Brain Supplements, Gut Bacteria, Tools & Games: The Honest Evidence Guide

Last updated: September 2026 · Evidence-first — verdicts follow the research, not the ads. Product links are Amazon affiliate links (we may earn a commission). Not medical advice.

AI, technology & your brain: augmentation, not IQ points

New: we turned this section into a full guide — Use AI to Get Smarter, with the research, four practical protocols, copy-paste Socratic prompts, and a sample AI thinking session.

Does using AI make you smarter?

Claimed: AI assistants upgrade your thinking itself.

Evidence: No study shows AI use raising IQ or general cognitive ability — and the evidence on passive use points the other way. Gerlich (2025, Societies, n=666) found frequent AI-tool use significantly correlated with lower critical-thinking scores, mediated by cognitive offloading, steepest in younger users. Lee et al. (2025, Microsoft Research + Carnegie Mellon, n=319 knowledge workers) found higher confidence in generative AI associated with less critical thinking applied to its output — while confidence in one's own ability predicted more scrutiny. The authors warn of "deterioration of cognitive faculties that ought to be preserved."

Verdict: Nothing here raises IQ; heavy, passive AI use is associated with less thinking, not more. The mechanism is offloading — handing the tool the reasoning, not just the lookup.

The "extended mind" argument — the serious case for AI

Philosophers Andy Clark and David Chalmers argued in 1998 ("The Extended Mind") that when an external process does work we'd call cognitive if done in the head, it can count as part of the cognitive system — the parity principle. A calculator never raised anyone's arithmetic IQ, yet numerate people with calculators vastly outperform those without. AI is the most aggressive version yet: not just storage (notebooks, search) but active reasoning, coupled to you.

Verdict: The strongest honest case: AI multiplies realized intelligence, not measured IQ. But the multiplier has a human-in-the-loop requirement — and the Lee et al. finding above is the empirical shadow of this argument: the coupled system is brilliant until the human stops checking, at which point it degrades to whatever the AI hallucinates.

"Your Brain on ChatGPT" (MIT, 2025)

Evidence: Kosmyna et al. (MIT Media Lab, n=54, EEG across 4 months of essay writing — preprint, small sample, one task): brain-only writers showed the strongest neural connectivity; ChatGPT users the weakest, and 83% of the ChatGPT group couldn't reproduce a single quote from their own essay minutes later (vs. 11% in the brain-only group). When groups swapped, writers who struggled first and then used AI showed higher engagement than those who outsourced from the start. Lead author Nataliya Kosmyna has explicitly cautioned against "ChatGPT makes you dumb" headlines — this measures engagement during the task, not brain damage.

Verdict: The cleanest neural evidence that offloading the thinking means the brain does less thinking in that moment — and that struggling first, then using AI, looks very different from outsourcing first.

AI tutors: better practice, not better brains (yet)

Evidence: The landmark RCT (Bastani et al., PNAS, ~1,000 students): unrestricted ChatGPT raised practice scores 48% — but students scored 17% worse on the unassisted exam than students who never had AI ("the crutch effect"). A hints-not-answers tutor design raised practice 127% and erased most of the harm — yet still produced no positive exam effect vs. control. The largest Khanmigo trial found the binding constraint was take-up: when the tutor withheld answers and asked questions instead, most students simply stopped using it.

Verdict: AI tutors reliably raise performance while the tool is present; evidence they raise learning that survives the tool's removal is thin to null. "Learns the subject better" is not yet proven; "raises IQ" is not even on the table.

The Flynn-effect interpretation

This one is interpretation, not a trial result — labeled as such. The Flynn effect: average IQ scores rose ~3 points per decade through much of the 20th century, likely because environments increasingly rewarded abstract, test-like thinking. The AI-era version: people who learn to decompose problems, verify outputs, and steer AI systems are training precisely those abstract-reasoning muscles — while people who only consume AI output train none of them. Plausible and consistent with history — but it cuts both ways: the era rewards AI fluency and punishes AI dependence, and the same tool produces both.

The GPS precedent: use it or lose it is measured

Dahmani & Bohbot (2020): greater lifetime GPS use predicted worse spatial memory, with a 3-year follow-up showing steeper decline in hippocampus-dependent navigation — and the authors checked direction, finding GPS use drove the decline rather than the reverse. The honest asymmetry: nobody's spatial memory loss made GPS a bad tool; it made navigating without ever practicing the risk. Same for AI.

