视频链接: https://www.youtube.com/watch?v=ddq8JIMhz7c
本视频由斯坦福大学神经生物学教授 Andrew Huberman 主讲,深入探讨了基于科学证据的高效学习与记忆策略。其核心主题在于:有效的学习并非直觉上的不断重复输入,而是一个旨在“抵消自然遗忘”的过程。Huberman 详细解释了神经可塑性的生物学机制,指出通过主动的“自我测试”(而非反复阅读)能最有效地强化神经连接并减少高达 50% 的遗忘。视频还强调了专注力、高质量睡眠(特别是 REM 睡眠)以及利用肾上腺素(如情感唤醒或冷水刺激)在记忆巩固中的关键作用。通过建立规律的学习计划、消除干扰、并在接触新知识后立即进行开放式测试,任何学习者都能从单纯的知识熟悉提升到深度的掌握与精通。
神经可塑性(Neuroplasticity)是神经系统根据经验发生改变的能力,是所有学习行为的生物学基础。其底层机制主要由三种细胞层面的变化构成:
"Neuroplasticity is this incredible feature of your nervous system... the ability for your nervous system to change in response to experience." "The major mechanism by which learning and memory occurs in humans is the strengthening of existing connections and the weakening of existing connections." "The improvement in motor coordination... is largely the reflection of the removal of neural connections as opposed to the formation of neural connections."
学习不应被视为“获取信息”,而应被视为“抵消遗忘”。大脑天然会丢弃绝大多数输入信息,高效学习的目标是进行“遗忘接种”。 核心工具:自测(Testing)。
"The best way to study and learn is to access components of your memory systems that offset forgetting." "Testing is not just a way of evaluating what knowledge you've acquired... It also turns out to be the best tool for offsetting forgetting of any kind." "When you self-test material, you halve the amount of forgetting that would normally occur compared to if you're just exposed to the material."
神经可塑性是一个两步走的循环过程,缺一不可:
"Focus and alertness... cues your nervous system that something is important... signals to the neurons: hey, you're going to have to change." "The actual reordering of the connections... occurs during sleep." "Sleep is without question the best nootropic... you're going to need to get your sleep right in order to be able to study and learn at your absolute best."
神经塑性(即学习的本质)是一个两步过程。第一步是清醒状态下的高度专注,这会释放神经化学信号,标记哪些突触连接需要改变。第二步是真正的重构,这仅在深度睡眠和睡眠样状态下发生。神经元之间连接的增强、削弱或新连接的形成,主要发生在快速眼动(REM)睡眠阶段,且这一阶段在睡眠周期的后半段更为集中。此外,“第一晚效应”至关重要:学习新信息后的第一个晚上,是记忆巩固的核心窗口。如果当晚睡眠不足,即使后续补觉,学习效果也会大打折扣。因此,睡眠不仅是休息,更是将临时信息稳定编码进神经回路的物理过程。
"The actual reordering of the connections, the strengthening of connections between neurons that underly learning, the weakening of those connections occurs during sleep in particular during rapid eye movement sleep." "The 'first night effect' is the experimentally observed phenomenon whereby information that you learn on a given day is mostly Consolidated during the night's sleep that you have on that first night after the learning occurs." "Sleep is without question the best neutropic... your ability to be alert and focused is going to be greater if you slept well the night before."
专注力是学习的前提,而专注力本身是一种可训练的技能。当感到学习“吃力”或“痛苦”时,实际上是大脑在释放肾上腺素等神经调节剂,这是神经塑性启动的必要信号。为了优化这一状态,可采用两种低成本行为工具:
"That straining that you feel, that encouraging or in some cases forcing yourself to attend... reflects in part the release of neuromodulators like epinephrine adrenaline in the brain and body which serve to cue the neural circuits that they need to change." "People who do a 10-minute meditation per day... improve their level of focus, they improve their memory and recall ability." "Visual attention and cognitive attention comes back to that visual Target over and over again... it is known to create significant improvements in your ability to focus."
针对约 700 名高成就医学生的调研揭示了高效学习的共有习惯。顶尖学生并非依靠天赋,而是依靠严格的行为协议。核心习惯包括:
"They set aside time to study. They literally schedule time to study... they are very effective or they make it a point of putting their phone away and off of isolating themselves." "The best performing students seem to study alone... they make it a point to learn material in isolation then bring that material to other students in the same course and teach them." "In my laboratory for years we used to have a saying... watch one, do one, teach one. That was an excellent means to learn... to develop proficiency and even Mastery."
