Three leading theories currently exist on how different brain structures and cell activity can give rise to consciousness: Neurorepresentationalism, Active Inference and the Integrated Information Theory. The goal of the project is to perform a set of experiments to test the different predictions of these three major theories. The project is part of a series of initiatives organized by the Templeton World Charity Foundation to advance our understanding of the neural basis of consciousness, by bringing together theoreticians and experimentalists in a collaborative framework. The Foundation has awarded 3.6M dollar to allow the researchers to perform, over three years, a set of three challenging experiments.

An adversarial approach to move our understanding of consciousness forward

A key aspect of the project will be its adversarial approach: the researchers will perform experiments that aim to test incompatible predictions of the three theories. Dr Umberto Olcese: ‘with this project we expect to provide evidence supporting one theory and at the same time challenging the other two, thus advancing the whole field of consciousness and not only a specific theoretical framework.’ Professor Cyriel Pennartz will be the project director and will act as theory lead for Neurorepresentationalism. Active Inference will be represented by Prof. Karl Friston (University College London, UK) and the Integrated Information Theory by Prof. Giulio Tononi (University of Wisconsin – Madison USA). The theory leads originally developed the three theories. In this combined project they will monitor progress together and have committed to accepting the experimental results, irrespective of whether these results support or challenge their own theory.

Three incompatible theoretical frameworks

The theories that will be tested in this project present three different accounts to explain how the brain generates consciousness. A short explanation of the different theories:

• Neurorepresentationalism is a theory of consciousness based on predictive processing, which is itself a major general framework to interpret brain function. Neurorepresentationalism postulates that perception arises from the construction of both high- and low-level inferential representations across hierarchical brain networks.
• Active Inference is also based on predictive processing, but, in contrast with neurorepresentationalism, states that overt or covert action (such as eye movements or attention, respectively) is necessary for conscious perception.
• Finally, the Integrated Information Theory claims that the consciousness is an intrinsic property of a computational system (such as the brain) that stems from the ability to process information in a way that is at the same time integrated and differentiated.

A broad international collaboration to study consciousness

One of the experiments will be coordinated by Dr Umberto Olcese, at the Cognitive and Systems Neuroscience Group at SILS-UvA, who also acts as co-director for the whole project. This experiment will study whether inactive (non-spiking) neurons play a role in the generation of conscious experience, as predicted by the Integrated Information Theory but not by the other two theories. The other experiments will be led by researchers in Melbourne (Australia), Glasgow (UK) and Madison (USA). A key feature of the “Accelerating Research of Consciousness” program is that each experiment will be simultaneously replicated in other laboratories, in this case in the USA, Canada and Germany. Thus, the adversarial studies on consciousness will be carried out across the globe.

More information on the Accelerating Research on Consciousness initiative from the Templeton World Charity Foundation can be found here.

Cognitive and System Neuroscienc Research Group