Four evidence-aligned ways to use AI so it builds thinking

1. Struggle first, outsource second. Draft your own answer before consulting AI, then use it to critique and improve. The Kosmyna cross-over finding makes order of operations the whole game.
2. Make the AI Socratic, not oracular. Ask for hints, counterarguments, and "what am I missing?" rather than final answers — the hints-not-answers tutor design is the only configuration that avoided learning harm.
3. Explain it back. After AI helps, close the tab and restate the reasoning in your own words. If you can't, you offloaded — you didn't learn.
4. Verify on purpose. Treat every AI output as a draft to check. Confidence in the tool is exactly what switches critical thinking off — build a habit of spot-checking one claim per answer.

Want to learn the tool deeply instead of just prompting it? Books on artificial intelligence on Amazon.

Supplements: what's claimed, what the evidence says

Eight of the most-marketed "brain" supplements, in rough order of how seriously the science takes them:

1. Creatine monohydrate

Claimed: "Brain energy" that boosts memory, focus, and mental performance.

Evidence: A 2024 meta-analysis of 16 randomized trials (Frontiers in Nutrition) found significant improvement in memory (moderate certainty) and faster attention and processing-speed task times (low certainty) — but no effect on overall cognitive function or executive function. An earlier memory-focused meta-analysis found the memory effect more robust in older adults.

Verdict: Probably a small, real memory and speed nudge — not higher intelligence, not better reasoning, and nothing here touches IQ.

Safety: Generally well-tolerated in healthy adults; water retention and mild GI upset at high doses. Caution with kidney disease; limited data for children, pregnancy, or very long-term cognitive use.

If you've read the verdict and still want to try it: creatine monohydrate on Amazon.

2. Caffeine + L-theanine

Claimed: The "smart drug stack" — calm, jitter-free focus, flow state, enhanced cognition.

Evidence: Multiple small randomized crossover trials: one (Appetite, 2010) found improved task-switching attention with the combo vs. placebo; a sleep-deprivation trial found improved hit rate and reaction time. A 2023 systematic review (Nutrients) concluded the combo reliably improves aspects of attention and mood, smoothest at moderate caffeine doses. Effects are acute — measured within hours. No long-term or IQ evidence exists.

Verdict: The best-evidenced acute focus combo on this list — it makes you perform better today, mostly by adding alertness and trimming caffeine's jitters. It does not make you smarter.

Safety: Stay within ~400 mg/day caffeine (the FDA's healthy-adult ceiling); pregnancy guidance is much lower (~200 mg/day). Watch for anxiety, insomnia, and heart racing; don't take it late in the day.

If you've read the verdict and still want to try it: L-theanine on Amazon (caffeine you probably already own — it's coffee).

3. Bacopa monnieri

Claimed: An Ayurvedic "memory herb" that improves recall, learning, and focus.

Evidence: A 2014 meta-analysis (9 RCTs) found moderate improvements in attention-task performance and processing speed, with high heterogeneity on memory. A 2012 meta-analysis found improved delayed word recall, most consistently at 300 mg/day for 12+ weeks — acute dosing shows little. But recent evidence is mixed: a 2025 RCT (n=101) found no cognitive difference vs. placebo, and a 2026 network meta-analysis found favorable memory outcomes but no attention effect.

Verdict: One of the more credible herbs — plausible memory improvement after 8–12 weeks of a standardized extract. Inconsistent, slow, and still not an IQ effect. One brand's study does not validate all bacopa products.

Safety: GI complaints are the most consistent issue (nausea, cramps, frequent stools). Possible interactions with anticholinergics, sedatives, blood thinners, and thyroid medications. Heavy-metal contamination has been reported in some low-quality products — source matters.

If you've read the verdict and still want to try it: bacopa monnieri on Amazon.

4. Magnesium

Claimed: "Crosses the blood-brain barrier" to improve memory, reverse brain aging, and fix sleep.

Evidence: Deficiency is common (roughly half of US adults under-consume), so repletion matters. In a 2025 RCT (n=100, 6 weeks), a cognition composite improved vs. placebo — but Raven's Progressive Matrices, a fluid-intelligence measure, showed zero difference (p=0.953). Most "memory" claims trace to animal work, which marketers present as human evidence.

Verdict: Reasonable to try if you're deficient or sleep poorly — the honest mechanism is restoring a deficit, which can sharpen daytime performance. Thin evidence for direct cognitive enhancement in well-nourished people; the Raven's null is the IQ-relevant data point.

