A new theoretical paper proposes that the most basic feeling of being a conscious self originates in the brainstem rather than the outer layers of the brain. By examining how the brain processes pain, anesthesia, and brain injuries, the authors suggest that raw sentience is computed in an ancient subcortical structure, located beneath the cortex, that acts as a gatekeeper for attention. The hypothesis, published in The Neuroscientist, challenges dominant models of the mind that place conscious awareness within the cerebral cortex.
The brainstem is a stalk-like structure at the base of the brain that connects the upper brain to the spinal cord. Leading theories of consciousness locate it in the cerebral cortex, the wrinkled outer layer of the brain responsible for complex thought and language. In the authors’ view, these theories acknowledge the brainstem’s importance but treat it mainly as a power supply that keeps the cortex running. Yet a 2025 adversarial study of human brain activity failed to confirm key predictions of the two leading cortex-centered theories of consciousness, leaving open the question of how and where the brain generates subjective experience.
To address this puzzle, the paper draws on a distinction from philosophy between two types of awareness: access consciousness and phenomenal consciousness. Access consciousness refers to the specific things a person is aware of at any moment, such as sights, sounds, or memories. Phenomenal consciousness, which the authors call raw sentience or the “me inside,” is the underlying feeling of existing as a conscious observer, a presence that remains completely intact even if a person closes their eyes and blocks their ears.
Clinical and animal observations have repeatedly challenged the idea that the cortex creates this raw sentience. For instance, the author of a 2007 review on brain organization argued that children born without a cerebral cortex still display conscious awareness and emotional responses. Similarly, a 2015 clinical case study of a neurological patient found that a person with extensive damage to the brain areas widely believed to process pain could still fully feel and react to painful stimuli.
Tiny disruptions in a specific part of the brainstem can have the opposite effect. In a 2001 rat experiment led by the new paper’s first author, delivering microscopic doses of an anesthetic drug directly into a specific pocket of the brainstem rapidly and reversibly switched off pain responses and other signs of consciousness. The new research, led by Marshall Devor of The Hebrew University of Jerusalem alongside colleague Mark Baron, brings these lines of evidence together into an evolutionary framework proposing that raw conscious feeling originated, and still resides, in the brainstem.
Rather than conducting a single new laboratory experiment, the authors developed a hypothesis-generating theoretical synthesis drawing on clinical neurology, evolutionary biology, and their own laboratory’s prior animal experiments. They first evaluated clinical evidence from human patients undergoing awake brain surgery, patients experiencing epileptic seizures, and individuals with strokes or other brain lesions.
In reviewing surgical data, the authors point out that applying mild electrical currents directly to the visual, auditory, or touch areas of an awake patient’s cerebral cortex reliably causes the patient to see flashes of light, hear sounds, or feel tingling. Yet stimulating the cortex almost never produces a feeling of pain. Even in the insula, a folded region tucked deep within the cortex whose rear portion is considered the strongest candidate for a cortical pain center, one study evoked pain at only 8 of 313 tested sites. Pain is also rarely reported as a warning sensation, or aura, at the start of epileptic seizures, even though seizures can involve almost any part of the cortex and often produce flashes of light, sounds, or smells.
By contrast, direct electrical stimulation of subcortical pain pathways, including in the brainstem, reliably triggers feelings of pain in patients and animals. The authors also note that while extensive damage to the cerebral cortex can erase specific abilities, such as causing blindness or eliminating awareness of one side of the body, it does not cause a loss of raw consciousness or prevent a person from feeling pain. Small strokes or injuries in the upper brainstem, however, often plunge human patients into a coma.
To pinpoint the exact location of this consciousness switch, Devor and Baron examined data from their laboratory’s rodent studies focusing on the mesopontine tegmentum, a region in the upper brainstem. The 2001 mapping study involved 828 microinjections in 182 rats. Only injections into one tiny node, now called the mesopontine tegmental anesthesia area, caused the animals to lose muscle tone, pain responses, and other signs of consciousness, and injections placed even slightly off target had no such effect. In later work, the team found that volumes of 10 nanoliters or less (a nanoliter is a billionth of a liter) could produce anesthesia within seconds.
The researchers also examined what happened when they permanently destroyed cells in this area with a chemical toxin called ibotenic acid. When the damage was limited to a small patch, the rats did not fall into a coma. Instead, they woke up from surgery, resumed sleep-wake cycles with some insomnia, and became less sensitive to anesthetic drugs given through the bloodstream. From this, the authors conclude that the anesthesia area itself is not essential for generating raw consciousness.
