研究人员发现大脑如何利用短期和长期的合成变化,将时间信号转化为稳定的记忆。 Researchers found how brains turn time-based signals into stable memories using combined short- and long-term synaptic changes.

天津大学的研究人员和国际合作者已经确定了一种大脑机制,将神经活动的序列随着时间推移转化为空间模式,从而能够有效地处理和储存信息。 Researchers at Tianjin University and international collaborators have identified a brain mechanism that converts sequences of neural activity over time into spatial patterns, enabling efficient information processing and storage. 该研究报告发表在PNAS中,表明长期和短期的合成变化共同有助于将动态投入(像旋律一样)转化为稳定的表象,在不扩大神经网络的情况下增强记忆和噪音抵抗能力。 The study, published in PNAS, shows that long-term and short-term synaptic changes work together to transform dynamic inputs—like a melody—into stable representations, enhancing memory and noise resistance without enlarging neural networks. 研究结果在来自鼠标和人类新动物的计算模型和电生理数据的支持下,揭示了用于时间处理的“合作代码”,为开发更有效、可解释的人工智能系统提供了见解。 Supported by computational models and electrophysiological data from mouse and human neocortices, the findings reveal a "collaboration code" for temporal processing, offering insights for developing more efficient, interpretable artificial intelligence systems.