测试效应的核心在于:测试不仅是衡量知识的尺度,更是构建知识、巩固记忆和对抗遗忘的最强工具。研究表明,单纯的“重复阅读”是一种被动的学习方式,效果远逊于“主动召回”。 在一项经典研究中,对比了三组学习方式:第一组连续阅读材料四次(SSSS);第二组阅读三次后自测一次(SSST);第三组仅阅读一次后连续自测三次(STTT)。最终的长期记忆测试显示,测试次数越多(STTT组),记忆表现越好。 测试能将遗忘率降低约 50%。其机制在于:自测强迫大脑进行“召回”(Recollection),这比简单的“辨识”(Recognition)更能深度编码神经环路。最高效的测试形式是开放式简答题,因为它要求学习者在没有提示的情况下从零构建答案,这种“费力”的过程正是触发神经可塑性的关键。
"Testing is not just a way of evaluating what knowledge you've Acquired... It also turns out to be the best tool for offsetting forgetting of any kind." "The more tests that you take as a way to expose yourself to the material, the better you're going to perform on that material at some later point." "Reading and rereading material is far less effective than reading material and then thinking about that material, testing yourself on that material, forcing yourself to bring that material to mind."
学习过程中的“主观感受”往往具有欺骗性。实验发现,那些通过“重复阅读”学习的学生,对自己的考试表现拥有极高的自信度,因为多次接触材料让他们产生了“熟悉感”(Familiarity)。然而,实际测试结果恰恰相反:这部分学生的表现最差。 与之形成鲜明对比的是,通过“自测”学习的学生往往感到非常吃力,在练习中可能大量出错,因此他们的主观自信度极低,甚至认为自己没掌握。但最终,这组学生的实际考试成绩却是最高的。 这种错觉被称为“流利度偏见”。学习时的“费力感”和“挫败感”实际上是神经系统释放肾上腺素和乙酰胆碱的信号,这些神经递质在触发突触连接的增强或减弱(神经可塑性)中起核心作用。如果你觉得学习很轻松,那大概率你并没有在有效学习。
"When you're exposed to material over and over and over again, you think you've learned the material... actually you haven't learned it at all compared to the people that are exposed to the material and then take tests." "The students in the study-test-test-test group report much lower confidence in the material... and guess what? The exact opposite is true." "That strain that you feel, that agitation is great. That's a cue that you are learning or that you set the learning process in motion."
遗忘在接触新信息的瞬间就已经开始,因此测试的时机至关重要。研究对比了三种模式:
"The best performance comes from being exposed to material and then taking a test very soon after... as opposed to the last group [cramming before the final] which performs worst." "Test yourself on the material that you were just exposed to very soon after your first exposure to it because that offsets the natural forgetting of new material." "This business of putting the testing soon after exposure to new material is about offsetting the forgetting of that material."
测试的核心功能并非仅仅是评估(Evaluation),而是作为学习工具来抵消遗忘(Offset Forgetting)。研究显示,即使只在首次接触知识后进行一次自我测试,也能减少50%的遗忘量。在测试形式上,开放式简答(Open-ended short answer)远优于多项选择题。多项选择主要依赖“熟悉感”或“辨认能力”,这种被动的认知并不等同于知识的“敏捷性”或“精通”。
相比之下,开放式问题要求大脑进行“主动提取(Active Recall)”,这会迫使神经系统在识别到知识盲点时产生挫败感,进而触发神经重塑信号。一个关键的心理学现象是:反复阅读材料的学生通常会有很高的“伪自信”,认为自己掌握了知识,但实际测试表现最差;而通过不断自测的学生虽然在学习过程中自信心较低且感到吃力,但其长期记忆的持久性和准确性最高。
"Rather than think about studying to learn and retain information, I want you to think about studying to offset the natural process of forgetting."
"Testing is not just a way of evaluating what knowledge you've acquired... It also turns out to be the best tool for offsetting forgetting of any kind."
"Reading and rereading material is far less effective than reading material and then thinking about that material, testing yourself on that material, forcing yourself to bring that material to mind."
针对近700名医学生的研究发现,表现最顶尖的学生都有一个共同习惯:向同伴讲解所学内容。这种教学行为遵循“看一个、做一个、教一个”(Watch one, do one, teach one)的经典模型。在准备教学的过程中,学习者必须将碎片化的信息整合为逻辑严密的系统,这实际上是对知识掌握程度的极限测试。
教学能帮助学习者精准识别自己的“认知断点”——即当你无法清晰解释某个环节时,那就是你尚未理解的部分。尽管在竞争环境下,学生可能担心教别人会给自己带来劣势,但数据证明,通过教学获得的知识深度和灵活性,远超那些仅靠独自钻研的学生。这种方法将原本被动的吸收转化为了主动的、社会性的输出,极大地强化了神经连接。
"The other thing that they do, and this is very important, is that they make an effort to then teach their peers to teach other students in the class."
