Cheongil Kim¹ (), Sang Chul Chong¹; ¹Yonsei University

Despite the general impression of a stable visual world, our visual performance fluctuates continuously—even within a second—due to dynamic shifts of attention. Specifically, attention varies transiently following exogenous cues (transient attention) and rhythmically in synchrony with intrinsic neural oscillations (rhythmic attention). Here, we asked whether people can monitor these subsecond attentional dynamics and what mechanisms support such metacognitive monitoring. Using an exogenous cueing paradigm combined with confidence judgments, we examined whether confidence tracks fluctuations in orientation discrimination (Experiment 1) and target detection performance (Experiment 2) across cue–target intervals. To this end, we decomposed fluctuations in performance and confidence into subcomponents reflecting transient and rhythmic attention and examined performance-confidence correlations within each component. In Experiment 1, we observed positive correlations between orientation discrimination performance and confidence for both transient and rhythmic attention, indicating accurate metacognitive monitoring. In Experiment 2, we aimed to further clarify the mechanisms underlying metacognitive monitoring using a target detection task (Kim & Chong, 2025), focusing on whether observers monitor perceptual evidence (e.g., the quality of target perception) or visual system states (e.g., attentional states). Confidence in detection tasks can be informed by perceptual evidence when judging target presence (e.g., high confidence when the target is clearly perceived), but not when judging target absence, where little perceptual evidence is available. Therefore, if confidence is based solely on perceptual evidence, it should correlate with performance fluctuations for target-present responses but not for target-absent responses. Consistent with this prediction, we found positive correlations between detection performance and confidence for target-present responses but little correlation for target-absent responses. These findings suggest that metacognitive monitoring of attentional dynamics relies primarily on perceptual evidence rather than direct readouts of internal states.

Acknowledgements: This work was supported by the Yonsei University Research Fund (Post Doc. Researcher Supporting Program) of 2025 (project no.: 2025-12-0208) and the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (RS-2022-NR070542).