What makes us consciously aware of some experiences but not others?
What makes some brain processes conscious and others subconscious? And how come, even when we are awake and attentive, our awareness can shift in ways that are independent of other aspects of mental function? Like when you are driving and your thoughts drift elsewhere, yet you are still perfectly capable of navigating traffic and handling the vehicle.
While scientists have made significant progress understanding specific brain functions like vision, memory, and decision-making – what philosopher and cognitive scientist David Chalmers refers to as the “easy problems” – the fundamental nature of conscious experience, the “hard problem” has been elusive.
The nature of consciousness - how the brain generates our subjective experience of the world - remains one of science's greatest mysteries and humanity’s greatest quests. And, while modern researchers have identified neural correlates of consciousness, studied phenomena like REM sleep, mapped the brain with Functional Magnetic Resonance Imaging (fMRI) and Magnetoencephalography (MEG), and developed complex theories about how subjective experience emerges from physical processes, fundamental questions persist about the nature of consciousness, its relationship to the physical brain, and whether it exists in forms beyond human experience.
Theories of consciousness generally fall into two broad categories: First-order and higher-order. First-order theories propose that consciousness emerges directly from the properties or strength of perceptual representations. For example, Global Neuronal Workspace theory suggests that consciousness occurs when a first-order representation is "globally broadcast" throughout the brain. Think of it as a theater of mental functioning. Consciousness in this metaphor resembles a bright spot on the stage of immediate memory, directed there by a spotlight of attention under executive guidance. Only the bright spot is conscious, while the rest of the theater is dark and unconscious.¹
Higher-order theories, by contrast, argue that consciousness requires an additional step, that a first-order state must be meta-represented by a higher-order representation.
“The idea is I have this pen in front of me and there's going to be a sort of neural pattern of activity in my occipital cortex, which is where visual input from the eyes and the LGNs [Lateral Geniculate Nucleus – a key structure located in the thalamus that acts as a relay station for visual information, receiving signals directly from the retina and transmitting them to the primary visual cortex for further processing] are coming through,” explains Project Co-lead Axel Cleeremans, Research Director with the National Fund for Scientific Research (Belgium) and Professor of Cognitive Science in the Department of Psychology at Université Libre de Bruxelles.
“We call that a first-order representation because it's the first representation from something that's out there, basically, in my brain,” says Cleeremans. “Higher-order thought theory is saying that to be conscious, that pattern of neural activity needs to be re-represented – it needs to be the object of another representation, which is probably located in frontal cortex, though nobody is certain.”
While first-order theories have received extensive empirical investigation in neuroscience, higher-order theories have been less thoroughly tested. Moreover, higher-order theories are often incorrectly treated as a single unified framework, when in fact they encompass multiple variants with distinct predictions about how consciousness works. Exploring these variants lies at the heart of ETHOS – Empirical Tests of Higher-Order Theories of Consciousness – one of five Structured Adversarial Collaboration research projects being funded by Templeton World Charity Foundation (TWCF).
Launched in the summer of 2024 under the leadership of Stephen Fleming, Professor of Cognitive Neuroscience at University College London and Axel Cleeremans, the ETHOS project aims to compare the different variants of higher-order theories of consciousness through systematic empirical testing along two crucial "axes" of disagreement.
“Within Higher-Order Theory, the different variants essentially disagree on what kind of monitoring process – [axis] – is required to make something conscious. So, we have two classes of theory,” says Fleming.
The first axis concerns whether higher-order representations are "rich" or "sparse" - a distinction that refers to the neural architecture supporting conscious experience rather than the experience itself. Sparse theories propose that conscious experience results from an interaction between first-order and higher-order states, with higher-order states playing a limited role in tracking the precision, intensity, or reliability of first-order states. Even within sparse theories, however, disagreements exist about the nature of the higher-order code - for instance, whether it fundamentally distinguishes between perception and imagination.
Rich theories, by contrast, suggest that conscious experience is fully determined by higher-order states, which must be as detailed and complex as the perceptual experience itself. Under this view, higher-order representations essentially duplicate the content of first-order processing.
“So, what that means in practice is, imagine that I'm trying to explain why you're conscious of seeing an apple. Under a higher-order theory, I need to account for why I not only process the apple in my visual system, but I also become aware of processing the apple,” explains Fleming. “A sparse variant will say, well, all the properties of the apple, the fact it's red and round and so on, that's in my visual system.
“All I need under a sparse higher-order theory is some higher-order state that points to that first-order state and says that's a reliable, strong representation of the apple. And so that's what makes it conscious. Whereas under a rich theory, the whole, all the content about the apple, the fact it's round and colored and so on, gets kind of put into that higher-order state as well.”
