On December 16 at 12:00, Kadi Tulver will defend her doctoral thesis „An investigation of individual differences in the effects of priors on visual perception“.
Supervisor:
Professor Talis Bachmann, University of Tartu
Opponent:
Professor Jamie Ward, University of Sussex, UK
Summary
An enormous body of work has been devoted to understanding how visual perception works. And yet, there is still a lot we do not understand about the underlying structure of visual perception and the many complex cognitive mechanisms involved in creating a seamless phenomenological experience of our surroundings. One promising theory that is inspiring new approaches to this field of research is the predictive processing framework, which posits that in addition to the features of objective sensory information, our past experiences, knowledge, expectations and context all play a role in determining the qualitative and quantitative nature of the subjective percept. Importantly, the predictive processing framework offers new ways to approach individual differences research in vision. A new perspective on top-down influences on perception allows us to consider that individual differences can result not only from differences in simple basic perceptual abilities, but also from dissimilar past experiences and beliefs, or even because of a trait-like cognitive bias in the relative weighting of prior expectations versus sensory information.
The general aim of the current dissertation was to study the expression and structure of interindividual differences in top-down effects on visual perception. To answer this question, four published studies are introduced which contribute to ongoing work in the field. The role of predictions in influencing the subjective perceptual experience was demonstrated across many different perceptual tasks and processing levels. To explore whether there is a general factor for the effects of priors, as individuals may vary along a dimension of relative reliance on priors, a battery of perceptual tasks was compiled where prior effects of subjective perception had been demonstrated. Results revealed that individual variance in those tasks was better described by two factors which reflected the different hierarchical levels of the priors evoked. Therefore, there is no general factor of priors without regard to the types of priors evoked by each specific task. Another study demonstrated that illusory perception of an absent stimulus can be reliably evoked in a dual-task setup by repeatedly strengthening the expectation to see two visual stimuli presented simultaneously. The results suggest that illusory perception is quite common under certain conditions, but also displays rather large differences between individuals as well as between analogous tasks. Lastly, questionnaire measures of the autistic and schizotypal trait were used to investigate whether the predictive processing account of suboptimal prior precision in autism and schizophrenia would find support within the neurotypical population. Overall, it was concluded that the proposed hypotheses are likely too simplistic to capture the complexities of perceptual atypicalities in these spectrums.