I expect, therefore I see: individual differences in visual awareness

Predictive processing theories posit that awareness of the visual world emerges as the brain engages in predictive inference about the causes of its sensory input. At each level of the processing hierarchy top-down predictions are corrected by bottom-up sensory prediction error to form behaviourally optimal inferences about the state of the visual world. Research suggests there may be individual differences in predictive processing mechanisms such that some individuals are more reliant on prior knowledge, whereas others assign more weight to sensory evidence. Predictive processing biases are thought to manifest in a range of typical and atypical perceptual experiences including proneness to perceptual illusions, sensory sensitivity in autism, and hallucinations in psychosis. The overarching aim of this thesis was to investigate whether in the general population predictive processing biases predict individual differences in visual awareness. Change blindness was selected as the central paradigm of investigation, as it can be conceptualised as a failure to incorporate a novel change into the current prediction about the state of the visual world. The empirical work in Chapter 2 aimed to characterise individual differences in visual change detection using naturalistic scenes and to identify the perceptual and cognitive measures that predict noticing ability. There were reliable individual differences in change detection that generalised to ecologically valid displays. The ability to notice visual changes was predicted by the strength and stability of perceptual predictions, as measured by the accuracy of visual short-term memory and attentional control in the face of distractors. In Chapter 3 I used voxel-based-morphometry to investigate whether inter-individual variability in brain structure predicts individual differences in visual awareness. The latter was assessed by the change blindness task as well as its strongest predictor measures (visual short-term memory, attentional capture, and perceptual rivalry). Regions of interest (ROIs) were selected in the parietal and visual cortices based on previous evidence that these areas are causally involved in the awareness of visual stimuli. This study aimed to discover whether the average grey matter density in the ROIs predict susceptibility to CB. The ROI-based analyses revealed the average grey matter density in left posterior parietal cortex predicted visual short-term memory accuracy but none of the other hypothesised relationships were significant. Chapter 4 aimed to measure individual differences in the reliance on prior knowledge by employing the Mooney face detection task. In this task participants disambiguated faces in two-tone degraded images before and after the presentation of the original versions of the images. Better change detection was predicted by Mooney face detection without any prior knowledge of the images, a measure of ‘perceptual closure’ or an ability to generate a gestalt of a scene. The attention to detail subscale of the autism spectrum also predicted superior change detection. Reliance on prior knowledge in visual perception (assessed by improvement in Mooney face detection after seeing original images) did not consistently predict atypical perceptual experiences associated with the autism spectrum or schizotypy. Chapter 5 was an investigation into, firstly, whether there is a general predictive processing bias, which manifests across different methods of inducing prior knowledge, or whether such a bias is paradigm-specific and, secondly, whether reliance on priors predicts perceptual experiences and traits. All prior manipulations in this study lead to an increased tendency to see the expected stimulus in a binocular rivalry display, except adaptation, which lead to a suppression of visual awareness. Attentional control, perceptual priming, expectancy, and imagery loaded onto a common factor, suggesting that the strength of selective attention is closely linked with the facilitatory effect of expectation. The strength of adaptation predicted superior change detection and perceptual priming predicted the propensity to experience perceptual illusions. Taken together, these findings suggest that there are reliable individual differences in visual change detection, and these are predicted by the strength of visual short-term memory representations, attentional control, perceptual closure ability, as well as the strength of low-level adaptation. Possessing expectations facilitates the entry of the corresponding percept into awareness, irrespective of the method of prior induction. The facilitatory effect that priors exert on visual awareness across different methods is closely linked with the ability to exert attentional control. This suggests that the effects of expectations on awareness may be attentional. However, predictive processing biases were method-specific in that a facilitatory effect using one prior induction method will not necessarily predict the magnitude of the effect using a different method. Some prior effects (e.g., perceptual priming, imagery, and adaptation) yielded correlations with perceptual experiences and traits in the general population. As the research in this thesis is correlational, future studies will need to delineate the effects of expectation, attention, and adaptation on visual awareness and explore the neural representations of these mechanisms.