Angela Shen¹ (), Megan A. K. Peters^1,2; ¹University of California, Irvine, ²Program in Brain, Mind, & Consciousness, Canadian Institute for Advanced Research

Does visual metacognition accurately track performance differences due to visual field location? It seems not always: some evidence suggests observers can overestimate their performance in the visual periphery (perhaps especially for near-threshold stimuli) when performance is matched between central and peripheral stimulus locations. However, this effect has only been demonstrated across few performance levels or visual field locations. Here, we investigated the relationship between performance and confidence across a matched range of performance and at multiple stimulus levels and locations. In a 2AFC discrimination task, observers simultaneously reported the tilt (+/- 45 degrees) of a grating and their confidence on a 4-point scale. The grating appeared at one of four visual eccentricities (2, 4.67, 7.33, or 10 degrees of visual angle) on each trial, and one of eight polar angles (cardinal: 0°, 90°, 180°, 270°; intercardinal: 45°, 135°, 225°, 315°) in each block of trials. Thus, we recorded performance and confidence at 32 visual field locations. We staircased stimulus contrast at each eccentricity to obtain discrimination thresholds, selected four contrast levels around each threshold value, and fit psychometric functions to performance and to confidence ratings for each location. We applied a novel analytic approach – relative psychometric function (RPF) analysis (Maniscalco et al., submitted) – to compute mean confidence rating for each location and across a matched range of performance levels (stemming from the 5 stimulus contrast levels). The RPF method provides an intuitive summary statistic, the area under the RPF curve (RPF AUC), that quantifies the relationship between two psychological processes – here, performance and confidence. Results comparing RPF AUC across visual field locations suggest that the quantitative relationship between confidence and performance varies as a function of eccentricity and polar angle. These findings represent a significant step towards understanding polar angle-based and eccentricity-based asymmetry in performance-controlled subjective confidence.

Acknowledgements: We acknowledge the support of the Natural Sciences and Engineering Research Council of Canada (NSERC; to AS) and the Canadian Institute for Advanced Research (CIFAR, Fellowship in Brain, Mind, & Consciousness; to MAKP).