A case for graded, early neural components as the most predictive correlates of consciousness

Invited talk by Kristian Sandberg, Institute of Cognitive Neuroscience, UCL

It is often argued that consciousness is an all-or-none phenomenon, and particularly neuronal global workspace theory predicts that consciousness is related to sudden “neural ignition” involving frontal areas at a relatively late time of processing (around 350-550ms after stimulus onset). In this talk, I provide behavioural evidence that consciousness is best viewed as a graded phenomenon, and I provide neuroscientific evidence from MEG studies that graded, relatively early neural components (130-320ms after stimulus onset) localised to mainly occipital/temporal areas of the cortex are the most consistent predictors of visual consciousness. These components can be used to decode conscious perception of specific stimuli, the presence of absence of conscious experiences on single trials, and graded experiences. The components also predict gradual developments in conscious perception across tens of seconds and reversals of ambiguous perception. They even generalize across years and are somewhat similar across individuals. Overall, I argue that we should move away from viewing late, frontal activity as a general neural signature of consciousness and instead view a given conscious experience as a specific, amplified pattern of activity extending across sensory areas.