Metacognitive expression activates attention networks and increases episodic recall detail during naturalistic film viewing

Continuous self-report during ongoing experience provides a high-resolution window into subjective processing, yet concerns that it may disrupt the neural and cognitive processes being measured have limited its use in neuroimaging. We tested this by examining how expressing metacognitive judgment influences neural activity and behavior during complex social inference. Thirty-five participants underwent fMRI while watching an episode of a murder-mystery show and evaluating their certainty of a central character’s guilt. Participants continuously rated their certainty during one half of the episode (expressive metacognition) and viewed the other half without providing ratings (reflective metacognition), followed by surprise free recall and character evaluations. Expressive metacognition increased activation across salience, dorsal attention, and control networks, including medial frontal, parietal, and lateral prefrontal regions, whereas reflective metacognition yielded greater activation within a single anterior vmPFC cluster. The two conditions were also associated with increased intersubject synchrony in control and default mode networks, respectively. Despite these differences, they did not differ in which or how many scenes were recalled, recall order, or socioemotional character judgments, indicating that core aspects of episodic organization and social evaluation remain intact. However, expressive metacognition yielded greater episodic detail in recall. These findings show that expressing judgments reshapes neural processing and shared responses while leaving many downstream outcomes unchanged. supporting continuous self-report as a viable tool for studying dynamic cognition in naturalistic settings, provided its engagement of attention and control networks aligns with the research question.