Published in Nature, April 30, 2025
Conscious content should be decodable from posterior cortex only. The prefrontal cortex is not necessary for conscious experience — it's involved in reporting and acting on experience, not having it.
Orientation information was ONLY decodable from posterior cortex. Prefrontal showed chance-level performance with strong Bayesian evidence for null.
Conscious content requires prefrontal involvement. All dimensions of conscious content should be represented in the global workspace, including both category and orientation.
Category (faces vs objects) was decodable from prefrontal, but orientation was NOT. This challenges GNWT's claim that prefrontal represents ALL conscious content.
Sustained activity throughout conscious perception. Gamma-band synchronization should persist within posterior cortex for the duration of the conscious experience.
NO sustained gamma-band synchronization found in posterior cortex. Activity was early, brief, and restricted to low frequencies — the opposite of IIT's prediction.
Brief "ignition" bursts at stimulus onset AND offset. The global workspace should fire when content enters and when it leaves consciousness.
Onset ignition appeared as expected, but NO offset ignition! Only 1 of 99 prefrontal electrodes showed any offset response — and it occurred earlier than predicted.
Sustained gamma synchronization within posterior regions. Consciousness requires ongoing integration within the posterior "hot zone" — not connectivity to frontal areas.
Connectivity was early and brief, restricted to low frequencies. No sustained gamma synchronization in posterior cortex was observed.
Connectivity between prefrontal and sensory areas. The global workspace broadcasts information from sensory regions to the whole brain via long-range connections.
Some connectivity observed but not in the pattern predicted. The expected frontoparietal network activation was inconsistent across modalities.
Posterior cortex finding supports IIT, but temporal predictions failed. No sustained gamma synchronization — a core requirement of the theory.
Prefrontal can't decode orientation — contradicting global workspace claims. No offset ignition — GNWT's temporal signature missing.
Both theories faced critical empirical challenges. The result is not a victory for either camp but a call for theoretical revision.
Despite no clear winner, the adversarial framework produced genuine knowledge: specific predictions were tested and some failed.
Both IIT and GNWT will need to account for these findings. The data is now public for testing other theories.
The hard problem of consciousness — why physical processes give rise to subjective experience — remains "as baffling as ever."