Investigating neural structures and behavioural biases in perceptual decision-making

Why do we behave as we do? The general mechanisms underlying visual perception and decision-making within a social context have been under scientific scrutiny for over half a century (Sherif, 1945; Asch, 1951; Berns et al, 2005). In spite of this, neither a definitive mechanism for how perceptual biases arise, nor a robust neural basis have emerged. In the first part of my thesis, I use a combination of visual behavioural testing and computational modelling to investigate the development of perceptual biases under social advice in children, and explore their potential mechanisms. In the second part, I use structural and functional magnetic resonance imaging in adults to examine the organisation of a cortical region that has been implicated in such visual perceptual decisions. First, I use a two-alternative forced choice visual task to probe the extent to which typically-developing children (between 6 and 14 years) conform to social advice, when making judgments about ambiguous and unambiguous structure-from-motion stimuli. Perceptual bias reported by typically-developing children with a single advisor was largely absent in the youngest children, but increases with age. When reaction time data were analysed with a drift diffusion model, the results suggested that the conforming bias of the older children was likely the product of a perceptual, not simply a decision, bias. The typically-developing children were then compared with age and IQ-matched children with autism. The autistic group showed a small conforming bias in decisions for the youngest children, while the effect of social advice decreased with age. Drift diffusion modelling showed no evidence of a perceptual bias, but rather a potential bias at later stages of processing, in the formation of the judgment. Second, I investigate how well visual area hMT+ can be mapped and characterised in individual human subjects. Visual cortical area hMT+ in humans is homologous to a macaque monkey region associated with perceptual decisions about the stimuli used in our visual psychophysics experiment with children (Krug 2004; Krug, 2013). By comparing two types of structural MRI scan sensitive to myelination (MP2RAGE at 7T and T1w/T2w at 3T) with functional localisers for hMT+, I found a consistent association of MRI contrasts indicating dense myelin with the functional localisers for group averaged datasets. However, regional patterns of myelination, and their association with functional hMT+, varied considerably between individual subjects. I went on to examine the internal functional organisation of area hMT+ with a high-resolution 7T fMRI scan. These scans provided evidence that in humans as in the monkey, cortical clusters respond selectively to specific stimulus aspects, such as specific directions of motion, or 3D depth. These behavioural experiments in conjunction with computational modelling suggest that advice given by others can bias perceptual decisions about visual stimuli. High field MRI of a cortical region implicated in these perceptual processes, visual area hMT+, indicate a columnar architecture that can be revealed using current technology. Combining the two approaches should allow us directly to probe the neural basis of perceptual changes under social influence in humans.