Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 65 of the 68 references it cites were found in a registry or an index; 9 of the 16 sources its claims cite were checked for support.

An adversarial collaboration protocol for testing contrasting predictions of global neuronal workspace and integrated information theory

Open access

Abstract

The relationship between conscious experience and brain activity has intrigued scientists and philosophers for centuries. In the last decades, several theories have suggested different accounts for these relationships. These theories have developed in parallel, with little to no cross-talk among them. To advance research on consciousness, we established an adversarial collaboration between proponents of two of the major theories in the field, Global Neuronal Workspace and Integrated Information Theory. Together, we devised and preregistered two experiments that test contrasting predictions of these theories concerning the location and timing of correlates of visual consciousness, which have been endorsed by the theories' proponents. Predicted outcomes should either support, refute, or challenge these theories. Six theory-impartial laboratories will follow the study protocol specified here, using three complementary methods: Functional Magnetic Resonance Imaging (fMRI), Magneto-Electroencephalography (M-EEG), and intracranial electroencephalography (iEEG). The study protocol will include built-in replications, both between labs and within datasets. Through this ambitious undertaking, we hope to provide decisive evidence in favor or against the two theories and clarify the footprints of conscious visual perception in the human brain, while also providing an innovative model of large-scale, collaborative, and open science practice.

Overview

CoChat Summary

TL;DR

This paper presents a preregistered adversarial collaboration protocol between Global Neuronal Workspace (GNW) and Integrated Information Theory (IIT) to test contrasting predictions about the location and timing of visual consciousness. It reports no empirical outcomes yet; it specifies two experiments, three recording methods, built-in replications, and planned analyses across six theory-neutral laboratories.

Key findings

  • The protocol contrasts five predictions: location of the neural correlates of consciousness (GNW: fronto-parietal/PFC; IIT: posterior hot zone), decodability of conscious content, temporal dynamics (GNW: transient ignition at onset/offset; IIT: sustained substrate during experience), pre-stimulus activity, and functional connectivity/synchronization patterns.
  • Two experiments are planned: Experiment 1 uses clearly visible stimuli and manipulates task relevance, stimulus duration (500, 1000, 1500 ms), category (faces, objects, letters, false-fonts), and orientation; Experiment 2 uses attentional manipulation/inattentional blindness to compare seen versus unseen faces and objects in distracted-attention (dAT) and attended (AT) tasks.
  • Planned sample sizes are a minimum of 50 subjects per experiment for fMRI and M-EEG, and 25 subjects per experiment for iEEG, for a total of 250 collected datasets in each experiment; the Discussion states the project aims to collect 500 datasets (250 subjects × 2 studies).
  • Data will be split in half: phase 2 uses the first half to optimize analyses, then protocols are locked; phase 3 uses the second half for replication. Both classical and Bayesian statistics are planned.

Methods

  • Adversarial collaboration between proponents of GNW and IIT; theory-neutral teams acquire and analyze data, with six data acquisition laboratories using fMRI, M-EEG, and iEEG.
  • Experiment 1 is a nested factorial design with factors stimulus category, task relevance, stimulus duration, and stimulus orientation; trials are 2 s, stimuli are supra-threshold at fixation (6°×6°), and gaze is monitored. Experiment 2 is a nested factorial design with factors stimulus category (faces, objects), location (left, right), and task (dAT, AT); visibility (seen/unseen) is measured via reports in dAT, with 2.3° high-contrast stimuli presented for 250 ms at four locations during a video game.
  • Planned analyses include multivariate decoding of category/orientation or category/location, activation patterns across stimulus durations, inter-areal synchronization, and putative NCC contrast/conjunction analyses; analysis details are to be optimized in phase 2 and locked before phase 3 replication.

Limitations

  • The paper states that GNW and IIT make fundamental assumptions that are challenging to test at the relevant granularity, and that “it is possible that no unambiguous answer may emerge from these experiments”; at minimum, results are expected to challenge specific aspects of either or both theories.
  • The paper states that the preregistration leaves open specification of analysis methods, so the current protocol includes a higher-level description of analyses but not all analysis details.
  • Evident from the text: no empirical results are reported because this is a study protocol; iEEG sampling is opportunistic, no behavioral pre-screening is applied to patients, and patient behavior is noted as likely to differ from neurotypical participants.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1371/journal.pone.0268577

