Peptide Health Lab

Goal

Cognitive Function

Also known as: Cognition, Brain fog, Memory

An honest map of a goal this library cannot serve. No compound reviewed on this site has human evidence of improving cognition, and the three that were tested against a cognitive endpoint each returned a null result.

Last reviewed August 3, 2026

What "cognitive" actually means in the research

This page begins with its conclusion, because burying it would waste your time: no compound reviewed on this site has published human evidence of improving cognition. Three have been tested against a cognitive endpoint. All three came back null.[1][3][6]

That is not the same as "no research exists," and the difference matters. It would be easy to write a page here about mechanisms: mitochondrial energetics, androgen receptors in the hippocampus, thyroid hormone in the developing brain. It would be just as easy never to mention that when the mechanism was carried into a trial and pointed at a memory test, nothing happened. This page is organized around the trials.

Note also what the trials were allowed to study. Cognitive research recruits people with something wrong: a testosterone concentration below 275 ng/dL combined with objectively impaired memory, treated hypothyroidism, or a diagnosis of mild cognitive impairment.[1][7] Nobody has run a published randomized trial of any compound in this library in healthy adults with normal laboratory values, measuring cognition. So the question most readers arrive with, will this make my thinking sharper, has not been asked, and the closest approximations to it were answered no.

What the evidence supports

That the compounds do things. This is worth saying plainly, because "no cognitive benefit" is sometimes heard as "the trial failed." These were not failed trials. Testosterone gel raised serum testosterone into the mid-normal range for men aged 19 to 40 and produced significant increases in sexual activity, sexual desire, and erectile function, along with modest improvements in mood and depressive symptoms.[2] Nicotinamide riboside raised blood NAD+ by 2.6-fold in older adults with mild cognitive impairment and, in a second trial, lowered plasma phosphorylated tau 217 by seven percent relative to placebo.[6][7] The molecules arrived, engaged their targets, and moved biomarkers.

That treating hypothyroidism is a real intervention with real limits. Desiccated thyroid extract and levothyroxine both keep serum thyrotropin within the reference range, and in a randomized crossover comparison across three regimens the biochemistry behaved as expected.[4] Where those trials found something, it was a preference and a small weight difference: roughly half of patients preferred the extract, and those who did had lost about three pounds on it.[3] Preference is a legitimate finding. It is not a cognitive one.

That a well-designed null result is information. The most-cited body of evidence on this page exists because someone funded twelve academic medical centres to answer a question properly, and the answer was no.[1] That is more useful to a reader than a mechanism paper, and it is the reason this page is short rather than speculative.

Where the evidence is weak

The primary endpoint result, stated exactly. Among 493 men with age-associated memory impairment and low testosterone, a year of testosterone produced an adjusted estimated difference in delayed paragraph recall of -0.07, with a confidence interval running from -0.92 to 0.79 and a p-value of 0.88. Mean scores rose from 14.0 to 16.2 in the testosterone group and from 14.4 to 16.5 in the placebo group. Both arms improved by the same amount, which is what taking the same test three times looks like. Visual memory, executive function, and spatial ability were all non-significant too.[1]

The thyroid comparisons found nothing to choose between. The 2013 crossover trial reported no differences in symptoms or neurocognitive measurements between desiccated thyroid extract and levothyroxine.[3] The 2021 three-arm crossover, which put the Wechsler Memory Scale IV among its primary outcomes, reported no differences for primary or secondary outcomes except a minor increase in heart rate on the extract.[4]

Correcting a borderline lab value did not improve symptoms. In 737 adults aged 65 and over with persisting subclinical hypothyroidism, levothyroxine lowered thyrotropin from a baseline mean of 6.40 to 3.63 mIU/L while placebo drifted to 5.48. The between-group difference in the Hypothyroid Symptoms score at one year was 0.0. No secondary outcome favoured treatment.[5] Anyone reasoning from "my number is slightly off, so my thinking is slightly off, so fixing the number will fix my thinking" has to get past that result first.

