Place · Level 3
Lion's Mane · Hericium erinaceus
网红补脑菇: 机制有意思, 但人体证据很薄——离逆转老年痴呆还差得远
Last updated
Story path
- 1White coral mushroom · tradition + TikTokWhite coral mushroom · tradition + TikTok
- 2Mechanism · NGF/BDNF pathwayMechanism · NGF/BDNF pathway
- 3RCT evidence · narrower than marketingRCT evidence · narrower than marketing
- 4Product quality · grain vs real mushroomProduct quality · grain vs real mushroom
- 5Decision tree · should I useDecision tree · should I use
Chapter 1
White coral mushroom · tradition + TikTok
White coral mushroom · tradition + TikTok
Lion's Mane = Hericium erinaceus — a white / cream filamentous fungal cluster:
Scientific name Hericium erinaceus (Hericiaceae family)Aliases: lion's mane / houtou (China, also 'monkey-head mushroom') / Yamabushitake (Japan) / pom-pom mushroom (North America) / bear's head mushroomNative to: temperate forests of East Asia + North America + EuropeBiology:Saprotrophic + weakly parasitic — grows on dead or senescent hardwoods (beech / oak / birch)Fruiting body = white drooping needle-like hyphal cluster, looks like a lion's mane / coral / the pom-poms on Japanese yamabushi (mountain ascetic monk) robesEdible — steamed / sautéed, taste and texture resemble crab / lobsterTraditional medicinal use: Chinese medicine (monkey-head mushroom) — gastric tonic / calming / treats gastric ulcer; Japan (Yamabushitake) — ascetic practice + longevity
Chemistry: this is the core of its 'theoretical appeal':
Hericenones A–H — found in the fruiting body (mushroom), benzofuran derivativesErinacines A–K — found in the mycelium, cyathane diterpenesCommon feature of both compound classes: in vitro and animal models show promotion of nerve growth factor (NGF) and brain-derived neurotrophic factor (BDNF) expressionPolysaccharides (Hericium polysaccharides) — immunomodulatory (animal + in vitro)Antioxidant + anti-inflammatory components
The critical distinction between 'fruiting body vs mycelium' (the most common marketing deception):
Fruiting body (mushroom body) = the 'mushroom' you see in supermarkets, the fungus's 'fruit'Mainly contains hericenonesContains β-glucan (immunomodulatory) plus proteinMycelium = the mushroom's 'root system', usually grown on grain substrate (wheat / rice)Mainly contains erinacinesCommercially sold as grain + mycelium mix, containing 50–80% grain (starch)Market fraud: some products are just grain powder containing mycelium, with actual Hericium content < 10%Testing: ConsumerLab + third-party tests find some brands' β-glucan content extremely low — basically grain starch
Why the sudden post-2020 surge:
2020–2024 TikTok + Huberman + Joe Rogan + Tim Ferriss push — 'natural nootropic / Alzheimer prevention'Paul Stamets (mycologist, Fantastic Fungi documentary) promoted itPersonal mother case appeared in media — 'lion's mane reversed Alzheimer' (actually an N = 1 anecdote, no RCT)Microdosed psilocybin + lion's mane + niacin 'Stamets Stack' — completely without RCT validation2020 sales ~$50M, 2024 ~$300M+ — marketing-driven
The problem: 'real evidence' is far narrower than 'TikTok belief' — the next scenes unpack.
Scientific name Hericium erinaceus (Hericiaceae family)Aliases: lion's mane / houtou (China, also 'monkey-head mushroom') / Yamabushitake (Japan) / pom-pom mushroom (North America) / bear's head mushroomNative to: temperate forests of East Asia + North America + EuropeBiology:Saprotrophic + weakly parasitic — grows on dead or senescent hardwoods (beech / oak / birch)Fruiting body = white drooping needle-like hyphal cluster, looks like a lion's mane / coral / the pom-poms on Japanese yamabushi (mountain ascetic monk) robesEdible — steamed / sautéed, taste and texture resemble crab / lobsterTraditional medicinal use: Chinese medicine (monkey-head mushroom) — gastric tonic / calming / treats gastric ulcer; Japan (Yamabushitake) — ascetic practice + longevity
Chemistry: this is the core of its 'theoretical appeal':
Hericenones A–H — found in the fruiting body (mushroom), benzofuran derivativesErinacines A–K — found in the mycelium, cyathane diterpenesCommon feature of both compound classes: in vitro and animal models show promotion of nerve growth factor (NGF) and brain-derived neurotrophic factor (BDNF) expressionPolysaccharides (Hericium polysaccharides) — immunomodulatory (animal + in vitro)Antioxidant + anti-inflammatory components
The critical distinction between 'fruiting body vs mycelium' (the most common marketing deception):
Fruiting body (mushroom body) = the 'mushroom' you see in supermarkets, the fungus's 'fruit'Mainly contains hericenonesContains β-glucan (immunomodulatory) plus proteinMycelium = the mushroom's 'root system', usually grown on grain substrate (wheat / rice)Mainly contains erinacinesCommercially sold as grain + mycelium mix, containing 50–80% grain (starch)Market fraud: some products are just grain powder containing mycelium, with actual Hericium content < 10%Testing: ConsumerLab + third-party tests find some brands' β-glucan content extremely low — basically grain starch
Why the sudden post-2020 surge:
2020–2024 TikTok + Huberman + Joe Rogan + Tim Ferriss push — 'natural nootropic / Alzheimer prevention'Paul Stamets (mycologist, Fantastic Fungi documentary) promoted itPersonal mother case appeared in media — 'lion's mane reversed Alzheimer' (actually an N = 1 anecdote, no RCT)Microdosed psilocybin + lion's mane + niacin 'Stamets Stack' — completely without RCT validation2020 sales ~$50M, 2024 ~$300M+ — marketing-driven
The problem: 'real evidence' is far narrower than 'TikTok belief' — the next scenes unpack.
