliably reflect brain content.
A higher blood value does not prove better memory or prevented dementia.
To establish a meaningful outcome, a study must measure that outcome with an appropriate comparison and timeframe. BDNF may participate in exercise-related plasticity, but it cannot carry the full causal explanation for cognition, mood, skill, brain structure or disease risk.
What exercise studies in people show
Acute exercise often produces a transient rise in circulating BDNF, particularly after moderate-to-vigorous cardiovascular work and when blood is sampled soon after the session. “Transient” matters: the value can rise and then move toward baseline. That is different from a permanently elevated resting level.
A 2022 meta-analysis of 21 randomised-trial reports with 809 healthy participants found higher circulating BDNF after both acute exercise and longer programmes. Variation between studies was large. A 2026 review of acute cardiovascular exercise included 22 studies and 628 participants and estimated a moderate pooled rise in circulating BDNF (Hedges’ g =0.48), while again identifying intensity, sample type and sampling time as important moderators.
What appears most effective for raising circulating BDNF?
Choose a mode. The ranking concerns the short-term blood response, not proven brain concentration, intelligence or long-term academic performance.
High-intensity interval exercise often produces the largest immediate peripheral rise.
A 2023 network meta-analysis ranked HIIT highest for the acute peripheral BDNF response in adults. Earlier meta-analysis also linked longer acute exercise duration with larger increases. This does not make maximal intervals the best choice for every person, every day or every cognitive goal.
Practical reading: intensity is one lever, but safety, training status and the ability to recover decide whether the session is useful.
Moderate-to-vigorous aerobic exercise has the clearest repeated human signal.
Running, cycling and other endurance modes dominate the literature. Acute rises are usually transient and are most often measured immediately after exercise. Duration, intensity, fitness and sampling method help explain why studies disagree.
Practical reading: a sustainable aerobic base is more useful for health than chasing the biggest possible laboratory spike.
Resistance training may change BDNF, but the acute literature is less consistent.
Protocol, muscle mass involved, rest periods and blood sampling vary widely. Resistance exercise remains valuable for strength, function and general health even if its immediate blood-BDNF response is smaller or less predictable than vigorous aerobic work.
Practical reading: BDNF should not be used to rank the total value of strength and aerobic training.
Weeks of exercise can alter resting BDNF on average, but results are heterogeneous.
A 2022 meta-analysis of 21 randomised-trial reports found positive acute and long-term pooled effects in healthy people. Study size, age, sex, aerobic mode and sample handling influenced the result. An elevated baseline value is not required for training to improve cognition or fitness.
Practical reading: train for outcomes you care about. Treat BDNF as a mechanism candidate, not the programme target.
Does BDNF simply fall with age?
Select the sample. Different biological compartments produce different age patterns, so there is no honest universal concentration curve.
Age changes the signal—but the sample changes the story.
CSF study direction. In 128 cognitively normal adults, CSF BDNF was lower at greater age. The curve visualises direction only; the paper does not establish a universal decline rate for an individual.
Illustrative association map, not reference values. The vertical axis has no concentration units because serum, plasma and CSF are not directly interchangeable and the selected studies used different methods and populations.
Human serum studies do not show one consistent age slope.
In 259 healthy volunteers, mean serum BDNF was about 32.7 ng/mL with a wide range; the study reported a small increase with age and substantial within-person variation. In a separate community sample of 1,230 middle-aged and older adults, serum BDNF declined with age in women but remained stable in men. Serum values depend strongly on clotting and platelet release.
Conclusion: do not compare a personal serum number with a simplified “young versus old” chart.
One study of 140 healthy adults found lower plasma BDNF with greater age and weight.
Participants were 20–60 years old. Platelet BDNF did not show the same age pattern. Even tiny platelet contamination can shift plasma results, which makes collection and processing decisive.
Conclusion: a declining plasma association cannot be converted into a known decline inside the hippocampus.
In 128 cognitively normal adults, cerebrospinal-fluid BDNF decreased with age.
The same study found higher CSF BDNF in women than men and associations with memory measures. It was observational and relatively small; lumbar-puncture samples are not a population reference curve.
Conclusion: this supports an age-related central signal, while leaving cause, individual prediction and the effect of training unresolved.
Living human brain-region BDNF cannot be tracked like a routine blood marker.
Animal tissue, post-mortem tissue, cerebrospinal fluid, genetics, imaging and peripheral blood answer different questions. Peripheral BDNF has not shown a reliable one-to-one relationship with content in a specific living human brain region.
Conclusion: age almost certainly changes BDNF biology, but a single blood concentration is not a brain-age score.
BDNF may help create conditions for plasticity. It cannot replace studying.
