Nino Sharvashidze¹, Matteo Toscani², Matteo Valsecchi¹; ¹University of Bologna, ²Bournemouth University

Peripheral appearance is heavily influenced by top-down assumptions about the statistics of our visual environment. But does metaperception, i.e. visual confidence take this into account? Evidence is mixed, studies (e.g., Odegaard et al., 2018; Solovey et al., 2015) reported overconfidence in peripheral detection tasks, but Toscani et al. (2021) reported underconfidence when observers discriminated the orientation of a peripheral patch. Here we tested whether underconfidence is always associated with peripheral discrimination or if it is contingent on whether the specific task conforms to the constraints of peripheral vision. We used a task where peripheral vision performs well – rapid scene categorization (Larson & Loschky, 2009). Our 12 participants categorized scenes as Desert, Beach, Mountain or Forest in a two-interval four-alternative forced choice task. In each interval, they viewed the scene either only in the periphery (scotoma) or only in the fovea (window). Subsequently, they indicated the interval in which they were more confident in their judgement. Task difficulty was manipulated by varying the scotoma size, with larger scotoma corresponding to less peripheral information, or the window size, with larger window corresponding to more central information. Accuracy decreased with the increasing size of the scotoma and increased with the increasing size of the window. We computed the probability of higher confidence in the periphery as a function of performance difference between the two conditions. Participants’ points of equal confidence were systematically shifted towards higher foveal perceptual performance, demonstrating overconfidence in the periphery. This suggests that changing the task from local orientation discrimination to global scene categorization, i.e. a task where peripheral vision outperforms foveal vision, reversed the metaperceptual bias. Periphery is suited for detecting objects and categorizing scene gist, but not for discriminating fine details or local features. Metacognitive judgements seem to follow these inherent capabilities and constraints of peripheral vision.

Acknowledgements: The study was funded by the Italian Ministry of University and Research under the PRIN 2022 programme, project number 2022HEKCWH.