研究揭示大脑“举一反三”的神经机制

Core Insights - The research reveals a unique neural organization pattern in the primate brain, which aids in understanding the mechanisms behind the ability to transfer learning across different tasks [1] - The study indicates that primates, similar to humans, can extract abstract rules from tasks and apply them to new problems, highlighting cognitive similarities [1] Research Findings - The research team conducted a series of training tasks with macaques, discovering that their brains form two independent representational spaces during task completion [1] - One representational space is responsible for stable core decision logic, while the other encodes specific sensory features of the current task, allowing adaptation to changing external conditions [1] - This neural organization enables the independent operation of stable knowledge and new information, facilitating both stable storage and flexible updating of knowledge [1] Implications - The findings deepen the understanding of learning mechanisms in primate brains and provide new design ideas for developing next-generation artificial intelligence systems with rapid learning and adaptability [1]