Cites

65 of 65 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig 1 Core predictions, analyses, and potential results for the two experiments. For each theory (white circles), the core predictions (white rectangles) and possible outcomes (colored rectangles) are summarized. Green squares denote positive outcomes, red ones denote negative outcomes, and yellow squares denote negative outcomes that are less conclusive, as they might stem from limitations of the methods or analyses. On the right, we conclude how combinations of results would reflect on the theories. Note that some predictions require time-resolved measures (M-EEG, iEEG), while others can be tested with all three techniques. PFC refers to Prefrontal cortex; RSA refers to representational similarity analysis [RSA; 48]; IT refers to inferotemporal cortex; LOC refers to lateral occipital cortex; V1 refers to primary visual cortex.
Fig 1 Core predictions, analyses, and potential results for the two experiments. For each theory (white circles), the core predictions (white rectangles) and possible outcomes (colored rectangles) are summarized. Green squares denote positive outcomes, red ones denote negative outcomes, and yellow squares denote negative outcomes that are less conclusive, as they might stem from limitations of the methods or analyses. On the right, we conclude how combinations of results would reflect on the theories. Note that some predictions require time-resolved measures (M-EEG, iEEG), while others can be tested with all three techniques. PFC refers to Prefrontal cortex; RSA refers to representational similarity analysis [RSA; 48]; IT refers to inferotemporal cortex; LOC refers to lateral occipital cortex; V1 refers to primary visual cortex.CC BY · PubMed Central
Fig 2 Experiment 1 design—In separate blocks (top & bottom rows), subjects will be asked to detect infrequent targets drawn from two categories: a specific face and a specific object (upper row), or a specific letter and a specific false-font (lower row). The stimuli will be shared across blocks, but the selection of the targets will determine to which of three trial types each stimulus belongs: task-relevant targets, task-relevant non-targets, task-irrelevant non-targets highlighted by red, blue and green frames respectively. Colored frames are shown for illustration purposes only. The duration (500, 1000, 1500 ms) and orientation (side view/front view) of stimuli will be manipulated to allow for specific analyses. Blank intervals between stimuli and intertrial intervals (truncated exponential distribution) are not depicted here.
Fig 2 Experiment 1 design—In separate blocks (top & bottom rows), subjects will be asked to detect infrequent targets drawn from two categories: a specific face and a specific object (upper row), or a specific letter and a specific false-font (lower row). The stimuli will be shared across blocks, but the selection of the targets will determine to which of three trial types each stimulus belongs: task-relevant targets, task-relevant non-targets, task-irrelevant non-targets highlighted by red, blue and green frames respectively. Colored frames are shown for illustration purposes only. The duration (500, 1000, 1500 ms) and orientation (side view/front view) of stimuli will be manipulated to allow for specific analyses. Blank intervals between stimuli and intertrial intervals (truncated exponential distribution) are not depicted here.CC BY · PubMed Central
Fig 3 Experiment 2 design–A) The four worlds of the game, with four levels in each world (left). In worlds 1 & 3 (middle) the player will control a shiny blue orb at the bottom of the screen to collect falling disks of the same color (on three vertical tracks), while avoiding falling disks of a different color. The size of the full game display in visual angle is indicated here (panel A, middle). In worlds 2 & 4 the colors will be swapped and players will control an orange orb (right). B) Examples of stimuli presented in the background (central portion of screen blown-up for display purposes): face (left), object (middle), blank (right). C) During gameplay, at 3-6sec intervals (4-7sec for fMRI), an object/face/blank will appear on one of four background squares (middle) for 500ms total, 250ms at full contrast, followed by 250ms of fade-out. On some trials (every 9-18sec), the stimulus will be immediately followed by an awareness probe (right) in which the game will pause and a small arrow will appear in the center instructing subjects to report whether they had just noticed a stimulus in that location. D) Schematic of the background animation demonstrating the trial timing, stimulus timing, and probe timing. For a video demonstration of the video game, visit the preregistration website: https://osf.io/b9nce/).
Fig 3 Experiment 2 design–A) The four worlds of the game, with four levels in each world (left). In worlds 1 & 3 (middle) the player will control a shiny blue orb at the bottom of the screen to collect falling disks of the same color (on three vertical tracks), while avoiding falling disks of a different color. The size of the full game display in visual angle is indicated here (panel A, middle). In worlds 2 & 4 the colors will be swapped and players will control an orange orb (right). B) Examples of stimuli presented in the background (central portion of screen blown-up for display purposes): face (left), object (middle), blank (right). C) During gameplay, at 3-6sec intervals (4-7sec for fMRI), an object/face/blank will appear on one of four background squares (middle) for 500ms total, 250ms at full contrast, followed by 250ms of fade-out. On some trials (every 9-18sec), the stimulus will be immediately followed by an awareness probe (right) in which the game will pause and a small arrow will appear in the center instructing subjects to report whether they had just noticed a stimulus in that location. D) Schematic of the background animation demonstrating the trial timing, stimulus timing, and probe timing. For a video demonstration of the video game, visit the preregistration website: https://osf.io/b9nce/).CC BY · PubMed Central
Fig 4 Timeline for the adversarial collaboration protocol. Number of subjects refers to the sum of subjects collected in the two experiments. Due to the COVID-19 pandemic of 2020–2021, phase 1 and 2 will be carried out in parallel and adjustments to the timing of some of the phases may be necessary.
Fig 4 Timeline for the adversarial collaboration protocol. Number of subjects refers to the sum of subjects collected in the two experiments. Due to the COVID-19 pandemic of 2020–2021, phase 1 and 2 will be carried out in parallel and adjustments to the timing of some of the phases may be necessary.CC BY · PubMed Central

