Adversarial testing of global neuronal workspace and integrated information theories of consciousness

Cogitate Consortium, Oscar Ferrante, Urszula Gorska-Klimowska, Simon Henin, Rony Hirschhorn, Aya Khalaf, Alex Lepauvre, Ling Liu, David Richter, Yamil Vidal, Niccolò Bonacchi, Tanya Brown, Praveen Sripad, Marcelo Armendariz, Katarina Bendtz, Tara Ghafari, Dorottya Hetenyi, Jay Jeschke, Csaba Kozma, David R. Mazumder, Stephanie Montenegro, Alia Seedat, Abdelrahman Sharafeldin, Shujun Yang, Sylvain Baillet, David J. Chalmers, Radoslaw M. Cichy, Francis Fallon, Theofanis I. Panagiotaropoulos, Hal Blumenfeld, Floris P. de Lange, Sasha Devore, Ole Jensen, Gabriel Kreiman, Huan Luo, Melanie Boly, Stanislas Dehaene, Christof Koch, Giulio Tononi, Michael Pitts, Liad Mudrik, Lucia Melloni · 2025 · View original paper

← Back to overview

Evidence (3)

State Transitions # Continue PAPER_TPL BIO

Predicted prefrontal ignition at stimulus onset/offset was tested; results showed transient onset responses but no offset ignition in PFC.

"In PFC ROIs, cross-temporal RSA revealed transient face–object categorical representation at stimulus onset, but not at stimulus offset. Consequently, no significant correlation was found with the GNWT onset and offset model (compare Fig. 3d, right, with the predicted pattern in Fig. 1a)."

Maintenance of conscious content, p. 139

Transient onset but absent offset responses in PFC argue against GNWT’s predicted ignition-based state transitions at both onset and offset in human data, constraining theories that tie conscious updating to abrupt PFC 'ignition' events.

"For GNWT, the most substantial challenge based on our preregistered criteria pertains to its account for the maintenance of a conscious percept over time and, in particular, the lack of ignition at stimulus offset."

Discussion, p. 141

The authors explicitly interpret the absence of offset ignition as a major challenge for GNWT’s state-transition account of consciousness, reinforcing the results-based constraint on ignition-like dynamics.

Figures

Fig. 1 (p. 135) : Fig. 1 formalizes GNWT’s predicted onset/offset ignitions; paired with the Results, the absent offset ignition challenges the predicted state transitions in PFC.

Limitations: Offset-related responses can be weaker and harder to detect; sampling of PFC with iEEG is sparse and MEG source localization may limit sensitivity to brief events.

Representational Structure # Continue PAPER_TPL BIO

Cross-temporal RSA showed sustained category/identity structure in posterior cortex, but only transient category structure in PFC.

"In posterior cortex ROIs, cross-temporal RSA revealed sustained face–object categorical representation, with larger correlation distances between categories (face–objects) than within category (face, object; compare Fig. 3d, left, with Fig. 1a). The RSA matrix matched the IIT model better than the GNWT model (see section 6.3 in Supplementary Information for results of all contrasts)."

Maintenance of conscious content, p. 139

Sustained representational geometry in posterior cortex supports stable content coding over time, speaking to how conscious content may be organized in persistent neural subspaces.

"In PFC ROIs, cross-temporal RSA revealed transient face–object categorical representation at stimulus onset, but not at stimulus offset."

Maintenance of conscious content, p. 139

PFC representations were brief and not sustained, suggesting limited or transient access to category-level content relative to posterior regions, which refines accounts of where and how conscious content is structured.

Limitations: Effects were strongest in iEEG; MEG results for some content dimensions were weaker or inconclusive, and only a subset of electrodes showed both category selectivity and duration tracking.

Temporal Coordination # Continue PAPER_TPL BIO

Interareal synchrony analyses found early, brief, low-frequency effects and limited gamma-band phase locking, challenging predictions of sustained posterior gamma or late phasic PFC coupling.

"We found increased category-selective synchrony between category-selective and V1/V2 electrodes... These effects were early and brief (for example, less than 0.75 s), and restricted to low frequencies (2–25 Hz). This synchrony was mostly explained by the stimulus-evoked response... These results fail to align with IIT’s predictions: the activity was neither sustained nor observed in the gamma frequency band... No content-selective PPC was found between face-selective and object-selective electrodes and PFC electrodes in the relevant time window, contradicting the prediction of GNWT."

Interareal connectivity, p. 141

Observed phase-locking dynamics were brief and low-frequency, not the sustained posterior gamma or late PFC phasic coupling expected by IIT/GNWT, refining candidate temporal coordination mechanisms for conscious processing.

Limitations: Phase estimates from macroscopic recordings are noise-sensitive; PPC may miss amplitude-based coordination, partially addressed with DFC but still subject to source and coverage constraints.