Interacting with consciousness: An investigation into the neural signatures of conscious processing using the global-local auditory task

Can conscious processing be inferred from electrophysiological measurements?

Bekinschtein et al. (2009) devised the global-local auditory oddball task as a way of investigating electrophysiological (EEG) responses to hierarchical violations in auditory regularity. It was argued that the detection of local violations in auditory regularity (arising within a short temporal window) occurs independently of consciousness, whilst the detection of global violations in auditory regularity (arising within a longer temporal window) occurs only with the presence of consciousness. For this to be the case, global and local effects must be assumed to be independent and thus share an additive relationship. Crucially, if a relationship is additive, then cognitive subtraction allows each effect to be understood in isolation. Within this thesis, we explore the notion that global and local effects may not be independent but actually share a multiplicative relationship, based on the principle that the brain is a non-linear system (Friston et al., 1996). We examine this by re-analysing the data of the original attention study by Bekinschtein et al. (2009) using a factorial design analysis. What is more, we extend this work to consider the presence of an interaction between global and local effects when consciousness is manipulated using healthy sedation. Our findings reveal an interaction between global and local effects that occurs within the region of the global effect and varies depending on the presence of direct attention, expectancy and sedation. The manifestation of an interaction between global and local effects is discussed in detail in relation to a predictive coding framework, whereby multiple levels of prediction exist within the brain (Auksztulewicz & Friston, 2015). Furthered analysis, directly comparing the findings of varied attention to the findings with healthy sedation is presented and discussed with a view to the brain's predictive nature. Methodological issues that were encountered, as well as a method we developed for re-aligning time-series data, are presented and explored in detail within the later Chapters of this thesis. Our research suggests that global and local effects may not be considered as independent and therefore the presence of a global effect is not sufficient to constitute a functionally isolated marker of conscious processing (as initially proposed by Bekinschtein et al. [2009]). Further research is necessary to better define the nature of an interaction between global and local effects in relation to conscious experience.