Safety: Loose stools/diarrhea (especially citrate and oxide forms); caution with kidney disease and with antibiotics or blood-pressure drugs.

If you've read the verdict and still want to try it: magnesium supplements on Amazon.

5. Rhodiola rosea

Claimed: An "adaptogen" that fights fatigue and sharpens focus under stress.

Evidence: A 2010 systematic review (11 RCTs) found rhodiola "may" benefit physical and mental performance — but studies were small with no independent replication. A 2023 meta-analysis (26 RCTs) found small-to-moderate endurance benefits. The cognitive signal is concentrated in stressed, fatigued, or sleep-deprived populations.

Verdict: Plausible as an anti-fatigue aid when you're running on fumes — not a general cognitive enhancer and certainly not an IQ booster. Evidence quality is modest.

Safety: Activating — take it in the morning or expect insomnia. Check with a clinician if you take antidepressants, diabetes, or blood-pressure medications.

If you've read the verdict and still want to try it: rhodiola rosea on Amazon.

6. Omega-3 / DHA / fish oil

Claimed: "Brain food" that improves memory, focus, and protects against Alzheimer's — claims the FTC has literally sued over.

Evidence: A 2026 phase-2a RCT (n=365, 2 g/day DHA for 24 months) found no change in brain volume or cognitive performance vs. placebo. Two other large RCTs found no effect on psychomotor speed, executive function, memory, or fluid intelligence. A 25-RCT meta-analysis found no effect on global cognition, only a small memory signal in healthy people. In 2014, the FTC barred DHA-based "improves adult memory" claims as deceptive (the BrainStrong order).

Verdict: In healthy adults, fish oil is not a brain booster. A small memory signal in some trials does not equal better cognition — and nothing here raises IQ.

Safety: Fishy burps, nausea, loose stools. Bleeding risk at high doses, especially with anticoagulant or antiplatelet drugs.

If you've read the verdict and still want to try it: omega-3 fish oil on Amazon.

7. Ginkgo biloba

Claimed: Improves memory, focus, and "brain circulation" — one of the most-sold "brain" herbs on Earth.

Evidence: In healthy adults, a 2012 meta-analysis (n≈2,600 across outcomes) found effect estimates non-significant and near zero. Cochrane reviews found probably little or no effect on cognition. The large GEM trial (n=3,069, JAMA 2009) found ginkgo did not prevent dementia or Alzheimer's.

Verdict: For a healthy person seeking a mental edge, ginkgo has essentially no credible evidence — one of the most famous "brain" supplements with the weakest healthy-person data on this list.

Safety: The notable risk is bleeding — it inhibits platelet-activating factor; generally avoid combining with warfarin, aspirin, or NSAIDs, and stop before surgery.

If you've read the verdict and still want to try it: ginkgo biloba on Amazon.

8. Lion's mane mushroom

Claimed: Stimulates nerve growth factor, grows new brain cells, sharpens memory — usually citing petri-dish and mouse studies.

Evidence: Mostly animal and cell-culture work (compounds raise NGF in lab dishes). Human trials are small and mixed: one 2009 trial (n=30) improved cognitive scales but scores reverted after stopping; a 2025 acute RCT (n=18) found no significant effect on global cognition or mood; an 87-person RCT found no outcomes beating placebo. A 2025 systematic review found only 5 small human RCTs amid 15 lab studies.

Verdict: Hype far exceeds data. Interesting, early, and genuinely under-studied — but right now it's mostly mushroom folklore plus animal research, not evidence of cognitive enhancement.

Safety: Short-term use appears well-tolerated; long-term safety is incompletely characterized. Caution for mushroom allergies; no reliable human safety data for pregnancy.

If you've read the verdict and still want to try it: lion's mane on Amazon.

Blood, circulation & oxygen: what's claimed, what the evidence says

Supplements marketed on the idea that more blood flow to the brain means a better-working brain — plus iron, the one where deficiency is genuinely common:

Beetroot juice / dietary nitrate

Claimed: Boosts nitric oxide, increases blood flow to the brain, sharpens thinking.

Evidence: The vascular logic is real, the brain part isn't. A 13-week RCT (n=47, up to 800 mg nitrate/day) found no change in cognition or cerebral blood flow vs. placebo. A 2026 14-day pilot combining beetroot with caloric restriction found improved microvascular function — preliminary and confounded by the diet itself.

Verdict: Improves blood-vessel function; no solid evidence it makes a healthy brain work better.

Safety: Safe as food; lowers blood pressure (caution with hypotension or BP meds); high oxalates — kidney-stone caution.

