The abstract is read at the publisher; the summary is JClub's.
Confidence in global motion direction judgments is better explained by perceptual reliability than by stimulus-based heuristics, with a normalized motion-energy model accounting for most variance in both decisions and confidence.
- Why it matters: The field debates whether metacognitive confidence tracks how reliable perceptual estimates are or instead follows simple stimulus statistics. This task removes shortcut cues by making global direction available only after integrating locally ambiguous motion signals.
- What they did: Observers reported global direction across two successive intervals and chose which interval was more likely correct, while confidence was compared against motion coherence. Two image-computable motion-integration models were tested: a standard motion-energy model versus a normalized motion-energy model with tuned inhibition and an untuned gain control.
- The result: Confidence tracked accuracy beyond what coherence alone explained, including higher confidence for opposite-direction errors and for chosen intervals at the same coherence. The standard model failed to reproduce opposite-direction errors and their confidence pattern, whereas tuned inhibition plus untuned gain in the normalized model accounted for most variance.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.