The brain, treated as biology.

Three things change, on different timescales. Synaptic strength shifts within minutes to hours — existing connections become more or less responsive through receptor trafficking. Structural remodelling takes days to weeks, as dendritic spines physically grow or retract. Myelination takes weeks to months, insulating the axons you use repeatedly so signals travel faster.

The practical consequence is that learning something once changes almost nothing durable. The repetition is what converts a transient strength change into a structural one, and sleep is when much of that conversion happens.

Brain-derived neurotrophic factor is the growth factor that supports neuron survival, dendritic branching and long-term potentiation — the cellular basis of learning. It behaves like a permission slip: without adequate BDNF signalling, repetition produces far less structural change.

What raises it

• Aerobic exercise — the most reliable non-pharmacological lever, with effects measurable after a single session and cumulative over weeks
• Deep sleep — consolidation windows coincide with slow-wave activity
• Intermittent fasting and ketone availability — beta-hydroxybutyrate appears to upregulate BDNF expression
• Novelty and effortful learning — difficulty itself is part of the stimulus

What suppresses it

• Chronic inflammation — inflammatory cytokines reduce BDNF expression, which is the mechanistic bridge between gut permeability and cognitive dullness
• Sustained cortisol elevation — chronic stress shrinks hippocampal dendritic complexity in animal models
• Sleep restriction — even partial, across consecutive nights

New synapses are physical structures, and their membranes are built substantially from phospholipids. DHA in particular is enriched in synaptic membranes and influences their fluidity, which in turn affects how receptors sit and signal.

This is why omega-3 status shows up in plasticity research even though it is not a stimulant and produces no felt effect. It is a materials constraint rather than a performance input — you notice its absence over years, not afternoons.

1. Sleep — seven to nine hours, with consistent timing. Nothing else on this list compensates for its absence.
2. Aerobic exercise — 150 minutes weekly at moderate intensity, or three sessions of 25 minutes at higher intensity.
3. Effortful, spaced practice — the difficulty is the signal; easy review does little.
4. Reduce chronic inflammation — which for most people means addressing the gut, not the brain.
5. Ensure the raw materials — omega-3, magnesium, B vitamins, adequate protein.

Note the ordering. Supplements sit at position five for a reason: they remove a ceiling rather than raise one.

The uncomfortable finding from metacognition research is that confidence and accuracy can move independently. You can be highly confident and wrong, or accurate and uncertain, and the neural signatures differ. Metacognitive sensitivity — how well your confidence tracks your actual performance — is a measurable trait, and it is trainable.

It also degrades before performance does. Under sleep restriction, people become poor at judging their own impairment well before their raw output falls off a cliff. This is why self-report is unreliable in exactly the conditions where it matters most.

Inflammatory cytokines cross into the central nervous system and alter prefrontal function. The subjective experience is not usually memory loss — it is a subtle loss of judgement about your own capacity. You take on the difficult task at 4pm because you cannot tell that you should not.

The chain runs: compromised intestinal barrier → bacterial fragments in circulation → low-grade systemic inflammation → reduced prefrontal efficiency → degraded self-monitoring. Every link is measurable, and the intervention point is the first one.

• Calibration training — before checking an answer, record a confidence number. Reviewing the gap between confidence and outcome, repeatedly, sharpens the signal.
• External state checks — because self-report fails when impaired, use objective proxies: hours slept, time since last meal, hours into the working day. Decide by rule rather than by feel.
• Deliberate pauses — brief interruption before consequential decisions reliably improves accuracy, largely by allowing the monitoring system to engage at all.

L-theanine increases alpha-band activity and reduces the physiological arousal that narrows attention. In practice it produces steadiness rather than stimulation — useful for sustained monitoring, less useful for raw speed.

It is not a substitute for sleep, and it will not repair the inflammation pathway. What it does reasonably well is take the edge off caffeine and hold attention flat across a long task.