If you've read the verdict and still want to try it: beetroot supplements on Amazon.

L-citrulline / L-arginine

Claimed: Nitric-oxide boosters for circulation, mental energy, and even mood.

Evidence: Near-zero human evidence for mood or cognition. The one positive: a tiny double-blind crossover trial (n=12 taekwondo athletes, 2016) found citrulline maintained reaction time across matches vs. placebo decline — an anti-fatigue effect, not enhancement. The "antidepressant" arginine paper going around is a rat study, not human evidence. Honest nuance: one observational study found elevated plasma nitric oxide in recurrent depression associated with worse cognition — the story isn't as simple as "more NO = better."

Verdict: Plausible circulation story, near-zero human mood/cognition evidence.

Safety: GI upset; can trigger cold sores; caution with blood-pressure medications.

If you've read the verdict and still want to try it: L-citrulline on Amazon.

Iron

Claimed: Fixes "tired blood," lifts brain fog, improves mood and focus.

Evidence: Real benefit only where deficiency exists. A 2010 meta-analysis of 14 RCTs found attention/concentration improved — and IQ rose ~2.5 points in anaemic groups only, with no effect in non-anaemic people. A 2018 review found reduced fatigue in iron-deficient non-anaemic adults — but a 2020 RCT (n=405 blood donors with low ferritin) found no fatigue or well-being effect despite raising ferritin. Translation: iron fixes iron problems; it is not a general brain or mood enhancer.

Verdict: If you're deficient, correcting it can genuinely help energy and attention. If you're not, extra iron does nothing for your brain — and iron is one of the few supplements here that can genuinely harm you.

Safety: Iron accumulates (haemochromatosis risk, especially in men and postmenopausal women), feeds pathogens, and blocks absorption of thyroid medication and some antibiotics. Get a ferritin blood test before supplementing — do not supplement blindly.

If a blood test says you're low: iron supplements on Amazon.

Gut bacteria & the gut-brain axis: what's claimed, what the evidence says

Probiotics / "psychobiotics" for mood

Claimed: Fix your gut, fix your mood — probiotics as natural antidepressants.

Evidence: The most credible gut-mood evidence — but it's strain-specific. L. helveticus R0052 + B. longum R0175: reduced anxiety/distress in healthy volunteers (2011, 30 days) and reduced depression scores in people with major depression on antidepressants (Kazemi 2019, n=110, 8 weeks). A 2025 meta-analysis found anxiety reduced in healthy/chronically-ill groups — but not in the depressive-disorders subgroup; Bifidobacterium and multi-strain combos worked, Lactobacillus-only did not. A 9-strain trial (2019, n=71) was null.

Verdict: Strain-specific, adjunctive, modest. Genus-level "probiotics help mood" is marketing — the evidence lives or dies on the exact strain.

Safety: Generally safe; rare bloodstream infection risk in immunocompromised people; contamination and mislabeling in cheap products.

If you've read the verdict and still want to try them: probiotics on Amazon — and check the label for the actual strains, not just the genus.

Gut bacteria and metabolism

Claimed: The right gut bacteria "boost metabolism" and melt fat.

Evidence: Mouse effects are dramatic; human translation is a fraction. Lean-donor fecal transplant (n=18, 2012): small transient insulin-sensitivity gain, no weight change. A 2020 trial of weekly FMT capsules in obesity achieved successful engraftment but no clinically significant metabolic effects at 12 weeks.

Verdict: Small transient glucose/insulin nudges, no meaningful weight loss. Rodent extrapolation is exactly how weight-loss probiotics get sold — the human data doesn't back it.

Safety: Fecal transplant is not DIY — the FDA issued a safety alert after a death from drug-resistant bacteria in FMT.

Gut bacteria and cognition / brain activity

Claimed: Probiotics sharpen the brain via the "second brain."

Evidence: Signals only in impaired populations. A meta-analysis of 10 RCTs (n=778, cognitively impaired) found a moderate effect — but didn't report which species were used. A 2026 meta-analysis (24 trials, 1,809 older adults) found high heterogeneity and results unstable to single-study removal. A 20-RCT older-adult meta-analysis: positive but not significant. One 90-person Alzheimer's RCT found B. longum R0175 improved cognitive scores vs. placebo — small, early, needs replication.

Verdict: An intriguing therapeutic lead for cognitive impairment — small trials, high heterogeneity. No credible evidence for healthy adults.

Safety: Same probiotic caveats as above.

