Scholarlink

<p>In reward foraging tasks, prefrontal neurons track reward history, yet animals also show persistent choice-history biases. How these histories are represented in prefrontal circuits and guide animals’ decisions remains unknown. We asked whether past rewards and choices are incorporated by leaky integration or carried as discrete, history-specific codes, and how these codes are recruited under different task demands. We recorded medial prefrontal cortex (mPFC) activity while mice performed probabilistic reward foraging task and fit a reinforcement-learning model whose decision variable, combining reward and choice histories, captured behavior. Neurons represented history-specific rewards and choices while integrating them consistent with their behavioral impact. We then altered reward contingencies and inter-choice intervals and transiently inactivated mPFC. Neural representations adapted to changing task demands, yet the behavioral impact of inactivation was sensitive to inter-choice interval and reward contingencies. We conclude that mPFC hosts redundant computations whose influence is gated by timing and task structure.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Time-dependent deployment of medial prefrontal cortical representations in male mice

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月12日

查看全文
  • Junior Samuel Lopez-Yepez,
  • Anna Barta,
  • Juliane Martin,
  • Maria Moltesen,
  • Tsz-Fung Woo,
  • Oliver Hulme,
  • Ebru Demir,
  • Duda Kvitsiani

摘要

In reward foraging tasks, prefrontal neurons track reward history, yet animals also show persistent choice-history biases. How these histories are represented in prefrontal circuits and guide animals’ decisions remains unknown. We asked whether past rewards and choices are incorporated by leaky integration or carried as discrete, history-specific codes, and how these codes are recruited under different task demands. We recorded medial prefrontal cortex (mPFC) activity while mice performed probabilistic reward foraging task and fit a reinforcement-learning model whose decision variable, combining reward and choice histories, captured behavior. Neurons represented history-specific rewards and choices while integrating them consistent with their behavioral impact. We then altered reward contingencies and inter-choice intervals and transiently inactivated mPFC. Neural representations adapted to changing task demands, yet the behavioral impact of inactivation was sensitive to inter-choice interval and reward contingencies. We conclude that mPFC hosts redundant computations whose influence is gated by timing and task structure.