UvA collaborators: Cyriel Pennartz (Director, theory lead), Dr. Umberto Olcese (Co-director, experimental lead), Kwangjun Lee, MSc (theory researcher), Dr. Kengo Takahashi (experimental researcher), Reinder Dorman, MSc (Project manager). See our staff list for contact details.

The INTREPID project is an adversarial collaboration aiming to test the contrasting predictions of three prominent theories of consciousness: the integrated information theory (IIT, lead: Giulio Tononi) and two theories related to the predictive processing (PP) framework: neurorepresentationalism (PP-NREP, lead: Cyriel Pennartz) and active inference (PP-AI, lead: Karl Friston). The project is funded by the Templeton World Charity Foundation under the Accelerating Research of Consciousness (ARC) initiative.

The contrasting positions of the three theories lead to testable, general predictions. The principle of adversarial collaboration is borne out by a complementary hypothesis-testing across three experiments. Each experiment is designed in a way that allows to test incompatible predictions of at least two groups of theories (e.g. IIT vs. both PP-NREP and PP-AI). Together, the experimental results taken together should establish which theories of consciousness hold up best against the data.

INTREPID is led by the University of Amsterdam. The project consists of theoretical specialists, experimental groups and laboratories performing replicas of the experiments. In Amsterdam, at the Cognitive and Systems Neuroscience group, director and theoretical lead Prof. Dr. Cyriel Pennartz and PhD student Kwangjun Lee form the theoretical team for Neurorepresentationalism. Also from the Cognitive and Systems Neuroscience group, co-director and experimental lead Dr. Umberto Olcese together with Dr. Kengo Takahashi perform experiment 1 in Amsterdam.

The other two theory leads are Prof. Dr. Karl Friston from the University College London, and Prof. Dr. Giulio Tononi from University of Wisconsin-Madison. Institutes involved in the experiments are the University of Glasgow (experiment 2a; lead: Lars Muckli), University of Wisconsin-Madison (experiment 2b; lead: Melanie Boly) and Monash University (experiment 3; lead: Jakob Hohwy). The following institutions will aim to replicate the main experiments: Columbia University (replication experiment 1; lead: Rafael Yuste), York University Glendon College (replication 2a; lead: Patrick Cavanagh), Schepens Eye Research Institute from Harvard Medical School (replication 2b; lead: Eli Peli) and Friedrich Schiller University Jena (replication 3; lead: Gyula Kovacs). The project is supervised by an advisory board, consisting of Lucia Melloni, Anil Seth, Wolf Singer, Nao Tsuchiya, Nicholas Schiff and Andy Clark, providing insights and recommendations throughout the project. Lastly, Conrado Bosman (UvA), Chris Klink (Netherlands Institute for Neuroscience) and Simon van Gaal (UvA) are involved as members of the Data Replication and Analysis Committee (DARC), validating the processes of replication and analysis.

Integrated information theory

The integrated information theory claims that the quality of consciousness is identical to the cause-effect structure specified by a set of mechanisms (neural, or other) forming a maximum of integrated cause-effect power (Oizumi et al 2014, Tononi et al. 2016, Albantakis et al. 2023). It is an intrinsic, fundamental property of such a system, and is determined both by the nature of the causal mechanisms that compose it and by their state. IIT predicts that changes in the causal mechanisms that constitute the system at a relevant spatio-temporal scale should modify its cause-effect structure - even if there is no change in their state - and should thus have a perceptual correlate (i.e. a change in the quality of consciousness).

Predictive Processing – Active Inference

Active inference is a predictive processing theory concerned with inference of policies for action for expected (in contrast to current) prediction error minimisation (Friston et al. 2017, 2020, Friston 2018). In brief, active inference entails a covert or overt sampling of the sensorium to reduce uncertainty about the causes of sensations. Overt sampling in the visual domain mainly occurs through eye movements or other movement. Uncertainty reduction manifests as belief updating, which can be read as perception. Put simply, something can only be consciously ‘seen’ when ‘looked at’ or ‘noticed’.

Predictive Processing – Neurorepresentationalism

Set within the general Predictive Processing framework, Neurorepresentationalism postulates that perception arises from the construction of both high- and low-level inferential representations which can be simultaneously characterized as perceptual hypotheses; the continuous interaction with bottom-up sensory inputs provides for updating of generative models of the causes underlying changes in sensory input (Pennartz 2015, 2022; Pennartz et al. 2023). Under this account, high-level representations emerge from low-level representations by combining feature information coded by single neurons and small assemblies, into richer aggregates of information coded by unimodal (e.g. visual) networks and multimodal networks, giving rise to phenomenal experience at the highest level. It states that neural activity is considered essential for conscious experience, whereas motor activity is deemed important but not essential.

References

• Albantakis, L., Barbosa, L., Findlay, G., Grasso, M., Haun, A. M., Marshall, W., ... & Tononi, G. (2022). Integrated information theory (IIT) 4.0: formulating the properties of phenomenal existence in physical terms. arXiv preprint arXiv:2212.14787.
• Friston K. (2018) Am I Self-Conscious? (Or Does Self-Organization Entail Self-Consciousness?). Front Psychol 9: 579.
• Friston K, FitzGerald T, Rigoli F, Schwartenbeck P, Pezzulo G. (2017). Active Inference: A Process Theory. Neural Comput29: 1–49, 2017.
• Friston KJ, Wiese W, Hobson JA. (2020) Sentience and the origins of consciousness: From Cartesian duality to Markovian monism. Entropy 22: 516, 2020.
• Pennartz CMA (2015) The Brain’s Representational Power – on consciousness and the integration of modalities. MIT press (382 pp.). ISBN: 9780262029315.
• Pennartz CMA (2022) What is neurorepresentationalism? From neural activity and predictive processing to multi-level representations and consciousness. Behavioral Brain Research 432:113969.
• Pennartz CMA, Oude Lohuis MN, Olcese, U. (2023) How ‘visual’ is the visual cortex? Interactions between the visual cortex and other sensory, motivational and motor systems to enable vision itself. Proc. Trans. Royal Soc. B. 378: 20220336. DOI: 10.1098/rstb.2022.0336.
• Oizumi M, Albantakis L, Tononi G. (2014). From the phenomenology to the mechanisms of consciousness: Integrated Information Theory 3.0. PLoS Comput Biol 10: e1003588.
• Tononi, G., Boly, M., Massimini, M., & Koch, C. (2016). Integrated information theory: from consciousness to its physical substrate. Nature Reviews Neuroscience, 17(7), 450-461.