L-theanine is structurally similar to glutamate and interacts with glutamatergic and GABAergic signalling, though the precise binding profile is still debated. What is consistent across EEG studies is an increase in alpha-band power, particularly under cognitive load.

Functionally that reads as a widening of attention rather than a narrowing of it — less useful for a sprint, more useful for four hours of careful work. It does not increase arousal, which is why it is not felt as a stimulant and why people sometimes conclude it is doing nothing.

The best-supported application is combination with caffeine, at roughly a 2:1 theanine-to-caffeine ratio — 200 mg with 100 mg. Caffeine supplies the arousal; theanine removes the jitter, the heart-rate rise and the attentional narrowing that come with it.

Trials on the combination show improvements in sustained attention and task-switching that neither compound reliably produces alone. If you drink coffee and dislike how it feels by the second cup, this is the intervention worth trying first.

Theanine does not sedate. What it appears to do is reduce sleep latency and improve subjective sleep quality by lowering pre-sleep arousal — you fall asleep because you have stopped being wound up, not because you have been tranquillised.

Paired with magnesium glycinate, which supports the same downshift through a different pathway, it forms the basis of a night formulation that does not leave a morning residue.

• Focus: 100–200 mg with morning caffeine
• Sleep: 200 mg roughly an hour before bed, with food or without
• Tolerance: none reported at these doses; no need to cycle
• Interactions: may modestly lower blood pressure — worth flagging if you take antihypertensives

Expect a subtraction rather than an addition: less noise, not more power. If you are looking for a felt lift, this is the wrong molecule.

NAD+ is the electron carrier that makes oxidative phosphorylation possible. It is also the obligatory substrate for two families of repair enzymes: sirtuins, which regulate mitochondrial biogenesis and stress resistance, and PARPs, which repair DNA strand breaks.

That dual role creates a competition. DNA damage consumes NAD+ through PARP activation, leaving less for sirtuins and for energy production. In a tissue with no fuel reserve and very high demand, that trade-off shows up as function rather than as a lab value.

People rarely describe it as fatigue. The more common report is a narrowing of capacity — the same work takes longer, the afternoon becomes unproductive rather than merely tiring, and recovery from a poor night stretches from one day to three.

It is worth being precise about what supplementation can and cannot address here. Restoring a declining precursor is a plausible intervention. Reversing accumulated damage is not.

Both nicotinamide mononucleotide and nicotinamide riboside raise blood NAD+ levels in human trials — that part is well established. What is far less settled is whether raised blood NAD+ translates into raised brain NAD+, and whether that produces cognitive benefit in healthy adults.

The honest summary: the mechanism is sound, the biomarker moves, the functional endpoints in humans remain thin. Trans-resveratrol is often paired on the reasoning that it activates the sirtuins the NAD+ then feeds, though the human data there is likewise mechanistic rather than outcome-driven.

1. Cardiovascular fitness — the single strongest modifiable predictor of cognitive ageing in longitudinal data
2. Sleep quality — glymphatic clearance of metabolic waste happens predominantly during deep sleep
3. Blood pressure and glucose control — small-vessel damage is cumulative and largely silent
4. Hearing — untreated hearing loss is among the largest single risk factors for cognitive decline, and correcting it is straightforward
5. Then the cellular-energy pathway

Neurodegenerative pathology accumulates for twenty to thirty years before symptoms appear. That is bad news for late intervention and good news for early intervention: the window in which small inputs compound is enormous, and almost nobody uses it.

The corollary is that interventions started in the thirties and forties have a fundamentally different expected value from the same interventions started at seventy. Not because the mechanisms stop responding, but because compounding needs time.

Two brains with identical pathology can present very differently. The concept of cognitive reserve describes the capacity to sustain damage before function fails — built through education, occupational complexity, bilingualism, musical training and dense social networks.