"Don't be afraid to be a teacher of your peers in order to test... and develop mastery of the material."
"Watch one, do one, teach one is an excellent means to learn... to develop proficiency and even mastery."
“间隙效应(Gap Effects)”是一种极高效率的神经辅助工具,指在专注学习或练习过程中插入短暂的随机停顿(约5到30秒)。神经科学研究发现,在这些微小的静止瞬间,大脑中的海马体(Hippocampus)会自动以比现实快20到30倍的速度回放刚刚接触的信息。
这种高频回放机制与深度睡眠(尤其是REM睡眠)中的记忆巩固过程非常相似。这意味着,通过在清醒的学习时段引入这些间歇,你实际上在诱导大脑在短时间内产生大量的“虚拟重复”。这种方法不需要增加额外的总学习时间,却能通过神经元的超速重演,极大地提升知识或运动技能的内化速度,是利用大脑底层生物学特性的“作弊码”。
"Gap effects are what I just did, which is to take periodic pauses... during which your hippocampus repeats information that you've been exposed to at a rate 20 to 30 times faster than typical."
"Those breaks... turn out to enhance overall learning ability probably... because the neural circuits are able to generate more repetition similar to gap effects."
"Pause and let your hippocampus generate more repetitions of that material than it would otherwise if you just tried to barrel through."
情绪高度波动的体验通常能实现“单次学习(One-trial learning)”,其核心机制在于肾上腺素(Epinephrine)和去甲肾上腺素的大量释放。这些神经调节物质能够标记当前的神经环路,促使大脑将其视为重要信息而非“垃圾”进行保留。中世纪曾有一种极端的教学实践:在向儿童传授重要信息后将其丢入冷水中,通过冷水诱发的急性应激(肾上腺素飙升)来巩固记忆。在现代学习中,适当的压力、专注产生的张力,乃至摄入咖啡因或课后的冷水浴/冷水脸刺激,都能通过类似的生理唤醒机制,增加神经系统的可塑性并抵消自然遗忘。
"Negative stuff is remembered typically the first time and every time and very durably over time as compared to positive experiences... because of these neuromodulators like epinephrine and norepinephrine." "Adults and kids were taught information and then thrown literally thrown into cold water why to deploy adrenaline and consolidate memory of the material they were exposed to." "That straining that you feel reflects in part the release of neuromodulators like epinephrine adrenaline in the brain and body which serve to cue the neural circuits that they need to change at a later time."
传统的线性灌输式学习(如“消防栓喷水”般持续输出)容易导致认知疲劳和干扰。交错学习通过在核心内容中插入看似无关的轶事、故事或非线性信息(例如Huberman在讲座中提到他的斗牛犬),为大脑创造“间歇”。这种做法有两个好处:首先,它类似“间隔效应(Gap effects)”,让海马体在极短的停顿中以20-30倍的速度回放刚学到的信息;其次,它通过引入随机元素,促使编码区域将新知识与已有的背景知识或远程关联知识进行整合。这种“非线性插入”不仅能缓解疲劳,还能提升知识提取的灵活性。
"By injecting other information that seems totally unrelated random or pseudo random it allows the brain areas that are responsible for encoding information to take whatever new information you're learning and to incorporate it with existing knowledge." "Those breaks of interleaving information not just getting a steady barrage like drinking from a fire hose... turn out to enhance overall learning ability." "Your hippocampus the neurons in your hippocampus repeat information that you've been exposed to for the first time at a rate 20 to 30 times faster than typical."
学习是一个从低到高的层级演进过程:1. 无技能(Unskilled):认知极其有限;2. 技能化(Skilled):能识别并基本运用信息;3. 掌握(Mastery):拥有深厚的知识背景,能够灵活、高效且稳定地运用所学;4. 艺术化境界(Virtuosity):这是学习的巅峰。在这一阶段,技能和知识已经完全“脚本化(Scripted)”进神经系统。处于该境界的个体(如顶级运动员或音乐家)不仅能完美执行,甚至能主动在表现中引入“不确定性”和“自发性”,通过即兴发挥和迎接“X因素”来创造出连自己都感到惊讶的卓越表现。
"Mastery typically means that we have... close to the full depth of knowledge in a given area and that we can use it pretty flexibly." "Virtuosity... is where we actually have such Mastery of material that we can use it in ways that we still don't even know how we can use... we even invite elements of uncertainty and kind of spontaneity into the use of that material." "The neural connections that remain become much more robust they become much more reliable."