The NAD+ cognitive endpoints were pre-specified and did not move. In the mild-cognitive-impairment pilot, the primary outcome was change in the Montreal Cognitive Assessment, and it remained stable throughout the study alongside the other neurocognitive and psychometric measures; the authors state directly that nicotinamide riboside did not alter cognition.[6] In the crossover trial, the primary efficacy outcome was a neuropsychological battery and there were no significant between-group differences on it, on digital gameplay scores, or on step counts, even while the tau biomarker moved.[7] The systematic review of the wider NAD+ literature describes exactly this pattern: reliable target engagement, heterogeneous and often null functional outcomes.[8]

Where the optimistic reading comes from, and why it is weaker. A 2025 systematic review and meta-analysis pooling 14 studies of androgen replacement reported a standardized mean difference of 0.454 for overall cognition, with larger effects for executive function and memory. Its own authors report significant publication bias on Begg's and Egger's tests, describe the effect sizes as modest, and attribute moderate heterogeneity to varying cognitive instruments and treatment protocols.[9] A pooled estimate drawn from a literature with detected publication bias is weaker evidence than a single adequately powered trial designed to answer the same question. That trial found nothing.[1]

Nothing here has been tested for enhancement, and no amount of aggregation changes that. Three compounds with three null cognitive results do not combine into a case for a fourth. That is the failure mode this page exists to avoid: adding compounds until the section looks substantial. The library holds nothing else with human cognitive-efficacy evidence, so nothing else is mapped.

How to decide between these

The honest version of this section is that there is nothing to decide between. None of the three compounds mapped here has a cognitive result to weigh against another's, so a comparison would be a presentation choice rather than a finding.

What is worth deciding is a prior question: whether the thing you are calling a cognitive problem has a cause that has been identified. Untreated overt hypothyroidism, obstructive sleep apnoea, depression, medication side effects, and anaemia all present as difficulty thinking, and all have workups and treatments with real evidence behind them. That is the branch where the literature is strong. The branch this page covers, reaching for a compound against a symptom without a diagnosis, is the branch where the largest and best-run trials came back empty.[1][5]

If a laboratory abnormality is found and treated, the evidence supports treating it on its own terms rather than as a cognitive intervention.[4] The stack assessment on this site is built to organize that conversation for a provider visit, and on this particular goal its most useful output is the observation that the library has nothing to offer.

Questions to bring to a provider

The productive version of this conversation is not "which peptide helps cognition." It is "here is what I am experiencing, what could be causing it, and what has been shown to change it."

  • Which of these is actually being noticed, and over what period did it change: word-finding, short-term memory, sustained attention, or processing speed?
  • What causes are worth excluding before any compound is discussed, and what workup would exclude them?
  • Given that a year of testosterone in men who were both hypogonadal and memory-impaired produced no measurable cognitive change, what would justify expecting one here?[1]
  • If a thyroid change is proposed, what is the target, and how would we know whether it improved anything beyond the laboratory value?[5]
  • If an NAD+ precursor is proposed, what outcome would it be expected to change, given that the trials measuring cognition found no change?[6]
  • What objective measure would we agree on in advance, so that "it is working" is answerable later?

Anti-doping status

One compound mapped on this page is named on the World Anti-Doping Agency Prohibited List in force for 2026:

  • Testosterone: section S1.1, anabolic androgenic steroids. It is named directly in the list of anabolic androgenic steroids prohibited when administered exogenously, and substances in class S1 are prohibited at all times, in and out of competition. They are non-Specified Substances, which carries different sanctioning consequences from a Specified Substance. That distinction is one to raise with a national anti-doping organization rather than settle from a web page. That testosterone is an approved prescription medicine, lawfully dispensed, changes none of this; a therapeutic use exemption is the mechanism that exists for that situation.

Neither desiccated thyroid extract nor NAD+ precursors are named in the 2026 list. Section S0, however, covers any pharmacological substance not addressed by a later section that has no current approval by any governmental regulatory health authority for human therapeutic use, and the substances it lists are given as examples rather than as a closed set. Whether a particular unapproved preparation falls inside it is a question for a national anti-doping organization, not for a reference page.

The section assignment above was read from the published Prohibited List itself rather than from the research literature. A review article is a source for what the literature shows and never a source for what a regulator has done. The trial evidence and the anti-doping position happen to point the same way here: raising testosterone in older men for a year produced a moderate benefit for sexual function and none for cognition.[2] The list is reissued annually and its sections renumber between editions, so a competing athlete should confirm the current year's list rather than this page.

What the evidence supports for this goal

Grades describe how strong the evidence is; the line under each grade describes what kind of studies it is. How we grade evidence.

Testosterone and memory in older men
Human RCT evidence human RCT

The Cognitive Function Trial within the Testosterone Trials randomized 493 men aged 65 and over who had a serum testosterone below 275 ng/dL and met criteria for age-associated memory impairment. After a year of testosterone gel titrated into the young-adult reference range, there was no significant change from baseline in delayed paragraph recall against placebo. The adjusted estimated difference was -0.07 and both groups improved identically. Visual memory, executive function, and spatial ability showed no significant difference either. Scores rose in both arms, which is what a practice effect looks like.