When they used larger amounts of the toxin to damage a wider area surrounding that spot, the outcome changed. Six rats that received these larger injections on target fell into an irreversible coma, whereas six rats whose injections landed nearby but off target woke up normally from surgery. Based on these results, the authors propose that the neural circuitry generating raw consciousness in rodents sits within a 3-by-4-millimeter zone of the mesopontine tegmentum that encloses the anesthesia area and overlaps with the parabrachial complex, a major hub for incoming pain signals.
Drawing on these findings, the authors propose an evolutionary model for how consciousness works. When pain fibers in the skin or organs detect tissue damage from a threat like a predator’s bite, that signal already carries direct survival meaning, which the authors call Darwinian fitness valence, as soon as it enters the nervous system. In the authors’ reading, the ancient brainstem evolved sentience hundreds of millions of years ago in early vertebrates to translate these raw tissue-damage signals into the subjective feeling of pain, though they acknowledge there is currently no sure way to know whether fish or reptiles consciously feel pain.
Other senses, such as vision and hearing, do not carry immediate survival meaning at the sensory organ; a pattern of light on the retina could be a harmless cloud or a charging rhinoceros. In the authors’ view, sophisticated analysis of sights and sounds requires the computing power of the mammalian cortex, which evolved roughly 300 million years after early vertebrates, though they acknowledge that animals without a cortex can detect threats by sight and sound at a more basic level. The authors suggest that rather than building a second consciousness system inside the new cortex, evolution wired the cortex down into the existing brainstem consciousness processor.
In this model, the cortex operates like an unconscious supercomputer or a self-driving car, analyzing sights and sounds and sending a summary of their survival importance down to the brainstem. The brainstem circuit acts as an adjudicator, comparing the survival value of incoming signals against whatever the brain is currently focusing on. If a new signal is important enough, the brainstem sends a chemical or electrical signal back up to the cortex, “marking” that stream of information with conscious awareness while leaving routine tasks to run automatically in the background.
These ideas fit with a human neuroimaging study that mapped an 18-node network linking the brainstem to higher brain regions to support wakefulness. That study charted arousal pathways rather than testing a theory of subjective experience, but it likewise placed brainstem circuitry at the foundation of human consciousness.
There are some caveats to consider when evaluating this theory. Much of the brainstem evidence is also consistent with conventional, cortex-centered views, which already hold that the brainstem is needed to keep the cortex awake. Coma after brainstem damage or anesthetic injections does not, by itself, show that subjective experience is generated there. The theory’s distinctive support comes mainly from pain, the one sense the authors argue does not need the cortex.
Even on the question of pain, it is difficult to completely rule out the possibility that when electrical stimulation or seizures in the human cortex occasionally trigger pain, the cortex itself is generating part of that feeling. The authors note that in some epilepsy patients, repeated stimulation of a specific site in the insula reliably triggers pain. Devor and Baron interpret such cases as signals spreading to brainstem pain pathways, but some researchers have argued that the posterior insula functions as a pain center in its own right.
In addition, much of the evidence relies on translating findings between nonverbal animals and humans, or interpreting behavior in patients who cannot speak. In experiments with rats or observations of infants born without a cerebral cortex, researchers must infer conscious awareness and pain from physical reactions and vocalizations, which cannot definitively establish what an individual is subjectively experiencing on the inside. The coma lesion results pointing to the proposed consciousness zone also rest on just a dozen rats. The paper describes them only briefly, giving the animal counts in a figure caption, with no statistical tests or citation to a fuller earlier report.
Finally, the paper presents a hypothesis rather than a definitive test. The authors have not yet identified the specific neurotransmitters or firing patterns the brainstem would use to “mark” cortical activity, and the model does not solve the philosophical “hard problem” of how biological cells produce a first-person feeling at all. They suggest that brainstem neurons might communicate by releasing chemicals that spread broadly through nearby tissue, much as dopamine does in parts of the brain, rather than through precise cell-to-cell connections. That kind of analog processing is not captured by today’s silicon computers, and the authors call for experiments to test these cellular mechanisms directly.
The study, “Brainstem Theory of Consciousness,” was authored by Marshall Devor and Mark Baron.