Adversarial collaborations like COGITATE are establishing new methodological standards in consciousness research while advancing understanding of how subjective experience arises from brain activity.
What is consciousness? And how are specific types of brain activity related to our deepest subjective experiences?
For millennia, consciousness lived in the realm of philosophy. But, with the advent of advanced neuroimaging techniques, scientists gained powerful new tools to study how conscious experience manifests in the brain – a shift that has led to the development of multiple competing theories about the neural basis of consciousness, with different research groups championing various explanations.
To better understand existing theories of consciousness, the Templeton World Charity Foundation is funding a series of ambitious research projects, known as Structured Adversarial Collaborations, to allow researchers to evaluate competing hypotheses through carefully designed experiments and rigorous cross-laboratory validation.
The first in the series, known as COGITATE – Collaboration on Global Neuronal Workspace Theory (GNWT) and Integrated Information Theory (IIT) Testing Alternate Theories of Experience focused on testing two theories that make distinct predictions about how conscious experience arises from neural activity. Specifically, GNWT proposes that consciousness requires global broadcasting of information by a fronto-parietal network, indexed by a non-linear “ignition” response that amplifies information across brain areas. IIT, conversely, proposes that consciousness emerges from integrated information primarily in posterior brain regions, and equates consciousness with a maximum irreducible, intrinsic cause–effect power.
“The idea of consciousness is such an enigma to us all,” said Tanya Brown, Scientific Project Manager at the Max Planck Institute for Empirical Aesthetics, where COGITATE is coordinated. “I think we all want to understand how our brains work.”
The experimental design, developed collaboratively by theory proponents and research teams from eleven participating institutions across the world – including the Max Planck Institute for Empirical Aesthetics, Reed College, Tel Aviv University, Harvard University, Yale University, NYU-Langone, the University of Wisconsin, the Commissariat à l'Energie Atomique et aux Energies Alternatives, (“CEA”), the Donders Institute, the University of Birmingham, and Peking University – involved data collection using three different neuroimaging methods: Functional Magnetic Resonance Imaging (fMRI); Magneto-electroencephalography (M-EEG); and Intracranial Electroencephalography (iEEG). The idea being a multi-modal approach would be more comprehensive in terms of its findings.
“Each neuroimaging modality has pros and cons, strengths and weaknesses,” said Brown.
fMRI, for example, is particularly good at spatial differentiation, which can indicate where in the brain something is happening with a great degree of accuracy. It does this by measuring blood oxygen level dependency. As the presence of oxygen represents changes in activity in the brain, a higher concentration of oxygen molecules is indicative of more activity happening
“fMRI is a tool that enables us to examine what parts of the brain are active during different tasks. It works by detecting changes in blood flow, known as the BOLD-response. When a brain area is more active, it needs more oxygen, which causes more blood to flow there. fMRI uses radio frequencies and powerful magnets to detect these changes and creates images showing which parts of the brain are involved in the given task.
Unlike fMRI, M-EEG (which is a combination of magnetoencephalography and electroencephalography) is less effective for spatial measurements, but is effective at recording time – e.g., it can track activity in the brain at the exact millisecond it occurs. Using the collection of these methodologies, therefore, enables researchers to track, with high accuracy, where and when in the brain things change for subjects performing the exact same task.
The third methodology, iEEG, is an invasive approach that requires the implantation of electrodes in a subject’s brain. For COGITATE, it was used on patients who were in one of the partner clinics undergoing treatment for epilepsy or other protocols and who willingly agreed to participate in the project.
“When you integrate these three different modalities, it gives a very clear, convincing picture in the brain, that when somebody is doing X, this is where you see a change and this is at what time,” said Brown.
To enable researchers to examine how the brain represents different aspects of conscious content, the project focused on two key experiments, the first of which had participants view images from four categories (faces, objects, letters, and false fonts) presented at different orientations and durations while searching for specific targets.
“The first experiment kind of pushed the theories to the limit because the stimuli are clearly seen, failing to confirm their predictions cannot be explained away by weak signals,” said project co-lead Liad Mudrik, whose lab at Tel Aviv University investigates consciousness and high-level cognition, both at the behavioral and neural level.
“At the beginning of each block, you see two targets. One could be a face and the other an object. And then you see a stream of stimuli presented for either 500 milliseconds, a second, or a second and a half. Whenever you see the targets, you're supposed to hit a button so, what this very simple manipulation does is differentiate between stimuli that are task relevant and those that are task irrelevant. So, if you're looking for a specific face and a specific object, then all faces and all objects are relevant, but not all are targets.
“Letters and false fonts, in this trial, would be task-irrelevant. This allows us to differentiate between neural patterns involved in task performance from those involved in consciously processing the stimulus”.
The second experiment took the form of an engaging video game where participants had to collect falling objects while faces and other stimuli appeared in the background. This clever setup allowed researchers to compare brain activity when participants consciously perceived these background stimuli versus when they missed them.
"Each data collection site had to have the exact same protocol and the exact same instructions,” said Brown, emphasizing the project’s rigorous standardization across research sites. “Even the way that you tell a subject to perform in a study, that can have an impact.”
Results from the first set of experiments, which were published in a preprint late last year, present a nuanced picture that both supports and challenges aspects of each theory. For example, evidence of prefrontal cortex involvement in some aspects of conscious content supported GNWT, though researchers did not observe the predicted "ignition" pattern at conscious experience offset and were not able to decode all aspects of experience from frontal areas. For IIT, the findings revealed sustained activation tracking the duration of experience, but did not include the expected sustained synchronization between posterior areas.
"Even before we collected the data, we made theoretical progress by forcing the theories to specify their predictions,” said Mudrik, speaking about the project’s impact beyond specific scientific findings. “And, for the first time, we now have a list of specific areas – we know it's not just front versus back. We have clear predictions about where exactly in the brain it should happen.”
In terms of potential clinical applications, COGITATE could also prove helpful for assessing consciousness in unresponsive patients, as both theories inspired tests for consciousness in clinical settings. According to Mudrik, recent research suggests that "one out of four patients who is non-responsive is actually conscious.” Enhanced understanding of neural signatures of consciousness could improve diagnostic accuracy and patient care. Additionally, it could also inform debates about consciousness in other contexts, from fetal development to artificial intelligence systems.
"The work that we are now doing might have practical implications with respect to understanding which systems or organisms are conscious and which are not,” said Mudrik. “Though its most immediate importance is the ability to better understand and evaluate the two tested theories.”
Looking forward, the researchers hope this approach will help build a more comprehensive evidence base for evaluating theories of consciousness.
"At the end of the day, I'm hoping that all the different adversarial collaborations will yield valuable data sets, and together might form a critical mass of data and evidence that would allow us as a community to arbitrate more and weigh the different theories better than we can do now,” says Mudrik.
By combining rigorous experimental methods, pre-registered analyses, and open science practices, adversarial collaborations like COGITATE are helping to establish new methodological standards for consciousness research while advancing understanding of how subjective experience emerges from brain activity.
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.