[1] [2]

Desiccated thyroid extract and neurocognitive outcomes
Human RCT evidence human RCT

Two randomized double-blind crossover trials measured this directly. In the first, 70 hypothyroid patients stable on levothyroxine were crossed over to desiccated thyroid extract for 16 weeks in each arm with biochemical and neurocognitive testing at the end of each period; there were no differences in symptoms or neurocognitive measurements between the two therapies, though patients lost about three pounds on the extract and roughly half preferred it. In the second, 75 patients were randomized across levothyroxine, levothyroxine plus liothyronine, and desiccated thyroid extract for 22 weeks with the Wechsler Memory Scale IV among the primary outcomes, and there were no differences on any primary or secondary outcome except a minor heart-rate increase.

[3] [4]

NAD+ precursors and cognition
Human RCT evidence human RCT · review

A placebo-controlled randomized pilot gave 20 older adults with mild cognitive impairment nicotinamide riboside or placebo for ten weeks; blood NAD+ rose 2.6-fold, confirming the molecule reached its target, and the Montreal Cognitive Assessment and the other neurocognitive and psychometric measures remained stable throughout. A second crossover randomized trial in 46 adults with subjective cognitive decline or mild cognitive impairment used a neuropsychological battery as its primary efficacy outcome; plasma phosphorylated tau 217 fell relative to placebo, and cognition did not change on either conventional or digital assessment. The systematic review covering the field reaches the same aggregate conclusion for healthspan-relevant outcomes generally.

[6] [7] [8]

Treating a borderline laboratory value to improve symptoms
Human RCT evidence human RCT

A double-blind, randomized, placebo-controlled, parallel-group trial in 737 adults aged 65 and older with persisting subclinical hypothyroidism found that levothyroxine lowered thyrotropin as intended and produced no difference against placebo in the Hypothyroid Symptoms score or the Tiredness score at one year, with no beneficial effects on secondary-outcome measures either. The authors concluded that levothyroxine provided no apparent benefit in this population. It is the clearest available demonstration that normalizing a number and improving how someone functions are separable results.

[5]

Cognitive enhancement in a healthy adult
Anecdotal reports only human RCT · review

There is no evidence in this library, of any grade, for improving cognition in a person without a diagnosis. Every trial above required an abnormality to enter: a testosterone below 275 ng/dL together with objective memory impairment, treated hypothyroidism, or mild cognitive impairment. No compound reviewed on this site has been given to healthy adults with normal laboratory values and measured against a cognitive outcome in a published randomized trial. The pooled literature that reads more positively is assembled from heterogeneous studies and carries publication bias its own authors detected and reported.

[1] [6] [9]

Compounds people map to this goal

Each card carries the compound's FDA status and its overall evidence grade, plus why it shows up on this goal in particular. Follow the card to the full library page for the claim-by-claim evidence review and what published research reported. Dosing and protocol ranges are never on this page.

Testosterone

Also known as: Testosterone replacement therapy, TRT

FDA approved Human RCT evidence

A research reference for testosterone therapy, the one compound in this library with a large randomized trial literature, and a prescription hormone that requires a diagnosis, a prescriber, and ongoing lab monitoring.

Why it's on this page

Mapped here for a specific reason: the largest randomized trial ever designed to ask whether testosterone improves memory in older men found that it does not. That trial enrolled men who were both hypogonadal by laboratory criteria and objectively memory-impaired, which is the most favourable population anyone has tested, and it still separated from placebo on nothing. Prohibited in tested sport at all times as an anabolic androgenic steroid under WADA section S1.

Last reviewed August 2, 2026

NP Thyroid

Also known as: Desiccated thyroid extract, Natural desiccated thyroid, Porcine thyroid extract

Rx via compounding Human RCT evidence

A research reference for NP Thyroid, a porcine desiccated thyroid extract that the FDA classifies as an unapproved biological product marketed without the required license, and that showed no advantage over levothyroxine on symptoms or cognition in two randomized crossover trials.

Why it's on this page

Mapped because "brain fog" is the complaint that brings most readers to this page, and desiccated thyroid extract is the intervention most often proposed for it. Two randomized double-blind crossover trials measured neurocognitive performance directly and found no advantage over levothyroxine. It is here as the negative result, not as an option.

