Source
Cogitate Consortium, Ferrante, Gorska-Klimowska, Henin et al. (2025), Nature 642:133–142, LIT-973, read in NOTE-812: “Interareal connectivity”, Discussion, Methods (“Functional connectivity analysis”), Extended Data Figs. 5–7.
What was actually shown
The prediction. Preregistered with IIT’s proponents: “sustained gamma-band connectivity within the posterior cortex, that is, between high-level and low-level sensory areas (V1/V2), throughout any conscious visual experience”. IIT “would be challenged in the absence of sustained content-specific synchronization between face–object selective areas and V1/V2”. It was a critical prediction.
The test. Pairwise phase consistency between face- or object-selective sites and V1/V2, on 1.0-s and 1.5-s task-irrelevant trials. In iEEG it was computed electrode to electrode within patients, and in MEG with GED-filtered fusiform sources.
The result.
- iEEG: content-selective synchrony with V1/V2 was “early and brief (for example, less than 0.75 s), and restricted to low frequencies (2–25 Hz)”, and “mostly explained by the stimulus-evoked response”. It disappeared when the evoked response was removed. Bayes factors favoured the null (BF01 = 1.15–4.9).
- MEG: the same pattern, with gamma-band BF01 > 3.
- Exploratory DFC (amplitude co-modulation, evoked response removed): object-selective coupling with V1/V2 that included gamma but “was brief”. Face-selective coupling was “scarce” (BF01 = 1.3).
- fMRI gPPI: FFA coupled face-selectively with V1/V2, a result of the kind both theories allow, with no time course.
Beside it. The same study found sustained category information and duration-tracking activity in posterior cortex (THEORY-102’s pattern). So content was held over time while content-specific coupling to early visual cortex was not. The consortium calls the coupling result “the most direct challenge” to IIT, being “incompatible with IIT’s claim that the state of the neural network, including its activity and connectivity, specifies the degree and content of consciousness”.
What could have come out otherwise: gamma phase consistency could have stayed above its across-category baseline for the whole image, as predicted.
What this does not say
- It does not test integrated information. Nothing computed Φ. The prediction belongs to IIT’s proposed implementation in posterior cortex, which the authors say a theory may revise while keeping its core.
- It does not exclude sustained coupling under denser sampling. Only 12 iEEG electrodes were in V1/V2, against 472 in prefrontal cortex. The authors grant the null “could stem from methodological limitations (for example, limited iEEG sampling of V1/V2 areas)”.
- It does not say posterior cortex does not sustain content. It did, for category and object identity. The finding is about coupling between areas, not about content within one.
- It does not favour the workspace by default. The preregistered PPC found no content-selective coupling with prefrontal cortex either. Exploratory DFC found brief prefrontal gamma coupling, and that is not the preregistered test.
- It is about clearly seen, unchanging images over at most 1.5 s.
Connections
- THEORY-098 puts the basis of perceptual consciousness in recurrent activity in sensory cortex. This result does not refute it, since recurrence need not show as sustained gamma phase-locking with V1/V2. But it removes one measurable form that sustained posterior recurrence was predicted to take.
- THEORY-102 and THEORY-245 hold what the same paradigm showed about content: sustained posterior category coding, and prefrontal content at onset only and without identity or orientation. Read together, sustained content in posterior cortex came without sustained content-specific coupling to early visual cortex.
- QUESTION-048. As a theory test, it shows how a design aimed at a
theory reaches only its implementation. That is the paper’s own point,
and it is why no
answers:is declared.