热度是怎么起来的 · 以及个人故事为什么不算数
至于为什么 2020 年之后突然爆红, 说白了是 TikTok、几位播客大 V 和一部蘑菇纪录片一起带的货, 再配上几个逆转老年痴呆的个人故事 (其实都是没有对照的单例)。这几年销量翻了好几倍——但热度远远跑在证据前面。为什么个人故事在这个题目上尤其靠不住
不是因为讲故事的人不诚实, 而是因为认知功能这个终点有三个性质, 刚好让单个案例失去判断力:
它天天在波动。 同一个人的记忆和专注度, 睡好睡坏、有没有感冒、当天心情, 都能拉开可观的差距。人几乎总是在状态低谷的时候才决定要做点什么, 而低谷之后大概率会自己回升 —— 这叫回归均值, 和你那天开始吃了什么无关它没有客观刻度。 血压有数字, 脑子清不清楚没有。而人对自己认知状态的评估, 恰恰是最容易被期待改变的那一类感受: 你花了钱、抱着希望、每天记得吃, 这三件事本身就会让自我评分变好改变很少单独发生。 一个决定认真对待自己脑子的人, 通常同时开始早睡、运动、少喝酒。这几件的证据等级都比蘑菇高得多, 但功劳往往被算在最新加进来的那一件上
所以这一岛后面的做法是: 不看故事, 只看有对照组、有客观量表、事先说好要测什么的试验。然后你会发现, 这样的试验非常少, 而且每一个都很小。
Chapter 2
Mechanism · NGF/BDNF pathway
Mechanism · NGF/BDNF pathway
Lion's mane neurotrophic pathway (Lai 2013 + Mori 2009, etc.):
Main path: NGF + BDNF expression upregulation:
NGF (nerve growth factor) = key protein maintaining neuron survival + synaptic plasticity + myelinationAging → NGF ↓ → neuronal declineAlzheimer / Parkinson / MS: NGF signaling pathway impairedBDNF (brain-derived neurotrophic factor) = similar function, highly expressed in hippocampus + prefrontal cortexExercise + adequate sleep + learning → BDNF ↑Depression / chronic stress: BDNF ↓
Hericenones and erinacines in cell experiments:
Increase NGF / BDNF mRNA expression in cultured cortical / hippocampal neuronsPromote neurite outgrowthReduce Aβ-induced cell apoptosis (Alzheimer model)
Animal models:
Mice / rats given lion's mane extract → hippocampal BDNF ↑ + learning and memory performance improved (Morris water maze)APP/PS1 Alzheimer mouse model → reduces Aβ plaques + improves cognitionAxonal regeneration model: after peripheral nerve injury, adding lion's mane → regeneration rate ↑
The 'mechanism story' here is beautiful — but the gaps from in vitro → animal → human are large:
Gap 1: blood–brain barrier: The 'security gate' on brain vessels that blocks most substances in blood from entering the brain. penetration:
Hericenones: small molecules, can cross the blood-brain barrier (BBB) — animal PK data supportsErinacines: similar, can crossPolysaccharides: large molecules, don't directly cross the BBB, work indirectly via immune / gut-brain axisCaveat: 'crosses the BBB' ≠ 'effective' — needs sufficient concentration + right cells + right timing
Gap 2: dose differences:
Animal effective dose: ~ 200–500 mg/kg — scaled to humans (70 kg, cross-species scaling factor ÷ 6–12): human needs 1.2–5.8 g/dayMost RCTs use 1–3 g/day — close but often too lowTikTok recommendation: 500 mg–1 g/day — may not reach effective dose at all
Gap 3: duration:
Animal models: weeks to months (already a large fraction of mouse lifespan)Human Alzheimer progression: 5–20 yearsRCT duration: 12–49 weeks — hard to judge long-term effects
Gap 4: extract form + standardization:
Animal studies often use purified hericenones / erinacinesHuman RCTs use whole-mushroom powder / standardized extract / grain + mycelium mixActive compound content varies enormously (10×+)
Real mechanism positioning:
Has signal: in vitro + animal data suggest a real pathway existsTranslation to humans: still early — most human RCTs are small C-tier evidenceNot 'completely ineffective', but also not 'miracle Alzheimer cure'
Recent human PK data (limited):
After single oral dose, erinacine A is briefly detectable in plasmaBrain tissue concentration: no direct human data — extrapolated from animal modelsBioavailability: estimated ~5–15% (similar to many plant terpenes)
Conclusion: lion's mane is 'a plausible-mechanism early clinical candidate', not 'a proven brain-health drug'.