The mechanism is thought to be redundancy: more alternative routes through the network means more damage tolerated before a function is lost. It is the one factor on this list you build rather than consume, and it appears to keep accruing throughout life.

Over a long enough horizon, chronic low-grade inflammation is one of the more consequential inputs to brain ageing. Because the intestinal barrier is a major determinant of systemic inflammatory load, gut health becomes a cognitive-longevity intervention rather than a digestive one.

This is the reasoning behind treating gut, skin and cognition as one pathway rather than three product categories. The gut is upstream, and upstream is where leverage sits.

• Move — aerobic and resistance work, weekly, permanently
• Sleep — protected, consistent, prioritised over almost everything else
• Metabolic health — blood pressure, HbA1c and lipids known, tracked and in range
• Hearing and vision — corrected promptly, not tolerated
• Inflammation — addressed at the gut rather than managed at the symptom
• Learning — something effortful and ongoing, chosen because it is difficult
• Connection — social density is not a soft factor in this literature

The constraint here is materials and recovery, not stimulation. Recover. supports the consolidation window through magnesium glycinate and L-theanine; the sleep is doing the work, and the formulation is removing obstacles to it.

For the inflammatory side of the same pathway, BIOME CLEANSE and the Gut Restore Synbiotic address barrier integrity, which is where most chronic inflammatory load originates.

See the supplements →

Recover. carries L-theanine at a functional dose. Taken in the morning alongside caffeine it steadies attention; taken at night it lowers pre-sleep arousal. Same molecule, different use, depending on timing.

Magnesium glycinate is included because magnesium insufficiency is common and its first symptoms are frequently cognitive and sleep-related rather than muscular.

Energize. supplies NAD+ directly at 500 mg per tablet. Revive. takes the precursor route with NMN alongside quercetin, CoQ10, trans-resveratrol and B vitamins — the pairing rationale being that raising NAD+ is only useful if the enzymes that consume it are also active.

Vitality. covers the antioxidant and hepatic side with milk thistle, glutathione and alpha-lipoic acid.

See longevity & energy →

They do not treat, prevent or reverse any neurological condition, and they are not a substitute for sleep, exercise or medical care. Several actives — NMN and trans-resveratrol among them — lack sufficient pregnancy data for us to recommend them at all in that context.

If you take prescription medication, particularly blood thinners or thyroid medication, bring the label to your pharmacist. Everything is printed on it, which is the point.

• Aerobic exercise — associated with preserved hippocampal volume and better executive function; the effect sizes are unusually good for a lifestyle intervention
• Sleep — glymphatic clearance during deep sleep removes metabolic waste including amyloid species
• Blood-pressure control in midlife — among the clearest modifiable risk factors in longitudinal cohorts
• Hearing correction — large observational association with reduced cognitive decline; cheap and reversible to test
• Education and cognitive complexity — reserve accrues and continues to accrue

• Omega-3 (DHA/EPA) — clearest benefit in those starting deficient; less compelling in already-sufficient populations
• NAD+ precursors — biomarker moves reliably, functional cognitive endpoints not yet established
• Magnesium — insufficiency is common and correcting it improves sleep architecture, which is itself tier one
• Polyphenols — consistent epidemiology, inconsistent trials, likely partly mediated through the microbiome
• Time-restricted eating — good animal data, mixed and confounded human data

A great deal of what is marketed aggressively sits here — senolytics, most peptide protocols, high-dose single antioxidants. The mechanisms are often genuinely interesting. The human outcome data is absent, and in some cases the animal data suggests a dose-dependent downside.

Our position is that mechanism is a reason to watch a compound, not a reason to sell it as an outcome.

No supplement in this range slows brain ageing, and we will not say it does. What some of them do is narrower: correct a deficiency, restore a declining precursor, or reduce an inflammatory load. Those are real and measurable, and they are also modest next to sleep and exercise.

If a brand tells you otherwise about a capsule, that is a reason to distrust the brand rather than to buy the capsule.