Abstract
Neural correlates of perceptual awareness have been difficult to distinguish from correlates of (1) unconscious processing of physical stimulus properties, and (2) task-related cognition involved in generating reports about one’s perception. To isolate awareness-specific neural activity from unconscious sensory encoding, Del Cul et al. (2007) used backward masking to manipulate a physical property of stimuli in a linear fashion. Stimulus onset asynchrony (SOA) between targets and subsequent masks was systematically varied such that each target fell below, at, or above the perceptual threshold. Participants reported after each trial whether or not they saw the target while scalp-EEG was measured. Amplitudes of early sensory responses (P1/N1) scaled linearly with the SOA manipulation, while the P3b displayed a bifurcated pattern of amplitude modulation corresponding closely with the seen/unseen reports. The authors concluded that the P3b is most likely a correlate of conscious perception, while earlier (P1/N1) sensory responses reflect unconscious accumulation of sensory evidence. Here, we designed a new experiment similar to Del Cul et al. (2007), adding a novel ‘no-report’ condition to circumvent the task-related confounds inherent in purely report-based paradigms. In the report condition, our results closely replicated Del Cul et al. (2007): a linear increase of P1 amplitudes tracked the SOA manipulation regardless of whether the targets were consciously perceived, and P3b amplitudes showed a bifurcated pattern corresponding well with participants’ visibility reports. In the no-report condition, however, while the linear P1 modulations were the same as in the report condition, the P3b and its bifurcation dynamics completely disappeared. Eliminating the P3b revealed that some mid-latency ERPs showed signs of bifurcation dynamics, suggesting that they might correlate with perceptual awareness. These results may help improve and refine contemporary models of conscious perception such as the Global Neuronal Workspace.