Behind the gleam of blinking screens, scientists at the Allen Institute for Brain Science staged one of the most intriguing confrontations in science. The goal was to uncover what makes us conscious—what transforms light bouncing off a tree or the sound of a bird into the rich and subjective experience of waking life.
To achieve this mission, since 2018, the scientists have been testing two of the most influential theories about the rise of consciousness.
Seven years later, the results are in.
The findings offer no final verdict on the nature of consciousness. But they deal a blow to both of the leading theories—Integrated Information Theory (IIT) and Global Neuronal Workspace Theory (GNWT)—while providing new clues about how the brain transforms sensory input into the experience of being.
“Adversarial collaboration fits within the Allen Institute’s mission of team science, open science and big science, in service of one of the biggest, and most long-standing, intellectual challenges of humanity: the Mind-Body Problem,” said Christof Koch, meritorious investigator at the Allen Institute. “Unravelling this mystery is the passion of my entire life.”
Two Theories, One Conscious Mind
The two theories try to explain the origin of consciousness in strikingly different ways.
IIT, spearheaded by Italian neuroscientist Giulio Tononi, claims consciousness arises from how much information a system integrates within itself. According to IIT, the “hot zone” for this process lies in the back of the brain—the posterior cortex—where networks supposedly weave together the web of experience. But according to IIT, it is not where the activity happens that matters, but how deeply it is integrated.
GNWT, championed by French cognitive neuroscientist Stanislas Dehaene, argues that consciousness emerges when information becomes globally accessible across a network of brain regions—a kind of neural spotlight, centered largely in the prefrontal cortex (PFC), that brings thoughts and perceptions into awareness.
The study began as an “adversarial collaboration”—a relatively rare scientific arrangement in which theorists with opposing views agree to a shared experiment, using standardized methods and blinded analyses. The goal is to reduce bias and settle long-standing debates not with words, but with data.
<!– Tag ID: zmescience_300x250_InContent_3
–>
The 2019 experiment tested both theories head-to-head. Involving 256 adult participants—a record-breaking number of people for this type of brain study—the research team used visual stimuli and a combination of brain imaging tools: functional MRI to track blood flow, EEG to capture electrical activity, and MEG to measure magnetic fields. These tools allowed them to observe the brain as it responded to conscious and unconscious visual experiences.
The researchers laid out three critical predictions where IIT and GNWT would clash:
- Where in the brain is conscious content represented?
- How is it maintained over time?
- Which regions synchronize to support it?
Participants viewed suprathreshold visual stimuli—faces, letters, objects, even fake symbols—presented at different angles and for different durations. They were asked to detect rare targets, but most of the time, they simply watched as their brains were scanned.
What this allowed was a full-spectrum view of the brain’s dance—from millisecond-level flickers of electric fields to broad patterns of blood flow—capturing consciousness from multiple angles at once.
The results didn’t hand victory to either theory. Both IIT and GNWT got some things right—and some very wrong.
“It was clear that no single experiment would decisively refute either theory,” said Anil Seth, a professor of cognitive and computational neuroscience at the University of Sussex and one of the study’s co-authors. “The theories are just too different in their assumptions and explanatory goals, and the available experimental methods too coarse, to enable one theory to conclusively win out over another.”
For instance, conscious content could be decoded in both the posterior cortex and the prefrontal cortex, aligning with both theories. But when it came to more detailed aspects like the orientation of a face, the prefrontal cortex came up short. That was a problem for GNWT.
At the same time, IIT predicted that conscious content would be maintained in the back of the brain for as long as a stimulus was visible. The evidence here was mixed. Some neural signatures in the posterior cortex did track stimulus duration, but only a minority of electrodes showed this sustained activity.
Even more damaging was the failure to find sustained synchronization within posterior regions—the key claim of IIT. Connectivity was brief and patchy, not the continuous integration the theory calls for.
And GNWT? It fared no better on its own turf.
The theory predicted distinct bursts—or “ignitions”—in the prefrontal cortex at both the beginning and end of a conscious experience. But those offset ignitions were missing. Though the prefrontal cortex lit up when a stimulus appeared, it stayed mostly quiet when it vanished.
“Much has been learned about both theories and about where and when in the brain information about visual experience can be decoded from,” said Seth. “Having said all this, the findings of the collaboration remain extremely valuable.”
A New Model for Investigating Consciousness
The greatest achievement of the study may not lie in what it found, but in how it was done.
Unlike typical neuroscience studies—often conducted by a small team aiming to support a hypothesis—this effort was born of open science. The predictions were publicly registered in advance. The experiments were replicated across labs and technologies. The theory proponents themselves stepped back from data analysis to reduce bias.
It was a scientific duel, but fought with mutual respect and transparency. This isn’t the end of the road for either theory, but it is a recalibration.
“Adversarial collaborations are a powerful social process, little used because of its challenging nature,” Koch added. “But it requires a great deal of cooperation and work.”
Maybe consciousness isn’t located strictly in the front or the back of the brain. Maybe it doesn’t live in bursts or seamless continuity. Perhaps it’s more like a shifting mosaic, stitched together by fleeting alignments of neural activity. More studies are already underway. The next confrontations may involve animal models, higher resolution brain scans, or radically new theories altogether.
The implications go beyond the philosophical. They may help doctors detect signs of consciousness in patients who appear unresponsive, such as those in comas or vegetative states. In 2023, a New England Journal of Medicine study estimated that up to 25% of such patients may retain “covert consciousness,” detectable only with brain scans.
In the meantime, the mind remains mysterious. But now, it has fewer places to hide.
The findings appeared in the journal Nature.