Main path: NGF + BDNF expression upregulation:
NGF (nerve growth factor) = key protein maintaining neuron survival + synaptic plasticity + myelinationAging → NGF ↓ → neuronal declineAlzheimer / Parkinson / MS: NGF signaling pathway impairedBDNF (brain-derived neurotrophic factor) = similar function, highly expressed in hippocampus + prefrontal cortexExercise + adequate sleep + learning → BDNF ↑Depression / chronic stress: BDNF ↓
Hericenones and erinacines in cell experiments:
Increase NGF / BDNF mRNA expression in cultured cortical / hippocampal neuronsPromote neurite outgrowthReduce Aβ-induced cell apoptosis (Alzheimer model)
Animal models:
Mice / rats given lion's mane extract → hippocampal BDNF ↑ + learning and memory performance improved (Morris water maze)APP/PS1 Alzheimer mouse model → reduces Aβ plaques + improves cognitionAxonal regeneration model: after peripheral nerve injury, adding lion's mane → regeneration rate ↑
The 'mechanism story' here is beautiful — but the gaps from in vitro → animal → human are large:
Gap 1: blood–brain barrier: The 'security gate' on brain vessels that blocks most substances in blood from entering the brain. penetration:
Hericenones: small molecules, can cross the blood-brain barrier (BBB) — animal PK data supportsErinacines: similar, can crossPolysaccharides: large molecules, don't directly cross the BBB, work indirectly via immune / gut-brain axisCaveat: 'crosses the BBB' ≠ 'effective' — needs sufficient concentration + right cells + right timing
Gap 2: dose differences:
Animal effective dose: ~ 200–500 mg/kg — scaled to humans (70 kg, cross-species scaling factor ÷ 6–12): human needs 1.2–5.8 g/dayMost RCTs use 1–3 g/day — close but often too lowTikTok recommendation: 500 mg–1 g/day — may not reach effective dose at all
Gap 3: duration:
Animal models: weeks to months (already a large fraction of mouse lifespan)Human Alzheimer progression: 5–20 yearsRCT duration: 12–49 weeks — hard to judge long-term effects
Gap 4: extract form + standardization:
Animal studies often use purified hericenones / erinacinesHuman RCTs use whole-mushroom powder / standardized extract / grain + mycelium mixActive compound content varies enormously (10×+)
Real mechanism positioning:
Has signal: in vitro + animal data suggest a real pathway existsTranslation to humans: still early — most human RCTs are small C-tier evidenceNot 'completely ineffective', but also not 'miracle Alzheimer cure'
Recent human PK data (limited):
After single oral dose, erinacine A is briefly detectable in plasmaBrain tissue concentration: no direct human data — extrapolated from animal modelsBioavailability: estimated ~5–15% (similar to many plant terpenes)
Conclusion: lion's mane is 'a plausible-mechanism early clinical candidate', not 'a proven brain-health drug'.
NGF 推高之后 · 神经元里在建什么
NGF 上去了这句话本身不解释任何事。要看懂它, 得跟着信号走完下面这一串。第一步 · NGF 得先被接住。 NGF 自己不进细胞, 它停靠在神经元表面一种叫 TrkA 的受体上。两个受体被同一批 NGF 拉到一起、彼此激活, 信号就从细胞膜上启动了。受体不够或者已经坏掉的神经元, 你把 NGF 堆得再高也没用 —— 这是第一个天花板, 也是多给一点就多长一点这种直觉错在哪里。
第二步 · 信号被送回细胞核。 接住的信号不是就地起效, 而是被装进小囊泡, 沿着轴突内部的轨道一路逆行运回细胞体。细胞核收到之后改变基因表达, 开始生产建材: 细胞骨架的零件、膜、突触上要用的蛋白。注意这一步的时间尺度 —— 运输、转录、翻译, 这是以小时到天计的流程。
第三步 · 生长锥往外推。 轴突的尖端有一个像手掌一样的结构, 叫生长锥。它伸出许多细丝在周围探路, 探到合适的方向就把细胞骨架往那边铺一段, 整根轴突于是长长一点。NGF 在这里同时干两件事: 提供建造许可 (让细胞有材料可用), 和提供方向 (哪边浓度高就往哪边长)。
第四步 · 碰到目标, 搭突触。 生长锥碰到合适的目标细胞就停下来, 双方各自把该有的蛋白摆到接触面上, 一个新的突触慢慢成形。到这一步, 两个神经元之间才真正多了一条新的连线 —— 而连线才是记忆和技能的物理形式。
第五步 · 用它, 否则失去它。 新连接不是建好就永久了。它必须被反复激活才会被加固, 否则会被修剪掉。神经系统在这件事上非常吝啬: 维持一条突触要持续花能量, 不用的就拆。
把这五步连起来看, 三个结论自己就掉出来了:
这是建造, 不是兴奋。 咖啡因那类东西是当场改变神经元的放电状态, 分钟级见效; 而上面这五步要转录、要运输、要搭结构, 时间尺度是周。所以吃了当天觉得更清醒几乎不可能来自这条通路 —— 那时工程还没动土停药之后为什么会退回去。 那项 MCI 试验里, 受试者停用之后效果在几周内消退。放进第五步就完全说得通: 外来的推力撤掉, 新建的连接不再被优先维持, 于是按正常的修剪规则退回基线。这恰恰是这条机制成立时应有的样子 —— 一个真正长出结构的干预, 如果结构的维持依赖持续输入, 撤掉就会回退。它不是假效应的证据, 但它确实说明这不是一次性修好为什么在已经有点问题的人身上信号更明显。 健康年轻人的这条通路本来就在正常运转, 瓶颈不在 NGF 的量上; 而衰退中的系统缺的恰恰是这一环。一个补短板的干预, 在没有短板的人身上看不到效果, 这不是矛盾, 是意料之中 —— 后面 RCT 那一幕里健康人群信号最弱, 根子在这里
细胞与动物数据 · 四道翻译坎
细胞实验里它确实做到了: 培养的皮层、海马神经元里 NGF、BDNF 的表达上去了, 神经突起长得更旺, 连阿尔茨海默模型里那种 Aβ 毒性引起的细胞凋亡也少了。动物身上也漂亮——小鼠、大鼠喂了提取物, 海马 BDNF 升高, 走迷宫的记忆表现变好; 在 APP/PS1 这种阿尔茨海默小鼠里, Aβ 斑块减少、认知改善; 周围神经被切断后补一补, 再生也更快。问题是, 从试管到小鼠再到人, 中间隔着好几道坎:
能不能进脑子: hericenones 和 erinacines 是小分子, 动物数据说它们能穿过血脑屏障; 但多糖那类大分子进不去, 只能靠免疫和肠脑轴绕着起作用。而且穿得过去不等于管用, 还得在对的细胞、对的时机攒够浓度。剂量对不对得上: 动物的有效剂量大约 200-500 mg/kg, 按体重和跨物种系数折算到人, 差不多要 1.2-5.8 g/天; 而多数人体试验只用 1-3 g/天, 已经偏低, TikTok 上常推的 500 mg-1 g/天, 很可能压根没到有效门槛。时间够不够长: 动物模型跑几周到几个月, 对老鼠已是一大段寿命; 可人的阿尔茨海默要走 5-20 年, 而试验通常只有 12-49 周, 长期到底怎样根本看不出来。吃的是不是同一个东西: 动物研究常用提纯的 hericenones、erinacines, 人体试验用的却是全菇粉、标准化提取物或谷物菌丝混合物, 活性成分含量能差十倍以上。
所以真实的定位是这样: 机制这条线是有信号的, 体外和动物都提示通路真实存在; 但翻译到人身上还很早, 多数人体试验是小样本的低级别证据。它不是完全没用的安慰剂, 也远不是逆转阿尔茨海默的神药——更准确的说法是, 一个机制说得通、可临床还没验透的早期候选。
第一道坎细看: NGF 自己过不了血脑屏障
上上页那条建造链, 是在神经元附近已经有 NGF的前提下走的。现在问一个更靠前的问题: 那些 NGF 是从哪来的?这里有一个几乎所有营销文案都跳过的事实: NGF 是一个大蛋白, 它过不了血脑屏障。
血脑屏障是脑血管内皮细胞之间那一圈焊死的接缝。身体别处的毛细血管壁, 细胞与细胞之间留有缝隙, 分子可以从缝里挤过去; 脑血管这里被紧密连接封住了, 只有小的、脂溶性的分子能直接穿膜, 大分子要么走专门的转运蛋白, 要么就进不去。NGF 这种尺寸的蛋白, 没有给它开的门。
所以补 NGF 这条路从一开始就是不通的 —— 你吞下去的任何 NGF 都到不了脑子, 更何况蛋白在胃里先就被消化成碎片了。
猴头菇声称的路径因此不是送 NGF 进去, 而是让脑自己多做。 完整的说法长这样:
猴头菇里的 hericenones 和 erinacines 是小分子, 不是蛋白小分子有可能穿过血脑屏障 —— 动物药代数据支持这一点进到脑组织之后, 它们作用于脑内的细胞, 让这些细胞自己合成并释放更多 NGF脑内新增的 NGF 再走建造链那五步
这条链的每一环都各自有证据, 但强度差得很远。 最薄的是中间那一环: 小分子在人的脑组织里到底到了什么浓度, 没有直接的人体数据。 人身上能测到的只是口服之后血浆里短暂出现的量; 从血浆到脑组织的那一跳, 目前是从动物模型外推的。
为什么这一环特别重要, 而不只是一句还没测。 因为它是整个故事的枢纽。前面那些漂亮的体外实验, 做法都是把分子直接加到培养的神经元上 —— 那等于默认分子已经到了。如果实际到达脑组织的量远低于培养皿里的浓度, 那么细胞数据一条都不假, 只是够不着。这不是造假问题, 是剂量到达问题, 也正是同一幕里那道剂量折算坎在说的同一件事。
说清楚这一点, 不等于说它没用。 诚实的表述是: 这是一条每一环都合理、但没有一环是在人脑里直接测到的链条。它比连机制都讲不出来的补剂强, 也比已经在人身上验证过的干预弱很多。
而知道断点在哪, 你就知道什么样的新研究才真正能推动结论 —— 不是再做一个小样本的量表试验, 而是有人真的把到达人脑的量测出来。下次看到猴头菇的新研究, 先看它测的是不是这一环。
Chapter 3
RCT evidence · narrower than marketing
RCT evidence · narrower than marketing
Complete lion's mane human RCT inventory (as of 2024):
C-tier (RCTs exist but small samples / preliminary results):
① Mori 2009 (Phytotherapy Research) — MCI trial (the most-cited):
N = 30 mild cognitive impairment Japanese adults, age 50–80Intervention: Yamabushitake fruiting body powder, four 250 mg tablets three times a day = 3 g/day × 16 weeksResults:Treatment group's HDS-R (revised Hasegawa Dementia Scale) score rose significantlyEffect dissipated 4 weeks after stopping — suggests continuous use neededCaveats:N = 30 very smallMCI ≠ AlzheimerShort durationNo large-sample replication (15 years later, still no N > 200 replication study)
② Li 2020 (Frontiers in Aging Neuroscience) — early AD pilot:
N = 49 mild Alzheimer adults (Taiwan)Intervention: erinacine A-enriched mycelium 350 mg × 3/day × 49 weeksResults: MMSE / CASI / IADL improved, neurofilament light protein decreasedCaveats:Pilot study, not phase 3Single-center / single-ethnicity / no replicationCompany-funded (Hericium Erinaceus mycelium producer)
C-tier (RCTs exist but small samples / preliminary results):
① Mori 2009 (Phytotherapy Research) — MCI trial (the most-cited):
N = 30 mild cognitive impairment Japanese adults, age 50–80Intervention: Yamabushitake fruiting body powder, four 250 mg tablets three times a day = 3 g/day × 16 weeksResults:Treatment group's HDS-R (revised Hasegawa Dementia Scale) score rose significantlyEffect dissipated 4 weeks after stopping — suggests continuous use neededCaveats:N = 30 very smallMCI ≠ AlzheimerShort durationNo large-sample replication (15 years later, still no N > 200 replication study)
② Li 2020 (Frontiers in Aging Neuroscience) — early AD pilot:
N = 49 mild Alzheimer adults (Taiwan)Intervention: erinacine A-enriched mycelium 350 mg × 3/day × 49 weeksResults: MMSE / CASI / IADL improved, neurofilament light protein decreasedCaveats:Pilot study, not phase 3Single-center / single-ethnicity / no replicationCompany-funded (Hericium Erinaceus mycelium producer)
早期阿尔茨海默那项试点 · 以及试点是干什么用的
另一篇常被当成逆转阿尔茨海默证据的, 是 Li 2020。49 位台湾的轻度阿尔茨海默患者, 每天吃一次 350 mg、一天三次的 erinacine A 强化菌丝体, 连吃 49 周, 结果 MMSE、CASI、IADL 几项认知量表都改善了, 一个反映神经损伤的血液标记 (神经丝轻链蛋白) 也降了。数据方向确实好看, 但它只是个探路性质的小试点, 单中心、单一族群、没人复制过, 而且钱是生产这种菌丝体的公司出的——这些都得记在账上。为什么公司出钱这一条要写在明面上
资助方不等于造假。但它会通过一些完全合规的途径影响结果的样子: 试验设计选哪个终点、多久测一次、以及阴性结果发不发表。所以行业惯例是披露, 而读者的正确反应也不是因此不信, 而是因此需要一个独立团队复制它。这项试点做完到现在, 复制还没有出现。
还有一个更基础的问题: 试点研究是用来干什么的?
试点 (pilot) 的目的通常不是证明有效, 而是回答这个试验做得下去吗 —— 剂量能不能耐受、受试者招不招得到、终点测不测得动、脱落率有多高。它的样本量是按这些问题算出来的, 不是按检出疗效算出来的。
