Abstract
Several influential theories of consciousness attempt to explain how, when and where conscious perception arises in the brain. It is generally accepted that the transition to conscious perception requires feedback from higher level regions. However, the extent of this feedback is still debated. One theory holds that the transition to conscious perception requires feedback from frontoparietal regions (the Global Neuronal Workspace theory; Dehaene & Naccache, 2001), whereas the other maintains that feedback within sensory regions is sufficient to enable conscious perception (the Local Recurrence Theory; Lamme, 2006). Here, we combined a challenging discrimination task with EEG decoding techniques to arbitrate between these competing models of consciousness. Participants discriminated at-threshold masked face vs house stimuli and reported confidence in their discrimination performance. A classifier was trained and tested to discriminate correctly identified face vs house stimuli, separately for high- and low-confidence responses, using both time-specific (on-diagonal) and temporal generalization methods (off-diagonal). The results reveal two distinct decodable patterns of late neural activity. One pattern reflects perceptual confidence, indicating a global cognitive maintenance of the stimulus representation that is consciously reportable (late, on-diagonal decoding). The other pattern (late, off-diagonal decoding) shows maintenance of category information through recurrent processes within visual cortex, and is independent from confidence. This suggests that two functionally independent category-selective representations of a stimulus can temporally co-exist. We confirm the predicted temporal dynamics of how visual stimuli gain access to consciousness and reconcile two competing consciousness theories, by demonstrating for the first time that local recurrence within sensory regions and global maintenance processes in frontoparietal regions temporally co-occur during conscious perception, but that only the latter reflects perceptual confidence.
Meeting abstract presented at VSS 2018