Research Project
Full description To generate faithful representations of the external world necessary to guide behaviour, the brain must combine sensory information with prior knowledge about the environment. Accordingly, there is ample convergent evidence that the outcomes of perceptual decisions and their associated actions can be biased via the manipulation of expectations. A common view is that the magnitude of these biases is flexibly weighted by the reliability of available sensory information such that expectations guide perceptual decisions more when the available sensory data are sparse or imprecise (de Lange et al., 2018; Summerfield & de Lange, 2014). Prominent 'predictive coding' theories of perception argue that prior expectations not only shape the decisions and behaviours that result from perceived stimuli, but also the neural representations of the sensory signals themselves. Recent evidence from magnetic resonance imaging (MRI) and electroencephalography (EEG) has broadly supported this view (Kok et al., 2012, 2017; Tang et al., 2018, 2023; Yon et al., 2019), but findings differ regarding how sensory representations are influenced by prior experience. Under so-called 'sharpening' accounts, the responses of neurons preferentially tuned to unexpected stimuli are suppressed, leading to an attenuated but sharper population response to expected stimuli (Kok et al., 2012; Yon et al., 2018). In contrast, 'dampening' accounts propose that it is the neurons tuned to the expected stimulus that are most suppressed leading to a diminished, but less selective, population response (Richter et al., 2022). A recent review has suggested that opposing mechanisms might occur over the time course of the population response. In such opposing process models the brain initially biases perception toward expected events (sharpening), before boosting the gain for new information that is different to what is expected (Press et al., 2020). Other investigators have questioned whether current evidence can be interpreted as supporting expectation effects in the first place (Feuerriegel et al., 2021). In this project we aim to collect behavioural data and electrophysiological recordings of brain activity while human volunteers perform a perceptual decision-making task in which prior expectations and stimulus reliability will be systematically manipulated.