这带来一个统计上的后果, 值得记住: 样本小的时候, 真效应和随机波动长得一模一样。所以一个小试点得出阳性结果, 既可能是真的, 也可能只是这一批人恰好如此 —— 而单靠这一项研究, 你没有办法区分。
它真正有资格支持的说法只有一句: 值得做一个更大的试验。
把这一条推广开, 你会得到一个很实用的筛子: 看到某个补剂有临床试验支持, 先问三句 —— 多少人? 有没有安慰剂对照? 有没有别的团队做出同样结果? 第三句最重要, 也是这一味蘑菇目前最答不上来的一句。
Other C-tier RCTs · mood + menopause + healthy memory
③ Vigna 2019 (Evid-Based Compl Alt Med) — depression-anxiety:N = 77 overweight / obese with anxiety + depressive symptomsIntervention: lion's mane 80% fruiting body + 20% mycelium, 500 mg × 2/day × 8 weeksResults: depression + anxiety scores ↓, pro-BDNF / BDNF ratio improvedCaveats:Composite-condition population (obesity + depression)Short durationSingle-center
④ Nagano 2010 — menopausal women's mood:
N = 30 randomised (26 analysed), mean age about 40, classified healthy by the paper — not a menopausal cohortLion's mane cookies, 4/day (2 g powder) × 4 weeksResults: CES-D and ICI fell within the treatment group, but only two ICI sub-items beat placebo. (The CMI score this page used to name is not an instrument the trial used)Caveats:Same small sample / short duration'Lion's mane cookies' — lion's mane content unclear
⑤ Saitsu 2019 — healthy middle-elderly memory:
N = 31 healthy adults age 50–80Lion's mane extract 3.2 g/day × 12 weeksResults: of the three tests, only the MMSE reached significance; the Benton visual retention test and the S-PA verbal paired-associate test did notCaveat: weak effect in a healthy population — and the test that moved is the bluntest one (MMSE is a screening scale, not a memory test)
C-D tier (in vitro + animal + case reports):
MS (multiple sclerosis) — case reports + animalPeripheral neuropathy — animalGastric ulcer — traditional use, some animal RCTsAntitumor — in vitro + animal
Marketing gaps + meta + cross-comparison
Popular promotions without RCTs:'Prevents Alzheimer' — no large-sample / long-term RCT data'Reverses Alzheimer' — Li 2020 is a pilot, doesn't constitute 'reversal' evidence; anecdotes don't count'Focus boost / faster learning' — no RCT validation of acute effects in healthy people'Stamets Stack enhances consciousness' — completely without RCT
Meta-analysis status:
As of 2024: no high-quality meta-analysis pools lion's mane RCTs — because studies are too heterogeneous and samples too smallThis itself is a marker of 'low evidence tier'
Cross-comparison:
Lion's mane human evidence: C-tierOmega-3 (DHA) for cognitive decline prevention: B-tier (PREDIMED, ACCORD-MIND)Mediterranean diet + exercise: A-tier (FINGER trial)Adequate sleep (7–9 h): B-A tier (multiple observational + RCTs)
Lion's mane RCT data is clearly weaker than the above
'Few RCTs = I should wait for more evidence' vs 'no harm, try it':
For 'wait for more evidence': money is better spent on A-tier things (exercise / sleep / Mediterranean diet)For 'try it': safety profile is good + reversible + subjective experience sometimes improvesBalanced view: if you want to try, use third-party-certified real fruiting body extract + 1–3 g/day + 8–12 weeks of objective measurement + no 'reverse Alzheimer' expectation
Chapter 4
Product quality · grain vs real mushroom
Product quality · grain vs real mushroom
The biggest pitfall in the lion's mane market: what you buy may not be lion's mane:
Pitfall 1: grain mycelium vs real fruiting body:
Grain-grown mycelium (GFM):Cultured on a rice / wheat bran substrateThe grain isn't separated → dried, powdered, and sold directlyReal Hericium content ~10–30%; the remaining 50–80% is grain starchβ-glucan content often tests < 10% (low)Erinacines (the target of animal PK studies) are present in low amountsReal fruiting body (FB):Mycelium grows out actual 'mushrooms' — harvested and driedβ-glucan content 30–50% (high)Hericenones content adequate3–5× higher price (longer growth time + equipment investment)
ConsumerLab + 2019 Hobbs survey:
30+ US market lion's mane products tested70% labeled 'mushroom' were actually grain mycelium powder~60% of products had real fruiting body content < 30%β-glucan / maltose ratio (M/G ratio): real fruiting body < 0.5, grain mycelium > 1.0 — usable to identify fraud