Fermented foods & prebiotics

Claimed: Kimchi, yogurt, and kefir calm inflammation and lift mood.

Evidence: The Stanford trial (Wastyk et al., Cell 2021, n=36, 10 weeks): a fermented-food diet increased microbiome diversity and decreased 19 inflammatory proteins (including IL-6), dose-dependently. Critical limit: the trial measured inflammation biomarkers, not mood or clinical outcomes — mood claims leap past the data. Prebiotic-specific mood RCTs are thinner and mixed.

Verdict: The best-evidenced gut intervention on this page — for inflammation markers. The mood link is a plausible mechanism with an unproven outcome. And it's food, not a pill: the cheapest, safest experiment on this entire page.

Safety: Food-safe for most; tyramine caution with MAOI antidepressants or histamine sensitivity.

Tools & games: what's claimed, what the evidence says

Three beloved categories, three deflations — and what each is actually good for:

Dual n-back training apps

Claimed: Training working memory raises fluid intelligence and IQ — the famous 2008 finding that launched the brain-training industry.

Evidence: The reference-standard meta-analysis (87 studies, 2016): working-memory training does not improve performance on measures of intelligence — gains are task-specific. A preregistered 2025 RCT found no transfer to fluid intelligence. A Canadian RCT's Bayesian analyses favored the null 4–40×. One possible exception: small trials suggest verbal working-memory gains in ADHD populations — remediation of a deficit, not enhancement.

Verdict: Dual n-back makes you good at dual n-back. For a healthy adult it does not raise IQ or general intelligence — one of the most thoroughly deflated claims in cognitive science. (We cover brain training in depth in Does Brain Training Work? and rate the actual apps in Best Brain Training Apps.)

Chess

Claimed: "Chess makes you smarter" — a common selling point of chess programs for kids.

Evidence: A 2016 meta-analysis found chess skill correlates with fluid reasoning and memory (~6% of variance each) — but correlation runs both ways: smarter people are drawn to and persist at chess. Full-scale IQ explained only ~1% of chess-skill variance. Chess instruction modestly improved math performance in some studies; IQ effects were inconsistent.

Verdict: Chess builds chess skill, pattern recognition, and disciplined thinking habits — real and valuable. It does not raise general IQ; the correlation mostly reflects who chooses to play.

If you want a board anyway: chess sets on Amazon.

Learning a musical instrument or a new language

Claimed: "Music makes kids smarter"; bilingualism gives a sharper brain; language apps rewire your mind.

Evidence: A 2020 multilevel meta-analysis (N=6,984, 254 effect sizes): once study quality was controlled, the effect of music training on cognition was null (g ≈ 0) — positive results came only from low-quality studies. A 2018 meta-analysis found no bilingual advantage in executive functions. An 18-month language-training RCT in healthy older adults found no cognitive benefit.

Verdict: Learn an instrument or a language for the skill itself — music, culture, travel, enjoyment, discipline. Do not learn them expecting a higher IQ; the science says the cognitive bonus is ≈ zero.

The honest bottom line

Rank-ordered by seriousness of evidence for any cognitive or mood benefit in healthy adults: caffeine + L-theanine (acute focus) > creatine (small memory/speed nudge) > bacopa (slow, mixed, plausible) > psychobiotics for mood (strain-specific, modest, adjunctive) > rhodiola (anti-fatigue under stress) > magnesium (fixes deficits) > omega-3 (small memory signal, not a booster) > beetroot (vessels, not brains) > iron (only if you're deficient — test first) > ginkgo (near-zero) > lion's mane (folklore plus petri dishes) > L-citrulline/L-arginine (near-zero human evidence) > gut bacteria for metabolism (no meaningful weight loss) > dual n-back, chess, and music for IQ (≈ zero transfer). AI belongs in its own category: it doesn't raise IQ, but used as a Socratic sparring partner it multiplies what your IQ accomplishes — used as an answer machine, the evidence says it hollows out the thinking it replaces. And none of it raises IQ — which is exactly what our honest guide to increasing IQ told you first.

Keep exploring: take the free IQ quiz, read can you actually increase your IQ?, does brain training work?, or browse the best brain training apps.

For entertainment and general information only — this guide is not medical, psychological, or professional advice. Talk to a qualified clinician before starting any supplement. Product links on this page are Amazon affiliate links: as an Amazon Associate, IQ Challenge earns from qualifying purchases, which helps keep the quizzes free.

© FUNGWA 峳華 LLC (fēng huá — “peaks of splendor”)