Abstract
Visual illusions offer the unique opportunity to examine the co-ordination of bottom-up and top-down processing by generating a predictable mismatch between sensory inputs and the subjective perception. As such, they represent useful tools in exploring anomalous perceptual experiences in clinical populations. Although still controversial, there is an overall trend in the literature suggesting an increased resistance to visual illusions in people with psychosis. Here, we quantified illusion susceptibility to a battery of 13 visual illusions in a clinical group of 25 young people reporting psychotic-like experiences and a control group of 74 participants. Depression, anxiety and stress levels were measured in both groups by means of a questionnaire, whereas frequency, appraisals and emotional responses to psychotic-like experiences were examined in the clinical group only by means of a semi-structured interview. In contrast to the general finding of reduced illusion strength in those with psychosis, we observed that 10 out of 13 illusions tested generated greater effects in the clinical group compared to the control group. However, such a between-groups difference disappeared once depression, anxiety and stress levels were controlled for, suggesting that the overall increased susceptibility reported by the clinical group could be due to the severity of mental health problems, namely high levels of depression, anxiety and stress, associated with the psychotic-like experiences. Specifically, stress level turned out to be the best predictor of illusion strength in the clinical group. Results also revealed that illusion susceptibility in the clinical group was unrelated to anomalous experiences, depression and anxiety. We conclude that the tendency for the clinical group to exhibit greater vulnerability to illusions than the control group can be explained by increased levels of stress experienced by those reporting psychotic-like experiences. This enhanced susceptibility might suggest a distinct perceptual style biased towards top-down messages carrying prior expectations.