Most cognitive enhancement marketing sells novelty. The compounds with the strongest evidence for acute focus and chronic neuroprotection are decidedly not novel. This guide walks through them in order of evidence strength, then addresses how to think about combining them.
The stacking framework: why combine compounds at all?
A stack is two or more compounds taken together, ideally targeting different mechanisms so the effects compound rather than redundate. The case for stacking is not that single compounds are weak; it is that single-compound effect sizes are usually small, and combining mechanism-distinct compounds at full doses can produce a perceptible additive (occasionally synergistic) shift.
The case against stacking is straightforward: every compound you add multiplies the troubleshooting cost. If you take three nootropics and feel different, you cannot tell which one moved the needle, which one is doing nothing, and which one (if any) is actively interfering with the others. The dose-response and timing curves of three compounds taken together are not the same as the sum of three compounds taken individually.
The honest middle ground: keep your stack small, add one compound at a time with a 2-week washout between changes, and measure something concrete (reaction time, a memory task, a subjective focus score). The goal of a stack is to reach a stable, predictable cognitive state, not to maximize compound count.
Tier 1: well-evidenced, low-risk
Caffeine + L-theanine is the most-replicated acute focus stack in the literature. Haskell 2008 (n=27) found 50 mg caffeine + 100 mg L-theanine improved attention-switching and reduced subjective tiredness compared with caffeine alone ( Haskell et al. 2008, n=27 ). Owen 2008 (Nutritional Neuroscience) replicated with similar effect sizes in a larger cohort ( Owen et al. 2008 ).
The mechanism is complementary. Caffeine antagonizes adenosine receptors, lifting alertness but adding sympathetic tone (jitter, BP, raised heart rate). L-theanine, an amino acid found in tea, increases alpha-wave activity and modulates GABA, taking the edge off the caffeine without blunting the alertness gain. The 1:1 ratio at 100 to 200 mg of each is the dose range most studies use. Stronger doses (300 mg + 300 mg) blunt rather than enhance the effect.
Onset is 15 to 30 minutes; duration 3 to 4 hours. Tolerance develops to the alertness component within 2 to 4 weeks of daily dosing; periodic 3 to 7 day washouts restore the acute response. Take with breakfast or 30 to 60 minutes before a focus block; avoid past 8 hours before bed because caffeine half-life (5 hours) is long enough to interfere with sleep architecture even when subjective sleep onset feels fine.
Creatine 5 g/day is primarily a muscle supplement, but a growing literature suggests cognitive benefit under stress: sleep deprivation, hypoxia, high cognitive load. Rawson and Venezia 2003 reviewed creatine effects on elderly cognition with modest positive signals ( Rawson & Venezia 2011 ). The mechanism is bioenergetic. The brain consumes ~20% of body energy at rest, and the phosphocreatine system buffers ATP under acute metabolic stress just as it does in skeletal muscle. Brain phosphocreatine pools turn over on the same timescale as muscle, which makes the cognitive-under-stress signal mechanistically coherent.
In well-rested omnivores at typical maintenance doses, the cognitive effect is small and inconsistent. Use creatine for the muscle case; treat cognitive benefit as a bonus that may show up under sleep debt or vegetarian baseline. See the creatine compound entry for the full evidence map.
Sleep, exercise, sunlight. Embarrassingly effective. A single bad-sleep night reduces working memory performance by 20 to 40% in controlled studies; no nootropic closes that gap. The anyone-selling-you-a-stack-while-ignoring-these heuristic is the cleanest single filter for telling cognitive-enhancement marketing from honest practice. The boring three are also free, which is useful in a market structured to sell you the alternative.
Tier 2: situational
Modafinil. 100-200 mg in the AM produces reliable 8-12 hour wakefulness extension with less rebound than stimulants. Good acute data in shift-worker and sleep-deprivation models; long-term data in healthy cognitive enhancement is thin. Schedule IV in the US; prescription-gated in most jurisdictions.
Racetams (piracetam, aniracetam, oxiracetam, phenylpiracetam). The original "nootropic" class. Human RCTs in healthy adults are sparse and effect sizes in clinical populations (dementia, stroke recovery) are modest. Piracetam is the most-studied; it's also the least-stimulating, which is either a feature or a dealbreaker depending on your goal.
Lion's mane (Hericium erinaceus). One notable RCT in mild cognitive impairment (Mori 2009 Mori K et al. 2008, n=30 , n=30) showed cognitive score improvements on supplementation that regressed after washout. Preclinical NGF-promoting mechanism is interesting; dose-response in humans is poorly characterized.
Ashwagandha. Better evidence for stress/anxiety reduction than for direct cognitive enhancement. If chronic stress is eating your cognition, this works on the upstream problem. See the ashwagandha compound entry.
Nicotine. Real cognitive effects on attention; commonly overlooked because of its other associations. Gum or patch, not combustible. Legal but not child-proof.
Tier 3: experimental
Cerebrolysin, Semax, Selank, methylene blue, the ever-expanding peptide nootropic list. Mostly regulatory grey zones in the US, EU, and most of Asia. Human RCTs in healthy adults are scarce. Covered in the compound directory for completeness, not endorsement.