Learning requires attention, encoding, retrieval practice, error correction and sleep-dependent consolidation. Exercise can change arousal, blood flow and molecular signals—including BDNF—and some small experiments link post-exercise BDNF changes with later learning. That is a plausible window, not a guaranteed cognitive shortcut.
Regular aerobic and resistance training support health. A brief moderate or harder aerobic bout may sharpen arousal for some people, but exhausting work can also impair the next task.
Explain material, solve problems and connect new facts to what you already know. Exposure alone is weak learning.
Testing yourself and spacing practice are direct learning tools. A BDNF rise is not evidence that information was encoded.
Sleep supports memory processing and next-day attention. Training that repeatedly damages sleep may undermine the academic goal it was meant to support.
Exercise can change circulating BDNF. The next causal step must still be tested.
The blood finding is consistent with a biological environment that supports plasticity. It does not show where the BDNF came from, how much reached a particular brain circuit, whether it caused a cognitive change, or which training prescription is best for one person.
One striking trial — then the wider evidence.
A good scientific story does not end with the most memorable result. It asks whether the result repeats.
older adults randomised for one year
Can aerobic training alter hippocampal volume?
- Who?
- Adults aged 55–80 without dementia; 120 entered the trial.
- Comparison
- Aerobic walking versus stretching and toning.
- Duration
- 12 months, with supervised training and progressive intensity.
- Measured
- MRI hippocampal volume, cardiorespiratory fitness, serum BDNF and spatial memory.
- Result
- The aerobic group showed about a 2% increase in anterior hippocampal volume, while volume declined in the control condition. Spatial-memory performance improved, and volume change was associated with fitness and BDNF changes.
Anterior hippocampal volume after 12 months of aerobic training in the 2011 trial.
Compelling, randomised, and frequently cited.Eight trials, 554 healthy older adults: no statistically significant pooled effect on hippocampal volume.
95% CI −0.01 to 0.21; p=0.073.Three lenses answer three different questions.
Strong control of the intervention; usually smaller samples and shorter follow-up.
Combines studies, while inheriting differences in methods, quality and populations.
Can observe disease incidence at scale, but cannot remove every confounder.
The 2024 review included supervised aerobic programmes lasting three to twelve months. Fitness improved, but the pooled hippocampal-volume effect did not reach statistical significance. Only the Erickson trial had hippocampal volume as its primary outcome. MRI methods, sample size, programme duration and adherence differed.
The brain remains responsive.
- A well-designed trial showed structural and memory changes after a year.
- Aerobic training reliably improved cardiorespiratory fitness across the pooled trials.
- The hippocampus is a credible research target.
“Cardio grows your brain” is too certain.
- The pooled structural evidence is currently inconclusive.
- MRI volume cannot identify which microscopic process changed.
- A group average does not predict one person’s response.
Three questions that must not be collapsed into one.
Can one bout alter mood, attention or reaction time?
Can training improve a tested cognitive domain?
Is activity associated with lower dementia incidence?
An acute exercise experiment can support a causal claim about a short-term test result. It cannot show that dementia was prevented decades later. A prospective cohort can follow hundreds of thousands of people long enough to observe diagnoses, but active and inactive people may differ in education, health, wealth and early disease processes.
A 2023 umbrella review of 20 reviews, covering 332 primary studies in adults aged 55 and over, found a small overall positive effect of exercise on cognition ( d =0.22), with very high variation across studies. Specific domains and exercise types did not respond uniformly.
A separate meta-analysis of 29 prospective cohorts, involving more than two million participants, reported that greater physical activity was associated with lower Alzheimer’s disease incidence. After the strongest available confounder adjustment, the pooled hazard ratio was 0.85. That means a lower relative rate in more-active groups; it does not prove that activity alone caused the difference.
“Lower risk” is not the same as “prevention.”
Dementia is an umbrella term for syndromes caused by different diseases. Risk accumulates through age, genetics, cardiovascular health, education, hearing, social and environmental factors, and disease processes that may begin years before diagnosis. Movement can be one modifiable part of that system.
Where genetics enters the picture
APOE is a gene involved in lipid transport. One variant, APOE ε4, is associated with a higher risk of late-onset Alzheimer’s disease, but it is not a diagnosis and does not determine an individual future. Reviews of physical activity and APOE report mixed genotype-specific findings. One meta-analysis found similar benefit on most brain-health outcomes for carriers and non-carriers, with possible differences in brain activation that still require confirmation.
This is a useful example of genotype × environment: biological variation can shape probability and response, while behaviour and environment also matter. It is not evidence that one exercise plan can “override” genetic risk.
Repeated exercise improves cardiorespiratory fitness and supports multiple cardiovascular and metabolic outcomes.
Controlled reviews suggest small average benefits, with substantial heterogeneity.
Large cohorts often find lower risk among more-active groups, but residual confounding and reverse causation remain.
