Feedforward connectivity patterns from visual areas to the front of the brain contain information about sensory stimuli regardless of awareness or report.

Current theories of consciousness disagree on whether the prefrontal cortex (PFC) is necessary for conscious perception. No-report paradigms, which separate task reporting from conscious experience, have shown reduced PFC activity but have not tested the PFC's role in monitoring sensory cortex patterns. Reanalyzing EEG data from a no-report inattentional blindness paradigm, researchers used Granger causality and multivariate classifiers to examine feedforward input from visual areas to the PFC. Unexpectedly, information about the sensory stimulus was present in these input patterns regardless of whether the stimulus was consciously seen. This suggests that theories of consciousness must be revised, as the contents of consciousness are decodable from posterior-to-PFC inputs even without awareness or report.

Study at a glance

Abstract

Current theories of consciousness can be categorized to some extent by their predictions about the putative role of the prefrontal cortex (PFC) in conscious perception. One family of the theories proposes that the PFC is necessary for conscious perception. The other postulates that the PFC is not necessary and that other areas (e.g., posterior cortical areas) are more important for conscious perception. No-report paradigms could potentially arbitrate the debate as they disentangle task reporting from conscious perception. While previous no-report paradigms tend to point to a reduction in PFC activity, they have not examined the critical role of the PFC in "monitoring" or "reading out" the patterns of activity in the sensory cortex to generate conscious perception. To address this, we reanalysed electroencephalography (EEG) data from a no-report inattentional blindness paradigm (Shafto & Pitts, 2015). We examined the role of feedforward input patterns to the PFC from sensory cortices. We employed nonparametric spectral Granger causality and quantified the amount of information that reflected the contents of consciousness using multivariate classifiers. Unexpectedly, regardless of whether the stimulus was consciously seen or not, we found that information relating to the current sensory stimulus was present in the pattern of inputs from visual areas to the PFC. In light of these findings, we suggest various theories of consciousness need to be revised to accommodate the fact that the contents of consciousness are decodable from the input patterns from posterior sensory regions to the PFC, regardless of awareness (or report).