Verification

Claims the paper makes

7 Valid · 3 Not valid · 2 Not checked · 58 Own results and interpretations

9 more claims

Create a free account to see all 70 claims and every source verdict.

See all claims — free account
  • Probably not a substantial claim. 49
  • It may state two separate claims as one. 17

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

1 of 1

Citations

Cites

Show all 68 references

Research works printed but not linked: 2

  1. G. Tononi, 2013, Integrated information theory of consciousness: an updated account, Archives Italiennes de Biologie, 150, 290
  2. A. Mack, 1998, Bradford Books series in cognitive psychology

Cited by

See all 114

Metadata

Authors
Melloni, L., Mudrik, L., Pitts, M., Bendtz, K., Ferrante, O., Gorska, U., Hirschhorn, R., Khalaf, A., Kozma, C., Lepauvre, A., Liu, L., Mazumder, D., Richter, D., Zhou, H., Blumenfeld, H., Boly, M., Chalmers, D. J., Devore, S., Fallon, F., de Lange, F. P., Jensen, O., Kreiman, G., Luo, H., Panagiotaropoulos, T. I., Dehaene, S., Koch, C., & Tononi, G.
Affiliations
  • Tel Aviv University (IL)
  • Reed College (US)
  • Harvard University (US)
  • University of Birmingham (GB)
  • Cairo University (EG)
  • University of Wisconsin–Madison (US)
  • Max Planck Institute for Empirical Aesthetics (DE)
  • Peking University (CN)
  • Radboud University Nijmegen (NL)
  • Inserm (FR)
  • New York University (US)
  • St. John's University (US)
Venue
PLoS ONE
Published
2023-02-10
Volume
18
Issue
2
Pages
e0268577
ISSN
1932-6203 (ISSN-L)
OpenAlex
W4319825978
Cited by
113
References
68
Citations by year
2026: 30 · 2025: 33 · 2024: 30 · 2023: 20
FWCI
29.12
Type
journal article
Language
en
Open access
Locations
  • PLoS ONE (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Radboud Repository (Radboud University) (submittedVersion)
  • HAL (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • University of Birmingham Research Portal (University of Birmingham) (submittedVersion, cc-by)
  • Radboud Repository (Radboud University) (submittedVersion, other-oa)
Funders
Templeton World Charity Foundation, Inc (TWCF0389, TWCF0345, TWCF0378)
Topics
  • Social Sciences › Arts and Humanities › Philosophy › Philosophy and Theoretical Science
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Visual perception and processing mechanisms
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Neural dynamics and brain function
Keywords
global neuronal workspace theory, integrated information theory, visual awareness, open science, adversarial collaboration, preregistration, replication, visual perception
MeSH terms
Brain, Consciousness, Electroencephalography, Humans, Information Theory, Visual Perception

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-09-28, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-09-28 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: An adversarial collaboration protocol for testing contrasting predictions of global neuronal workspace and integrated information theory. https://app.cochat.ai/papers/10.1371/journal.pone.0268577
The paper
Melloni, L., Mudrik, L., Pitts, M., Bendtz, K., Ferrante, O., Gorska, U., Hirschhorn, R., Khalaf, A., Kozma, C., Lepauvre, A., Liu, L., Mazumder, D., Richter, D., Zhou, H., Blumenfeld, H., Boly, M., Chalmers, D. J., Devore, S., Fallon, F., . . . Tononi, G. (2023). An adversarial collaboration protocol for testing contrasting predictions of global neuronal workspace and integrated information theory. PLoS ONE. https://doi.org/10.1371/journal.pone.0268577

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.