Purchase this article with an account.
Vincent de Gardelle, Pascal Mamassian; Is confidence amodal? Confidence comparison across dimensions of a visual stimulus. Journal of Vision 2013;13(9):57. doi: https://doi.org/10.1167/13.9.57.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Introduction: Our perceptual judgments in everyday life or in laboratory experiments are always accompanied with an introspective, subjective feeling of confidence: for a given response, we can estimate whether we have just given a random guess, or whether we are certain about the choice. However, the characteristics and mechanisms of these confidence judgments are still poorly understood. Here, we ask in particular whether confidence judgments in human observers are made in an abstract and generic frame of reference, or whether they are specific to a task. Methods: On each trial, observers were presented with two Gabor patches and had to judge both the change in orientation (clockwise vs. counter-clockwise) and spatial frequency (low-to-high vs. high-to-low) across the two patches. Then they had to report which perceptual decision (i.e. along which stimulus dimension, orientation or spatial frequency) was associated with higher confidence. Results: Observers’ ability to discriminate two orientations (or two spatial frequencies) naturally increased with the orientation (spatial frequency) difference between the two Gabor patches. Critically, we found that confidence judgments interacted with perceptual performance: perceptual decisions associated with high confidence choices formed a psychometric function with a steeper slope. To further characterize this interaction between confidence choices and discrimination sensitivity, we assessed the efficiency of confidence judgments by deriving the ideal observer performance in this procedure. Conclusion: Human observers have the ability of judging and comparing confidence across different discrimination tasks applied to the same stimulus. This finding demonstrates that there is an amodal (i.e. task-independent) component to confidence evaluation. Future studies will aim at assessing whether confidence is fully or only partially amodal, by assessing the potential cost (in terms of metacognitive efficiency) of a switch between two dimensions during the confidence comparison judgment.
Meeting abstract presented at VSS 2013
This PDF is available to Subscribers Only