How to Track Nootropic Effectiveness: A Self-Testing Method

By Peter Benson, Cognitive Enhancement Researcher  ·  Updated October 2026  ·  Citations re-verified

Short answer: Don’t rely on how you feel. Take four baseline measurements on a simple objective test, change one thing, and treat a change as a signal only if your scores move clearly outside your baseline range. Fast-acting compounds can be compared on and off over two weeks; slow ones need 8 to 12 weeks and then a period without them.

Most people never measure whether a supplement works. They take something for a fortnight, notice a good Tuesday and conclude it helps, or notice nothing and move on to the next one. Both conclusions come from the same source: a vague impression of how the last few weeks felt.

That impression is unreliable, and there is good research on why. This article explains what goes wrong and sets out a method anyone can run at home. It is not as strong as a clinical trial, and it says where it falls short. For which compounds are worth testing, see the Nootropics & Supplements guide and best nootropics for beginners.

Why “I feel sharper” is weak evidence

Winkler and Hermann (2019) gave 75 healthy adults an inert nasal spray. One group was told it was a performance-enhancing drug, another that it was a substance that would impair them, and a third group received nothing. Those expecting enhancement rated their own performance as far more improved than those expecting impairment, a large effect (d = 1.16), and reported feeling less tired (d = 0.96).

Their measured performance did not change. The authors had predicted that expectation would shift scores on tests of alertness, working memory and sustained and divided attention. It did not, in either direction.

In practice: belief moves how you rate yourself and leaves your test scores alone. If self-rating is your only measure, you are measuring what you hoped for when you took the capsule.

Feelings also miss real impairment

The error runs the other way too. Van Dongen and colleagues (2003) restricted 48 healthy adults to four, six or eight hours in bed for 14 nights. Performance in the restricted groups fell steadily. Sleepiness ratings rose at first, then showed only small further increases, and did not separate the six-hour group from the four-hour group. People who were measurably getting worse felt much the same. That is also why sleep has to be logged during any test you run; see sleep and cognitive performance.

Four things that create a false result

Switching to an objective test is necessary and not sufficient. Four things can still produce an improvement that has nothing to do with what you are taking.

1. Practice effects

You get better at cognitive tests by taking them. Bartels and colleagues (2010) tested 36 healthy adults seven times over a year with no intervention. The largest gains came between the first and second sessions, two to three weeks apart: about 14% in executive function, 13% in learning and memory and 12% in attention. Scores kept improving while testing was frequent, through the third month, and then levelled off.

The gains do not go away. Rijnen and colleagues (2018) tested 158 healthy adults on a computerised battery and retested them at three and twelve months. Cognitive flexibility, processing speed and reaction time were higher at three months, and there was no significant change between three and twelve. Scores settle at a new, higher level.

How large the effect is, and how long it lasts, varies from test to test. The Bartels authors recommend choosing tests that are less prone to it.

In practice: an improvement in your first few sessions is the expected result of repetition. Several baseline sessions before you start absorb the steepest part of it. Where a test offers different versions, such as a fresh word list each time, use them, so that you are not simply learning the material. That reduces the problem without removing it.

2. Regression to the mean

People tend to start a supplement when they feel at their worst. Unusually bad measurements are usually followed by ones closer to average, whatever you do. Barnett, van der Pols and Dobson (2005) describe how this makes natural variation look like real change, and recommend taking several baseline measurements to reduce it.

In practice: build your baseline from measurements spread over two weeks, not from the bad week that prompted the purchase.

3. Other changes you did not record

Sleep, alcohol, illness, workload and stress all move test scores. If you are not recording them, any one can pass for the effect of your supplement. A minimum log is hours slept, caffeine amount and timing, alcohol, and a 1–10 stress rating. Change one thing at a time, the same rule that governs sensible stacking.

4. Expectation

Covered above, and the hardest to remove, because you know what you took. An objective test reduces its influence. Blinding removes it, and for fast-acting compounds blinding is achievable at home.

What counts as a real change?

Clinical psychology has a formal answer. The Reliable Change Index, introduced by Jacobson and Truax (1991), divides the difference between two scores by the standard error of that difference; a value beyond about 1.96 indicates a change unlikely to be measurement error alone. The Rijnen study above published formulae of this kind for exactly this reason: without them, a change in one person’s score cannot be interpreted.

Home tests rarely publish the figures needed for that calculation, so the practical substitute is a personal baseline band: the lowest and highest of your baseline scores. Be clear about what this is. Four measurements give a rough band that would shift if you took four more, and it takes no account of how reliable the test is. A score outside the band is a screening signal that deserves a closer look. It is not statistical evidence of reliable change.

Decide what “working” means before you start

Choose one main outcome in advance. If you are testing Bacopa for memory, a memory test decides the result, not whichever of several scores happens to rise. Record other measures if you like, but do not promote one to the headline after seeing the data. A review behind the reporting guideline for N-of-1 trials found that 79% of published reports did not say which outcome was the primary one (Vohra et al., 2015).

Decide the second number too: how much improvement would make this supplement worth its cost, the daily habit and any side effects? A change can be real and still too small to matter to you.

Choose your design by how fast the compound acts

A before-and-after comparison is not an N-of-1 trial. Zenner, Böttinger and Konigorski (2022) define N-of-1 trials as “multi-crossover self-experiments”: you switch between conditions several times, which is what separates an effect from a trend over time. Whether you can do that depends on how quickly the compound acts and wears off.

Fast-acting compounds: on and off, over days to two weeks

If the effect appears within an hour or two and is gone by the next day, as with L-theanine and caffeine, you can alternate. Take it on some mornings and not on others, test each time, and compare the two sets across ten or more mornings. Do not simply alternate day by day. Set the order at random before you start, for example by tossing a coin for each pair of days to decide which comes first, so that a pattern in your week cannot line up with the pattern of dosing.

You can also blind yourself. Buy empty capsules, fill half with the compound and half with something inert, and ask someone else to load them into a pill organiser in the random order and keep that order written down. Open the record only when all testing is finished. This step removes expectation, and very few people do it.

Slow compounds: baseline, 8 to 12 weeks, then withdrawal

Compounds such as Bacopa monnieri were tested over 12 weeks in their trials and cannot be switched on and off. The best available design is a reversal: baseline, 8 to 12 weeks of daily use, then several weeks without it, testing throughout.

This is stronger than before-and-after. For an effect you would expect to be reversible, a move back toward baseline after you have been off the compound long enough strengthens the case that it contributed. The reverse does not hold: scores that stay up do not prove the earlier gain was practice, since some effects may persist and a short break may not be long enough. It is much weaker than a blinded comparison. You know what you are taking, there is one reversal instead of several, and a change in season or workload across three months can still mislead you. Treat its results as suggestive.

Self-testing methods ranked by strength

New to this? The Getting Started roadmap puts these steps in order.

Frequently asked questions

Scientific references