Towards an integrative neuroscience of metacognition.

Structured transformations of uncertainty support metacognitive self-evaluations in humans and other animals, linking dynamic evidence accumulation and prefrontal integration to confidence-based self-beliefs.

• Why it matters: Metacognition is central to adaptive learning, social coordination, and forming self-beliefs, yet its computational and neural basis—especially how uncertainty becomes confidence—remains to be unified.
• What they did: The work integrates convergent behavioral, computational, and neural evidence across rodents, non-human primates, and humans, proposing evidence-accumulation mechanisms for local confidence and multimodal prefrontal integration for global beliefs.
• The result: This architecture is proposed to enable metacognitive control across timescales and to explain how dysfunctional metacognition can yield reality-monitoring and insight disorders with aberrant confidence.

AI summary of the abstract · may contain errors · not medical advice · read the paper.