中文导读

人们常以为恐惧会让大脑把危险与安全分得更清，但麦吉尔大学与多伦多大学士嘉堡校区的小鼠研究给出不同图景：在恐惧学习后，腹侧海马中两类场景的神经表征更易重叠。研究团队用增强现实在安全与危险环境间切换，并用微型显微镜观察海马神经元活动；重叠或有助于更快提取威胁记忆，以速度换精确度，也可能在客观上安全的情境中触发强烈恐惧。ScienceDaily 于 2026-11-03 据麦吉尔材料报道，研究刊于 Nature Neuroscience，获 CIHR 与 NSERC 资助。阅读时可留意 rather than、yet、may、suggests 如何区分常见假设、动物发现与人类适用边界。

正文

Fear may blur the line between danger and safety in the brain, rather than make those memories more distinct. ScienceDaily reported the finding on 3 November 2026, drawing on materials from McGill University. The mouse study, led with the University of Toronto-Scarborough, appeared in Nature Neuroscience and focused on overlapping representations in the ventral hippocampus.

本段译文

恐惧或许会模糊大脑中危险与安全的界限，而不是让这些记忆更加分明。ScienceDaily 于 2026 年 11 月 3 日报道了这一发现，材料来自麦吉尔大学。这项与多伦多大学士嘉堡校区合作的小鼠研究发表在《自然·神经科学》上，关注腹侧海马中重叠的表征。

For years, many scientists believed that fear helps the brain separate threatening experiences from harmless ones. The new work challenges that idea. Researchers trained mice to link one environment with danger and another with safety. Instead of forming clearer, more separate neural maps, activity patterns for the two places began to overlap, especially in the ventral part of the hippocampus.

本段译文

多年来，许多科学家认为恐惧能帮助大脑把威胁经历与无害经历分开。这项新研究挑战了这一看法。研究人员训练小鼠把一种环境与危险联系，把另一种环境与安全联系。两组神经活动模式并未形成更清晰、更分离的脑内地图，反而开始重叠，尤其是在海马的腹侧部分。

“Our findings suggest fear may blur the boundary between danger and safety rather than sharpen it,” said senior author Mark Brandon, an associate professor of psychiatry at McGill and a researcher at the Douglas Research Centre. Lead author Robert Rozeske, an assistant professor of psychology at the University of Toronto, helped design an unusual test that combined augmented reality with real-time brain imaging.

本段译文

“我们的发现表明，恐惧可能模糊危险与安全的边界，而不是使其更清晰，”资深作者、麦吉尔大学精神病学副教授、道格拉斯研究中心研究员马克·布兰登说。第一作者、多伦多大学心理学助理教授罗伯特·罗泽斯克参与设计了一项不寻常的实验，把增强现实与实时脑成像结合起来。

In a specialized arena, sights and sounds could virtually teleport the animals between safe and dangerous settings without moving them. Miniature microscopes recorded hundreds of neurons in the hippocampus, a region central to learning and memory. That setup offered a moment-by-moment view of the brain switching between competing maps of threat and safety — something standard fear conditioning rarely shows so clearly.

本段译文

在专门场地中，视觉与声音能在不实际移动动物的情况下，把它们虚拟“传送”到安全或危险场景之间。微型显微镜记录了海马中数百个神经元的活动；海马是学习与记忆的关键区域。这一设置提供了大脑在威胁与安全的竞争地图之间切换的逐时画面——这是传统恐惧条件反射实验很少能如此清楚展示的。

Overlapping memories may sound like a flaw, yet the team argues they can serve a survival purpose. Shared patterns appeared to help the brain retrieve past threats more rapidly, trading some precision for speed. A false alarm may still be safer than missing a real danger. The ventral hippocampus, which processes emotionally charged experiences, may support this “better safe than sorry” response.

本段译文

记忆重叠听起来像缺陷，但研究团队认为它可能具有生存意义。共享模式似乎帮助大脑更快提取以往威胁信息，以一定精确度为代价换取速度。误报有时仍比错过真实危险更安全。处理情绪相关体验的腹侧海马，或许支持这种“宁可信其有”的反应。

The same idea may help explain anxiety and post-traumatic stress disorder. A harmless sight, sound or place can trigger intense fear when it pulls up a traumatic memory, even though the person is objectively safe. Brandon suggests this may be an overactive version of a normal survival mechanism, not simply a broken ability to keep memories apart. Funding came from Canadian Institutes of Health Research and the Natural Sciences and Engineering Research Council of Canada. Because the experiments used mice, further work is still needed in humans. Even so, the study reframes fear as a system that can shift into danger mode too quickly.

本段译文

同一思路或许有助于解释焦虑与创伤后应激障碍。一个本无害的景象、声音或场所，一旦勾起创伤记忆，就可能引发强烈恐惧，即便当事人客观上安全。布兰登提出，这或许是正常生存机制过度活跃的版本，而不仅仅是记忆分离能力损坏。研究获加拿大卫生研究院以及加拿大自然科学与工程研究理事会资助。由于实验使用小鼠，人类机制仍需进一步研究。即便如此，这项研究把恐惧重新理解为一个可能过快切换到“危险模式”的系统。

新闻来源

• ScienceDaily · The brain’s hidden fear mechanism may explain anxiety and PTSD (3 Nov 2026)
• McGill Newsroom · Survival mechanism in the brain may help explain PTSD
• Nature Neuroscience · Overlapping representations in the ventral hippocampus support rapid fear memory retrieval (Rozeske, Brandon et al.)

✳更多文章 →