Pitfall 1: grain mycelium vs real fruiting body:
Grain-grown mycelium (GFM):Cultured on a rice / wheat bran substrateThe grain isn't separated → dried, powdered, and sold directlyReal Hericium content ~10–30%; the remaining 50–80% is grain starchβ-glucan content often tests < 10% (low)Erinacines (the target of animal PK studies) are present in low amountsReal fruiting body (FB):Mycelium grows out actual 'mushrooms' — harvested and driedβ-glucan content 30–50% (high)Hericenones content adequate3–5× higher price (longer growth time + equipment investment)
ConsumerLab + 2019 Hobbs survey:
30+ US market lion's mane products tested70% labeled 'mushroom' were actually grain mycelium powder~60% of products had real fruiting body content < 30%β-glucan / maltose ratio (M/G ratio): real fruiting body < 0.5, grain mycelium > 1.0 — usable to identify fraud
市场实测 · 一个比值怎么验出掺没掺谷物
这不是个别现象。ConsumerLab 联合 2019 年 Hobbs 那次调查测了美国市场 30 多款狮鬃菇产品, 结果 70% 标着蘑菇的其实是谷物菌丝粉, 约 60% 产品的真子实体含量不到 30%。好在造假能查出来: 真子实体的麦芽糖和 β-葡聚糖比值 (M/G) 低于 0.5, 而掺谷物的菌丝粉会高过 1.0——一测便知。为什么一个比值就能验出真假
这个检测的逻辑很干净, 值得单独讲一下 —— 它是看穿成分标签的一个通用范式, 换个品类照样能用。
蘑菇细胞壁的主要结构材料是 β-葡聚糖; 谷物 (大米、麦麸) 的主要储能物质是淀粉, 而淀粉水解之后放出的是麦芽糖。两种原料各自带着一个别人没有的标记物。于是:
一份货里 β-葡聚糖占比高、麦芽糖低 → 它主要是蘑菇反过来麦芽糖高、β-葡聚糖低 → 里面主要是谷物
关键在于用比值, 而不是用绝对量。 绝对量可以被稀释、被浓缩、被换算单位糊弄; 而两个成分的比值反映的是原料本身的身份, 掺什么进去都躲不掉 —— 你往里加谷物, 分子分母同时朝一个方向动, 比值立刻暴露。
这也是为什么标签上写多少毫克几乎没有信息量。 毫克数说的是粉末的重量, 而粉末到底是什么, 正是这里要验的那个问题。一勺谷物粉和一勺子实体提取物, 在秤上完全一样。
所以你要找的不是剂量数字, 是成分身份的证明: 明确写子实体、给出 β-葡聚糖的标定值、附得上第三方报告。这三样齐了, 毫克数才开始有意义 —— 在那之前, 它只是在告诉你袋子有多重。
More pitfalls · dual-extraction + polysaccharide + fake erinacine A
Pitfall 2: 'dual extraction' obfuscation:Claim: 'water + ethanol dual extraction extracts all actives'Reality:Water extracts β-glucan (high molecular weight polysaccharides)Ethanol extracts small molecules (hericenones / erinacines / triterpenes)'Dual extraction' sounds like getting both classes at once, but in practice the ratio / standardization of the two solvents is hard to controlNot a scam, but 'perfectly preserves all actives' is marketing exaggeration
Pitfall 3: marketing confuses 'β-glucan' and 'polysaccharide':
β-glucan = mushroom cell wall polysaccharide, the immunomodulatory active compound'Polysaccharide' includes grain starch — the latter has no pharmacological activityA label of '50% polysaccharide' may mean 50% is starch, with mushroom β-glucan content unknownLook at β-glucan standardization, not 'polysaccharide'
Pitfall 4: fake 'erinacine A content':
Erinacines only exist in the mycelium, not in the fruiting body'Fruiting body + erinacine A marker' = contradiction, mostly fakeReal mycelium + standardized erinacine A may be legitimate (e.g. Li 2020 used erinacine A-enriched mycelium)
Quality identification checklist:
Label clearly says 'Fruiting Body / 子实体'Third-party test report (β-glucan ≥ 25%, low maltose content)USP / NSF / Eurofins / NPN (Canada) certificationCultivation method is explicit (organic / country of origin / substrate)Reasonable price: real fruiting body 60 g powder ~$25–50; below that likely grain mycelium
Reputable brands (2024 US market):
Real Mushrooms (Skye Chilton's company) — fruiting body only, third-party verifiedNammex (B2B raw material supplier, also sells consumer products) — industry benchmarkHericium Erinaceus (HEM) mycelium used in Li 2020 from China — but hard to obtain in consumer marketsAvoid: 'MycoStack composite mushroom' / 'Lion's Mane Coffee' / 'lion's mane chocolate' and other processed products — actual lion's mane content is tiny
Safety + interactions + processed-product reality
Safety profile (the relatively clean side of lion's mane):Acute side effects rare — GI discomfort / rash (~1–3%)No notable hepatorenal toxicity (unlike red yeast rice / tongkat ali / kava)No major drug interactionsMushroom-allergic individuals: use with caution — cross-reactivity possiblePregnancy / lactation: insufficient data, conservative avoidance recommended
Theoretical concerns:
Autoimmune disease: polysaccharide immune activation could theoretically worsen — data mixedAnticoagulation: weak in vitro antiplatelet signal — stop 2 weeks before surgery, caution on warfarin / DOACLong-term use (> 12 months): no data