The counter-view
Andrew Huberman is more enthusiastic about specific compound protocols (alpha-GPC, phenylpiracetam, acetyl-L-carnitine) than the evidence base supports in strict terms; many of these have mechanistic plausibility but thin trial data at the doses he recommends. Scott Alexander (SlateStarCodex) is more skeptical and would argue most nootropics are "works just barely enough that motivated reasoning does the rest". The honest middle: caffeine + L-theanine + creatine + sleep, and be skeptical of anything newer than those.
Goal-specific stacks
Different cognitive goals respond to different mechanism mixes. None of these are prescriptions; they are starting points for an n=1 trial.
| Phase | Dose | Notes |
|---|---|---|
| Daily focus (work / study) | 100-200 mg caffeine + 100-200 mg L-theanine, AM | Onset 20 min, duration 3-4 hr. Add 5 g creatine for chronic baseline. Skip 1-2 days/week to manage tolerance. |
| Sleep-deprived cognitive crunch | Caffeine + L-theanine + creatine 5 g + 200 mg modafinil | Modafinil if available + acute justified; 8-12 hr wakefulness. Acute use only; long-term data thin. |
| Chronic stress + foggy thinking | Ashwagandha 300-600 mg KSM-66 + creatine 5 g + sleep fix first | Stress reduction is upstream of cognition here; the stack ladders down to cognition via better sleep + lower cortisol. |
| Memory / learning under load | Caffeine + L-theanine + lion's mane 750-1000 mg + creatine 5 g | Lion's mane evidence is in MCI populations; effect in healthy adults is mostly mechanistic. 8-week minimum to assess. |
| Anti-fatigue without stimulant edge | L-theanine 200 mg + lion's mane 750 mg + creatine 5 g | Caffeine-free option for stimulant-sensitive users. Smaller acute lift but sustainable. |
The pattern is the same across goals: one acute lever (caffeine, modafinil, or none), one chronic-baseline lever (creatine, ashwagandha, or lion's mane), and the boring three (sleep, exercise, sunlight) that nothing replaces.
How to run an 8-week n=1 stack experiment
Most "nootropic doesn't work for me" reports come from poorly-designed self-trials: too many compounds at once, no measurement, premature evaluation. A clean n=1 protocol takes 8 weeks per compound and gives you a defensible answer.
- Pick one compound to test. Skip stacking for the first round.
- Establish baseline (week 0): log a measurement once daily for 7 days before starting. Examples: a 2-minute reaction-time test (any free web tool works), a working-memory task (n-back app), or a 3-question subjective focus score (1 to 10, energy / clarity / mood).
- Start at the lower end of the dose range (week 1). Maintain measurement cadence.
- Plateau check (week 4): by now you should see a stable signal in your measurement, either a clear shift or no movement. If no movement at the lower dose, escalate to mid-range for weeks 5 to 6.
- Washout (week 7): stop the compound for 7 days. Continue measurement. The reverse-direction shift (or absence of one) is the cleanest evidence the compound was doing something.
- Decision (week 8): keep, drop, or layer in the next compound. If you keep it, it stays as the new baseline; the next experiment tests something on top of it.
This is not the same as a randomized blinded trial. It will not protect you from placebo, expectancy, or confounders (sleep, stress, nutrition). Run during a stable life stretch, not during a job change or holiday. The goal is to filter compounds that produce zero detectable signal for you, not to prove statistical efficacy.
What not to stack
Some combinations either redundate (waste money) or interact badly. The most common mistakes:
- Caffeine + modafinil: stacks the cardiovascular load (BP, HR) without proportional cognitive benefit. Modafinil alone is sufficient for the wakefulness use case.
- Three-plus stimulants (caffeine + nicotine + yohimbe + tyrosine): inverted-U for adrenergic activation; you pass the peak and get worse.
- Racetams + acetylcholine precursors at high dose: the Tier-2 racetam protocols often pair piracetam with alpha-GPC or citicoline. At high doses this can produce headache and irritability via acetylcholine excess. If you stack, start at the lowest of both doses.
- Ashwagandha + benzodiazepines or other GABAergics: additive sedation. Same caution applies to ashwagandha + alcohol.
- 5-HTP + SSRIs: rare but real serotonin-syndrome risk. Skip if on SSRI / SNRI.
- Modafinil + hormonal contraception: modafinil induces CYP3A4, lowering hormonal contraceptive efficacy. Use a backup method or a non-CYP3A4-impacted nootropic.
Run a quick interaction check on every compound you add. The compound directory entries list known interactions per compound.
FAQ
Stacking principles
- Start with a single variable. Add one compound at a time, 2-week washouts between changes.
- Measure something. Reaction time, a simple memory task, subjective focus score. Even a weekly 3-question self-report beats vibes.
- Dose-escalate slowly. More is not more; several nootropics have inverted-U dose-response curves.
- Stop if nothing changes after 4 weeks.
