Neeranuch Kittikiatkumjorn¹ (), Waragon Phusuwan^1,2, Payachana Victoria Chareunsouk^1,2, Chaipat Chunharas^1,3,4, Sedthapong Chunamchai^1,2,3,4; ¹Cognitive Clinical and Computational Neuroscience Center of Excellence, Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand, ²Medical sciences program, Division of Neuroscience, Faculty of Medicine, Chulalongkorn university, Bangkok, Thailand, ³Division of Neurology, Department of Medicine, Faculty of Medicine, King Chulalongkorn Memorial Hospital, Thai Red Cross Society, Bangkok, Thailand, ⁴Chula Neuroscience Center, King Chulalongkorn Memorial Hospital, Bangkok, Thailand

Despite naturally noisy and limited inputs, the human visual system rapidly and robustly detects faces, extending to illusory face percepts in inanimate objects—known as face pareidolia. One hypothesis proposes a filling-in process that integrates visual inputs with prior knowledge. Although previous studies have demonstrated that misperception at the periphery—with increased sensory uncertainty due to degraded visual acuity— reflects greater reliance on surrounding low-level contextual information (e.g., contrast), it remains unclear whether high-level contextual information (e.g., faceness) similarly biases illusory face perception, and whether such modulation would vary with eccentricity. To address this gap, we used Mooney faces with quantified faceness scores, controlling for low-level properties, in a face detection task. We manipulated high-level contextual information by varying mean display faceness across multiple combinations of high-, low-, and scrambled non-face images. Eighteen participants completed two 300-trial blocks. Images were presented for 150 ms across four spatial locations at 3° or 6° eccentricity. Participants reported whether they perceived a face at each randomly probed position. At non-face locations, false-alarm rates were significantly higher at 6° than at 3° eccentricity, F(1,17) = 11.63, p < .01. Contextual faceness significantly interacted with eccentricity, F(6,102) = 2.51, p < .05. The eccentricity effect was strongest under low contextual faceness but diminished as contextual faceness increased, with convergence at higher contextual faceness. In contrast, contextual faceness did not differentially affect miss-rates across eccentricities, suggesting that the pattern was specific to false alarms rather than a uniform response shift from degraded visual acuity. These findings demonstrate that illusory face perception is jointly shaped by sensory uncertainty and high-level contextual information, with contextual effects depending on eccentricity. This pattern is consistent with the Bayesian framework in which the influence of priors increases with higher sensory uncertainty.