构建高效学习工作流的核心逻辑在于“对抗遗忘”,而非简单的信息获取。神经塑性的本质是神经元间连接的强化或削弱,这一过程分为两个阶段:触发期(专注与警觉)和巩固期(深度睡眠)。
1. 前置准备与专注力训练:
2. 学习期间的策略:
3. 核心驱动力:测试效应(Testing Effect):
4. 巩固与教学:
"Rather than think about studying to learn and retain information, I want you to think about studying to offset the natural process of forgetting that everybody experiences." (与其想着为了学习和保留信息而读书,我希望你将学习视为抵消每个人都会经历的自然遗忘过程。)
"That strain that you feel, that agitation is great, that’s a cue that you are learning or that you set the learning process in motion." (你感到的那种压力感和焦虑感其实非常棒,那是你的神经系统正在学习或者你已经启动了学习过程的信号。)
"When you self-test material, you halve the amount of forgetting that occurs compared to if you're just exposed to the material." (当你对所学材料进行自我测试时,与仅仅接触该材料相比,你发生的遗忘量会减少一半。)
根据神经生物学家 Andrew Huberman 的研究,学习的本质并非简单的信息累积,而是一个“抵消自然遗忘过程”的行为。神经塑性(Neuroplasticity)是这一过程的核心,它指神经系统根据经验改变自身的能力。 在生物学层面,这种抵消遗忘的机制主要通过三种方式实现:首先是加强现有的突触连接,使神经元间的通信更高效;其次是削弱或移除不必要的连接(这在运动技能学习中尤为重要,通过去除冗余动作实现精准);最后是极少数情况下的神经发生(即产生新神经元)。学习过程中,专注力会释放如肾上腺素等化学物质,为神经回路打上“需要改变”的标记,从而在后续的休息中稳固这些连接,使原本会被大脑自然清理的信息得以长期保留。
“自测”之所以优于“反复阅读”,是因为它从根本上区分了“熟悉感”与“掌握力”。
高效学习是一个由“触发”和“执行”组成的双步过程:
这种现象源于大脑对“熟悉感”(Familiarity)与“回忆能力”(Recollection)的混淆。研究表明,反复阅读或接触同一材料的学生通常表现出极高的自信度,因为随着接触次数增加,大脑识别信息变得轻松,产生了一种“我已经掌握”的错觉。然而,这种被动的重复仅仅停留在识别层面,而非真正的提取能力。
相比之下,采用“学习1次+测试3次”模式的学生,虽然在测试过程中会经历痛苦的挣扎、意识到自己的知识盲区,甚至因表现不佳而自信心受挫,但这种“提取练习”强制神经系统建立更牢固的连接。由于测试暴露了知识的缺失,学生能更准确地识别哪些内容尚未掌握,这种挑战性的过程虽然降低了主观自信,却能极大程度地抵消自然的遗忘过程,最终在实际考试中表现远超那些仅凭“熟悉感”而自我感觉良好的学生。
最佳的自测时间点是在首次接触新材料后的极短时间内,最好是当天或次日。科学研究显示,学习并非简单的获取,其核心在于“抵消遗忘”。在信息刚进入大脑时立即进行测试,可以起到一种“接种疫苗”的作用,防止大脑将这些新信息作为无用数据剔除。
如果测试时间过迟(例如在学习后很久才进行首次自测,即使是在正式考试前几天),即使学生在测试后得到了正确答案的反馈,其最终表现依然是最差的。这是因为自然的遗忘曲线在长时间的空窗期内已经抹去了大部分记忆痕迹。一旦信息已经从神经回路中流失,后期的补救性测试就很难再发挥出那种早期巩固的效果。因此,早期的测试能锁定记忆,而延迟的测试则是在试图重建已经坍塌的建筑。
“间隙效应”是指在学习或练习过程中,有意识地插入极短的停顿(通常为5到30秒)。这些看似无意义的空白期实际上是神经系统的高效处理期。神经科学数据发现,在这些微小的间隙中,大脑的海马体(记忆的关键区域)会以比正常处理速度快20到30倍的频率,对刚刚接触的信息进行“快速回放”。
