Are vividness judgments in mental imagery correlated with perceptual thresholds?

Abstract

Previous work in neuroimaging suggests that exteroceptive visual perception and mental imagery activate similar brain areas within the ventral visual stream (e.g., Horikawa & Kamitani, 2017). However, it is still unclear to what extent vividness judgments and perceptual thresholds are determined by this overlap. We measured perceptual thresholds and mental imagery vividness for a set of 20 images from the Natural Scene Dataset (Allen et al., 2022). In the imagination task, we asked participants to report, for each individual trial, the vividness of their mental image (40 trials per image). In the perceptual threshold ABX task, we asked participants to recognise among two unaltered visual scenes (A and B), a target image (X) embedded in Gaussian noise. The Quest procedure (Watson & Pelli, 1983) was used in order to establish the contrast threshold per image (40 trials per image). Preliminary results (N = 8) indicate no, or very weak, within-subject correlations between the contrast thresholds and the mean vividness scores (mean correlation = -0.072; SD = 0.168; min = -0.240; and max = 0.250). This suggests that most of the variance in perceptual thresholds and vividness judgments of mental images is not determined by the same mechanism. These findings contribute to ongoing modeling and theoretical efforts aimed at deciphering how the brain can generate our rich and vivid mental images.