'Lion's Mane Coffee' / 'Lion's Mane Chocolate' / 'Stamets Stack':
Most are marketing packagingLion's mane content < 100 mg/serving — far below the 1–3 g/day used in RCTs'A cup of coffee with brain boost on the side' — psychological satisfaction, no pharmacological meaningIf you really want to try, buy a pure fruiting body extract, not 'coffee with lion's mane in it'
Chapter 5
Decision tree · should I use
Decision tree · should I use
Lion's mane practical decision:
Scenarios worth considering (based on C-tier evidence + safety):
① Age 50–80 + MCI / subjective cognitive decline + lifestyle already addressed
Sleep ≥ 7 h / exercise / Mediterranean diet / social engagement / hearing screening already doneSubjective sense of cognitive fog → 8–12 week trialObjective testing (MoCA / simple memory score) at baseline + 12 weeksIf improvement ≥ 20% → continue; no change → stop
② Anxiety / depression adjunct (mild-moderate)
Already under professional careWant a 'natural adjunct' — weaker signal than ashwagandha / rhodiola, but possibly better safety profileDoesn't replace antidepressants / therapy
③ Early Alzheimer (Li 2020 indication)
Must be under neurologist guidanceDoesn't replace donepezil / memantineAs adjunct — some clinicians are willing to try
④ Nerve injury recovery (extrapolated from animal models)
Peripheral nerve injury / post-chemo neuropathyNo human RCT validation, but safe → can try
Scenarios worth considering (based on C-tier evidence + safety):
① Age 50–80 + MCI / subjective cognitive decline + lifestyle already addressed
Sleep ≥ 7 h / exercise / Mediterranean diet / social engagement / hearing screening already doneSubjective sense of cognitive fog → 8–12 week trialObjective testing (MoCA / simple memory score) at baseline + 12 weeksIf improvement ≥ 20% → continue; no change → stop
② Anxiety / depression adjunct (mild-moderate)
Already under professional careWant a 'natural adjunct' — weaker signal than ashwagandha / rhodiola, but possibly better safety profileDoesn't replace antidepressants / therapy
③ Early Alzheimer (Li 2020 indication)
Must be under neurologist guidanceDoesn't replace donepezil / memantineAs adjunct — some clinicians are willing to try
④ Nerve injury recovery (extrapolated from animal models)
Peripheral nerve injury / post-chemo neuropathyNo human RCT validation, but safe → can try
另两种可考虑的情形 · 恢复到底指什么
早期阿尔茨海默, 也就是 Li 2020 那个适应症。这一条必须在神经科医生指导下进行, 绝不能拿它替代多奈哌齐、美金刚这些正规药, 顶多当个辅助——有些医生也愿意一试。神经损伤的恢复, 比如周围神经损伤或化疗后的神经病变。这是从动物模型外推的, 人体还没有试验撑腰, 但因为它够安全, 想试也无妨。最后这一条里的恢复, 和补脑不是一回事, 值得分开说
周围神经和脑内神经元的处境差别很大。周围神经被切断或损伤之后, 本来就有一套天然的再生程序: 断端远侧的部分被清理掉, 髓鞘细胞排成一条通道, 轴突从断端重新长出来, 沿着这条通道往回长, 一直长到原来的目标。这条路是通的, 只是慢 —— 轴突生长的速度以毫米每天计, 所以一段较长的神经要按月来算。
NGF 在这个过程里的角色, 恰好就是机制那一幕讲过的那一套: 给生长锥提供建造材料和方向信号。所以在周围神经这个场景里, 神经营养因子的机制假说是最顺的 —— 它推的正是一条本来就在跑的程序, 相当于给一台已经启动的机器多送点料。
而脑内的情况相反。 中枢神经系统的成熟神经元再生能力弱得多, 而且周围环境里还存在抑制轴突生长的信号 —— 那不是设计缺陷, 是成年大脑为了保住已经学会的东西而刻意维持的稳定性。同样一个推高 NGF的干预, 在这两个地方遇到的阻力完全不同。
把这一层想清楚, 你就明白为什么这一条虽然连人体试验都还没有, 机制上却比健康人补脑那一条更站得住脚 —— 也正是它值得被单独列出来、而不是和补脑混成一句的原因。
注意: 机制更顺不等于有效。这一条目前的证据仍然只是从动物模型外推, 真要用在临床上 (尤其化疗后的神经病变), 必须先和你的主治医生讲 —— 化疗方案本身就复杂, 自己加东西不是小事。
Not worth / not recommended
Scenarios not worth or not recommended:① Healthy younger adults pursuing 'brain optimization'
RCT signal is weak in healthy populationsThe same budget put toward exercise, sleep, reading, learning a language, or learning an instrument returns far more than a supplement
② Treating it as 'lifelong Alzheimer prevention'
No large long-term RCT dataLifelong waste likelyPrioritize the A-tier evidence things (exercise + Mediterranean diet + social + sleep + hearing + BP/glucose control)
③ Moderate-to-severe late Alzheimer
Neurodegeneration is severe and the NGF pathway repair window is smallShould use acetylcholinesterase inhibitors + memantine + comprehensive careUsing lion's mane = misplaced hope
④ 'Stamets Stack' (lion's mane + microdosed psilocybin + niacin)
No RCT validation at allPsilocybin is illegal in most countriesFollowing KOLs isn't medicine
⑤ Mushroom allergy / pregnancy / lactation