The second axis of disagreement concerns whether higher-order representations can "misrepresent" their first-order targets. For example, can someone consciously experience the color green even when their first-order perceptual system is signaling red? This possibility of misrepresentation is central to non-relational higher-order theories, which propose that conscious experience follows the higher-order state even when it mismatches first-order processing. Relational theories, however, preclude such strong misrepresentation, arguing that first-order states supply the mental content that higher-order states merely monitor.
To test these competing ideas, the ETHOS team is preparing three groundbreaking experiments using diverse methodologies. The first, led by Rachel Denison at Boston University and Megan Peters at UC Irvine, will use psychophysical adaptation techniques (the scientific study of how physical stimuli create sensations and perceptions in the human mind) to distinguish between rich and sparse theories. The second experiment, headed by Nadine Dijkstra at UCL in collaboration with Jorge Morales at Northeastern University, will use brain imaging to explore how the brain handles perception versus imagination. A third experiment led by Zoltan Dienes will employ hypnotic suggestion to test whether higher-order representations can misrepresent first-order information.
“So, in the case of the apple, we could hypnotize someone to think a red apple is green,” says Fleming. “And then we can look at whether the higher order representation can change in the brain, changing the visual information. And if that happens, then that would give support to these rich theories that allow misrepresentation.”
“Progress comes by designing really robust experiments and challenging these theories,” says Dawid Potgieter, former Director of Programs in Discovery Science at TWCF who designed and launched the organization’s Accelerating Research on Consciousness Initiative.
“When I launched the initiative, the intention was to make progress by challenging scientists, and to do it in a very collaborative way.”
ETHOS aims to complete its experimental designs by July 2025, with pilot work ongoing until then. To ensure objectivity, Fleming and Cleeremans, while contributing their own theories, have deliberately separated themselves from conducting the experiments. The project will follow high standards set by previous adversarial collaborations, collecting large samples and using hold-out datasets for replication to reduce false positives in complex brain and behavioral data. All experimental procedures, datasets, and analysis tools will be made freely available to the broader scientific community, facilitating future research on theories of consciousness.
“We are currently developing a platform called Accelerating Research where this project will feature alongside four other adversarial collaborations,” says Potgieter, referring to the website he and his team have designed. “We expect the combined outputs from these projects to have a tremendous impact in the field.”
What makes ETHOS (and TWCF’s other projects) particularly innovative is precisely this – an embrace of adversarial collaboration. Rather than working in theoretical "micro-universes" where researchers tend to design experiments supporting their own theories, ETHOS brings together scientists with competing viewpoints to design decisive tests that could challenge – and eliminate – their own ideas.
“Over the past decade or so, there has been a feeling that there's a sort of uneasy stasis that has taken hold of the field,” says Cleeremans. “There's a sense that each of the theories that are involved in these adversarial collaborations have a little bit of the truth, but nobody knows how to move forward. So, this initiative is really meant to be an attempt at reducing the number of contenders by actually proving them wrong.”
While the researchers acknowledge that no single project will fully solve the mystery of consciousness, ETHOS represents a significant step forward in how scientists approach this fundamental question. By bringing together competing theoretical perspectives under the framework of structured adversarial collaboration, ETHOS may not only advance our understanding of consciousness but also provide a model for how scientists can work together to tackle some of nature's most profound mysteries.
“Consciousness is arguably the most complex topic under scientific investigation,” says Potgieter. “It will not be understood through any single experiment. It cannot be fully explained by any existing theory. It will remain a mystery for decades.
“The field has seen many theoretic and empirical innovations. To maximize progress, innovations in conceptual frameworks and experimental methodologies should be supplemented by improvements in how research projects are structured and financed. The process described here is one such example, but there remains ample scope for improvement. There is no way of predicting how the ARC initiative projects will turn out, but I believe the field would be best positioned for any eventuality by encouraging open and carefully considered discussions between researchers representing incompatible theories.”²
¹ Bernard J. Baars, Global Workspace Theory of Consciousness: Toward a Cognitive Neuroscience of Human Experience, Progress in Brain Research, vol. 150. https://www.sciencedirect.com/science/article/abs/pii/S0079612305500049#:~:text=GW%20theory%20may%20be%20thought,theater%20is%20dark%20and%20unconscious
² Potgieter, Dawid. 2024. “Accelerating Research on Consciousness: The Structured Adversarial Collaboration Process.” PsyArXiv. October 24. doi:10.31234/osf.io/rdq52. https://osf.io/preprints/psyarxiv/rdq52
To better understand existing theories of consciousness, Templeton World Charity Foundation (TWCF) funds a series of 5 ambitious research projects, known as Structured Adversarial Collaborations, to allow researchers to evaluate competing hypotheses through carefully designed experiments and rigorous cross-laboratory validation.
This blog is part of a series that complements the launch of Accelerating Research, a unique platform developed by Dawid Potgieter and DataCite with TWCF funding, showcasing results from each Structured Adversarial Collaboration project as it progresses.