这种现象类似于我们在深度睡眠(特别是REM睡眠)中所经历的记忆巩固过程,但在清醒的学习状态下即可发生。通过这种快速回放,神经回路在短时间内获得了远超实际练习次数的重复机会。例如,如果你在学习一段新知识或练习一个新动作时停顿10秒,你的海马体可能已经自主复习了几十次。这种机制极大地提高了记忆的编码频率,使大脑能在不增加实际学习时长的情况下,更深层地强化神经连接。
开放式问题(如简答题或自我陈述)要求大脑进行“主动回忆”(Recall),而多项选择题通常只涉及“辨认”或“熟悉感”(Familiarity)。在神经生物学层面,识别正确答案是一种相对被动的过程,容易产生“我已经懂了”的错觉。相比之下,开放式问题强制神经系统从零开始检索和构建信息,这种深度的认知努力能更清晰地暴露知识盲点,并强化神经元之间的突触连接。通过这种方式,学习者不仅能记住事实,还能培养对知识的敏捷性和灵活性,从而实现从简单的信息获取到真正的“知识精通”(Mastery)的跨越。
情绪波动和压力会促使身体分泌肾上腺素(Epinephrine)和去甲肾上腺素,这些神经调节物质是记忆形成的强力“固化剂”。当这些化学物质在接触新信息后的一定时间内大量释放时,它们会向大脑发出强烈信号,标记该信息具有极高的生存或应用价值,从而加速神经可塑性过程。历史上甚至有在中世纪通过将学生丢进冷水中来强化记忆的极端做法,其原理正是利用冷水激发的肾上腺素飙升来锁定刚刚学到的知识。这种生理反应能显著抵消自然的“遗忘过程”,使记忆在神经回路中更稳固、更长久地保留。
研究表明,顶尖的学习者(如优秀的医学生)通常具备高度的“空间孤立”和“时间规律”习惯。在环境上,他们倾向于独处并彻底消除干扰,例如关闭手机或断开社交联系,以保护有限的注意力资源。在时间分配上,他们通常每天投入3至4小时进行深度学习,但并非连续苦读,而是将其拆分为2至3个集中的时间段。更重要的是,他们利用大脑的昼夜节律,在每天固定的时间点进行学习,这种规律性可以诱导神经系统在特定时段自动进入高专注状态。此外,他们在独立掌握知识后,往往会通过“教导他人”来进一步验证和深化自己的理解。
“交替学习”和插入轶事信息并非简单的“休息”,而是一种高度有效的神经生物学策略,通过以下几个核心机制优化大脑的编码和整合过程:
打破“消防栓式”的信息过载(Firehose Effect): 当大脑持续接受高强度、单一类型的知识输入时,神经环路容易产生疲劳,导致编码效率下降。通过在核心学习内容中穿插一些看似无关或低难度的轶事、故事或伪随机信息,大脑获得了一个“功能性间歇”。这种“交替”能够像“间隙效应”(Gap Effects)一样,触发海马体以比平时快 20 到 30 倍的速度回放刚刚学到的新知识,从而在潜意识中增加了重复练习的次数。
建立跨领域的神经锚点(Neural Anchoring): 单纯的抽象知识往往是孤立的,难以被大脑长期保留。插入轶事或故事能促使负责信息编码的脑区将“新信息”与“现有知识”或“疏远但熟悉的记忆”联系起来。例如,Huberman 提到在讲授神经科学时插入关于他的斗牛犬 Costello 的轶事,这促使听众的大脑将复杂的生物学概念挂钩到“狗”这一已有的、稳固的神经环路上。这种“联想式整合”为新知识构建了一个脚手架,使其不再是孤立的神经放电,而是现有神经网络的一部分。
利用神经调制物质的“标记”作用(Tagging by Neuromodulators): 故事和轶事通常带有情感色彩或叙事张力,这会触发肾上腺素(Epinephrine)和去甲肾上腺素(Norepinephrine)的释放。这些神经调制物质不仅能提高警觉性和注意力,还具有“标记”作用。当大脑处于一种由于故事带来的轻微情感波动或警觉状态时,它会向神经系统发出信号,表明当前的信息是“重要的”。这种化学环境会显著增强突触连接的强度,从而帮助新知识更稳固地“固化”在现有的神经回路中,极大地抵消了自然遗忘的过程。
促进知识的灵活性(Flexibility and Mastery): 交替学习强迫大脑不断地在不同类型的信息间进行切换。这种切换过程要求大脑不仅仅是“识别”信息(熟悉感),而是要主动“提取”信息(精通感)。这种深度的认知加工不仅加深了记忆,还赋予了知识灵活性,使学习者未来能够将这些新知识应用于不同的背景或解决复杂的问题,而非仅仅死记硬背。