Avoid
Quality choice:
Must be labeled 'Fruiting Body / 子实体'Standardized to β-glucan ≥ 25%Third-party test report publicly availableTypical dose: 1–3 g/day fruiting body powder or equivalent extractTake 1–3 months and watch for change — stop if nothing
'Is it working?' test + cross-comparison + bottom line
The truth behind 'it really makes me feel sharper':Possibility 1: real effect (neurotrophic signaling + anti-inflammatory)Possibility 2: placebo — trust + expectation + 'I'm doing something' feelingPossibility 3: confounders — when you started lion's mane, you also improved sleep / reduced stress / started exercising — the latter is what's workingHow to distinguish: Use objective cognitive tests as baseline and follow-up metric (not 'feeling')Self-deception check: stop for 4 weeks, see if you return to baseline
Cross-comparison (for ranking your 'brain health investment'):
| Intervention | Evidence tier | Monthly cost | Time |
|---|
When budget + time are limited, lion's mane sits at C-tier; exercise + sleep are A-tier
Wrapping up:
> Lion's mane is in the category of 'plausible mechanism, C-tier evidence, good safety profile' early clinical candidates — not 'reverse Alzheimer', not 'add 20 IQ points'. The marketing narrative from TikTok / Huberman / Joe Rogan and similar voices is typically several orders of magnitude ahead of the real evidence.
>
> It can serve as an adjunct option, but shouldn't be the main path. If you're truly worried about cognitive health, exercise + sleep + Mediterranean diet + social engagement + hearing screening + BP and glucose control combined go far beyond any supplement.
>
> The biggest problem with 'natural nootropic' marketing is that it gets people to skip A-tier evidence things to buy C-tier evidence things — a classic opportunity-cost misalignment.
References · 6
- Lai, P.-L., Naidu, M., Sabaratnam, V., Wong, K.-H., David, R. P., Kuppusamy, U. R., et al. (2013). Neurotrophic properties of the Lion's mane medicinal mushroom, Hericium erinaceus from Malaysia. International Journal of Medicinal Mushrooms, 15(6), 539-554. Erinacines and hericenones cross the BBB in animals and stimulate NGF/BDNF expression in cultured neurons. 10.1615/IntJMedMushr.v15.i6.30
- Mori, K., Inatomi, S., Ouchi, K., Azumi, Y., & Tuchida, T. (2009). Improving effects of the mushroom Yamabushitake (Hericium erinaceus) on mild cognitive impairment: a double-blind placebo-controlled clinical trial. Phytotherapy Research, 23(3), 367-372. 10.1002/ptr.2634
- Li, I.-C., Chang, H.-H., Lin, C.-H., Chen, W.-P., Lu, T.-H., Lee, L.-Y., et al. (2020). Prevention of early Alzheimer's disease by erinacine A-enriched Hericium erinaceus mycelia pilot double-blind placebo-controlled study. Frontiers in Aging Neuroscience, 12, 155. N=49 mild AD adults × 49 wk → MMSE/CASI/IADL improved vs placebo. 10.3389/fnagi.2020.00155
- Vigna, L., Morelli, F., Agnelli, G. M., Napolitano, F., Ratto, D., Occhinegro, A., et al. (2019). Hericium erinaceus improves mood and sleep disorders in patients affected by overweight or obesity: could circulating pro-BDNF and BDNF be potential biomarkers? Evidence-Based Complementary and Alternative Medicine, 2019, 7861297. N=77, 8 wk → depression + anxiety scores ↓ vs placebo. 10.1155/2019/7861297
- Saitsu, Y., Nishide, A., Kikushima, K., Shimizu, K., Ohnuki, K. (2019). Improvement of cognitive functions by oral intake of <i>Hericium</i><i> erinaceus </i>. Biomedical Research, 40(4), 125-131. Three tests: MMSE, Benton visual retention, S-PA verbal paired-associate. ONLY the MMSE reached significance — the two memory tests did not. The story used to print the reverse. 10.2220/biomedres.40.125
- Nagano, M., Shimizu, K., Kondo, R., Hayashi, C., Sato, D., Kitagawa, K., et al. (2010). Reduction of depression and anxiety by 4 weeks Hericium erinaceus intake. Biomedical Research, 31(4), 231-237. 30 women randomised (26 analysed), mean age ~40 and classified healthy — not a menopausal cohort. Instruments were KMI, CES-D, PSQI and ICI; the story credited a CMI score, which was not used. 10.2220/biomedres.31.231