BDNF, blood flow and plasticity are plausible pieces; no single pathway explains every human outcome.
Train for the whole system.
Brain health does not require a mysterious “neuro” workout. It benefits from the same long-term foundations that support cardiovascular, metabolic and functional health.
The most defensible message is to build a pattern you can repeat: aerobic work that challenges circulation, strength training that preserves capacity, varied movement, sleep and recovery, and gradual progression. The exact mix depends on health, preference, training history and access.
- Use more than one mode.
Aerobic and strength training answer different physical questions. Neither should be sold as a complete brain-health treatment. - Progress from your current capacity.
A tolerable dose repeated for months is more informative than one heroic session. - Protect the basics.
Sleep, cardiovascular risk management, hearing, social connection and medical care matter alongside exercise. - Keep claims proportional.
Exercise can support health and may contribute to lower risk. It cannot promise immunity from cognitive decline or dementia.
Read the evidence,
not only the headline.
- Erickson KI et al. Exercise training increases size of hippocampus and improves memory. PNAS, 2011. PubMed
- Balbim GM et al. Aerobic exercise training effects on hippocampal volume in healthy older individuals: a meta-analysis of randomized controlled trials. GeroScience, 2024. PubMed
- Liu P et al. The effect of physical exercise on circulating BDNF in healthy subjects: a meta-analysis of randomized controlled trials. Brain and Behavior, 2022. PubMed
- Demurtas J et al. Exercise and cognition in healthy adults aged 55 years and older: an umbrella review. European Review of Aging and Physical Activity, 2023. PubMed
- Zhang X et al. Physical activity and Alzheimer’s disease: meta-analysis of 29 prospective cohort studies. Ageing Research Reviews, 2024. PubMed
- Pearce AM et al. APOE genotype moderation of the association between physical activity and brain health. Frontiers in Aging Neuroscience, 2022. PubMed
- World Health Organization. Risk reduction of cognitive decline and dementia: WHO guidelines. 2019. WHO
- Huberman Lab with Wendy Suzuki was used as question-finding and explanation input. Claims above were checked against original studies and reviews. Episode & show notes
- Hallett M. Volitional control of movement: the physiology of free will. Clinical Neurophysiology, 2007. PubMed
- Natali AL et al. Neuroanatomy, Corticospinal Cord Tract. StatPearls, updated 2026. PubMed
- Proske U, Gandevia SC. The proprioceptive senses: their roles in signalling body shape, position, movement and force. Physiological Reviews, 2012. PubMed
- Hodson-Tole EF, Wakeling JM. Motor unit recruitment for dynamic tasks. Journal of Comparative Physiology B, 2009. PubMed
- Shishmarev D. Excitation–contraction coupling in skeletal muscle: recent progress and unanswered questions. Biophysical Reviews, 2020. PubMed
- Estephan EP et al. Electrophysiological evaluation of the neuromuscular junction: a brief review. Arquivos de Neuro-Psiquiatria, 2023. PubMed
- Sasi M et al. Neurobiology of local and intercellular BDNF signalling. Pflügers Archiv, 2017. PubMed
- Nicolini C, Nelson AJ. Methodological pitfalls in the assessment of exercise-induced peripheral BDNF. Frontiers in Neuroergonomics, 2021. PubMed
- Neurospecific molecules measured in the periphery: how do they correlate with brain levels? A systematic review. 2022. PubMed
- Moreau D, Chapple-De Lange BT. Acute cardiovascular exercise and circulating neurotrophic factors: a systematic review and meta-analysis. Brain Research, 2026. PubMed
- Dinoff A et al. The effect of acute exercise on blood concentrations of brain-derived neurotrophic factor in healthy adults: a meta-analysis. European Journal of Neuroscience, 2017. PubMed
- Huang T et al. Effectiveness of high-intensity interval training on peripheral BDNF in adults: a systematic review and network meta-analysis. 2023. PubMed
- Polacchini A et al. Measuring and validating the levels of brain-derived neurotrophic factor in human serum. eNeuro, 2018. PubMed
- Lommatzsch M et al. The impact of age, weight and gender on BDNF levels in human platelets and plasma. Neurobiology of Aging, 2005. PubMed
- Bus BA et al. Serum BDNF: determinants and relationship with depressive symptoms in a community population. World Journal of Biological Psychiatry, 2012. PubMed
- Li G et al. Cerebrospinal fluid BDNF concentration and cognitive function in non-demented subjects. PLoS One, 2009. PubMed
Evidence changes. This page distinguishes direct experimental findings from observational associations and proposed mechanisms. It is educational and does not diagnose, prevent or treat a medical condition.
Ask what was unclear.
Your question can help shape a future explainer. Do not include private medical information in the message.