Last reviewed August 2, 2026

NAD+

Also known as: Nicotinamide adenine dinucleotide, NAD

Rx via compounding Anecdotal reports only

A research reference for NAD+, a coenzyme whose oral precursors have a real randomized-trial literature, and whose injected form has none.

Why it's on this page

Mapped because NAD+ precursors are the most heavily marketed molecule in the cognitive-longevity category and because, unusually, the question has actually been asked. Two randomized placebo-controlled trials gave oral nicotinamide riboside to older adults with mild cognitive impairment and measured cognition as a stated outcome. Neither found any change in it.

Last reviewed August 2, 2026

How to decide between them

These are the questions that actually change the answer. Work through them with a licensed provider. None of them can be answered by a page that does not know your labs, your history, or your medications.

  1. Is there a diagnosed condition here, or are you trying to make a healthy brain work better?

    This is the question that decides whether this page has anything for you, and for the second case the answer is no. Every human trial cited here required a diagnosis or a documented abnormality to enrol: low serum testosterone plus objectively impaired memory, treated hypothyroidism, mild cognitive impairment. Not one compound in this library has been tested for cognitive enhancement in a healthy adult with normal labs, in any published randomized trial. If that is what you are looking for, nothing here has been studied for it and this page cannot become useful by trying harder.

    Human RCT evidence [1] [6]

  2. Has anyone tested your exact question, this compound against this cognitive outcome, or only the compound?

    Applying that filter is what makes this page thin, and the thinness is the finding. Three of the compounds in this library have been carried into a trial with a cognitive endpoint attached. All three came back null: no change in delayed paragraph recall, visual memory, executive function, or spatial ability with a year of testosterone; no difference in neurocognitive measurements between desiccated thyroid extract and levothyroxine; no change on the Montreal Cognitive Assessment or on a neuropsychological battery with ten weeks and with eight weeks of nicotinamide riboside. A null result from a well-run trial is stronger information than an absence of trials, and it points the same way.

    Human RCT evidence [1] [3] [6]

  3. If a laboratory value is low, does correcting it reliably fix how you think?

    Not on the evidence here, and this is the assumption most worth interrogating before anything else. A double-blind trial randomized 737 adults aged 65 and over with persistent subclinical hypothyroidism to levothyroxine or placebo; the thyrotropin level came down as intended and neither hypothyroid symptoms nor tiredness improved, with no benefit on any secondary measure. Restoring a number is a different intervention from restoring a function, and the trials that separated the two found the number moving alone.

    Human RCT evidence [5] [4]

  4. Are you weighing the largest trial designed for the question, or a pooled estimate assembled afterwards?

    Both exist here and they disagree, which is worth understanding before you meet them separately. A 2025 systematic review pooled 14 heterogeneous studies of androgen replacement and reported moderate benefits for memory and executive function. Its own authors reported significant publication bias on both Begg's and Egger's tests and described the effect sizes as modest. The single trial purpose-built to answer the question randomized 493 memory-impaired hypogonadal men for a year and found an adjusted difference on its primary memory outcome of -0.07. When a designed trial and a pooled estimate with detected publication bias point in opposite directions, the designed trial is the one to reason from.

    Human RCT evidence [9] [1]

  5. Do you compete in a tested sport, at any level?

    Then one of the three compounds mapped here is already unavailable to you regardless of what the evidence said. Testosterone is named in the anabolic androgenic steroid section of the WADA Prohibited List in force for 2026 and is prohibited at all times, in and out of competition. The trial evidence is a separate matter and it points the same direction: raising testosterone in older men for a year produced a moderate sexual- function benefit and no cognitive benefit at all. Read the anti-doping section at the end of this page, and confirm the current year's list rather than this page.

    Human RCT evidence [2]

Questions for your provider

Bring this goal page to a licensed provider. A provider can order labs, review your medications and history, and tell you whether anything here applies to your situation. This goal page cannot. Peptide Health Lab does not prescribe and does not sell peptides.

A goal is not a diagnosis. The most useful version of this conversation usually starts with what is actually driving the symptom, not with which compound to try.

Citations

9 sources · every identifier checked against PubMed

Before you act on any of this

This page is educational only and is not medical advice. It does not diagnose, treat, or prescribe. Review it with a licensed provider before making any health decision. Peptide Health Lab does not sell peptides.

How we grade and verify evidence · Terms of Use · Privacy Policy

Want a structured starting point for that conversation?

Open the Stack Builder