Evidently you are conscious and, better yet, you are indulging one of its useful privileges: cognition. But how does your brain achieve your current experience of integrating sensory information—the words and images on this page—with the internal knowledge, motivations and reasoning you are using to understand them?
Picower Professor Earl K. Miller and Edward Hood Taplin Professor Emery N. Brown think about that question (and related ones) a lot, both independently and in close collaboration. But they do so for opposite reasons. As a cognitive neuroscientist, Miller’s job is to discern how the brain endows us with intellectual abilities such as attention, working memory and reasoning. As an anesthesiologist, Brown’s job is to reliably induce, maintain and then conclude a safe and appropriate level of unconsciousness for his patients.
“I’m interested in unconsciousness because I’m interested in consciousness, and I think Emery would say the opposite,” Miller quipped.
By comparing and contrasting brain activity in both states, Brown and Miller are building an understanding of how unconsciousness and consciousness each become manifest. The crux of their findings is that consciousness and cognition require the transmission of information via brain waves, which arise from the rhythmic electrical activity of coordinated groups of neurons. Think of brain waves as biological Wi-Fi signals. They integrate brain regions that have different information processing responsibilities into functional networks.
These waves can vary in their power, frequency, location and degree of alignment (or “phase”). Miller and Brown have shown that these properties differ markedly and systematically when we are conscious vs. when we are not. When brain wave patterns aren’t conducive to information exchange via synchronized, aligned, higher-frequency signals, the experience of consciousness in which external sensation and internal thought feel integrated, falls apart, Miller said.
“Consciousness is a unified experience of the sights, sounds, feelings, knowledge, etc in any given moment. Many major theories of consciousness involve knitting together networks across the brain so they can create that unified experience.” said Miller, who on April 23 delivered a keynote address remotely to the 30th Annual Science of Consciousness Conference at the University of Arizona. “And for the same reason these [brain wave] dynamics can organize thoughts, they can also knit together the unified experience of consciousness. We found that loss of consciousness is associated with the dramatic alternations of these dynamics through the different effects of different anesthetic drugs.”
Unconscious lessons
Brown has long studied how anesthetic drugs like propofol, ketamine, or dexmedetomidine produce states of unconsciousness that differ from that of sleep by profoundly (but only temporarily) impairing sensory and cognitive processing. He and colleagues have traced how the drugs’ molecular effects on neurons in specific brain regions alter normal oscillatory activity in key brain circuits.
Brown’s work has shown that each drug produces a distinct brain wave signature in patients that systematically varies with factors such as drug class, patient age and patient state of health. Monitoring these signatures with scalp-mounted electroencephalogram (EEG) electrodes in the operating room in real-time, a practice that he employs and advocates, reduces the guesswork of inferring how unconscious the patient is. Why rely solely on physical signs such as a lack of movement and steadiness of heart rate and blood pressure when you can also directly measure brain state? With a brain-based indicator of unconsciousness, anesthesiologists can refine anesthetic dosing, preventing the administration of too little or, more commonly, too much.
Last year Brown’s lab published a clever method for assessing unconsciousness while volunteers received dexmedetomidine. Speaking with volunteers can prolong wakefulness and accelerate reawakening. The “breathe-squeeze” test required volunteers to squeeze a ball every time they breathed. Once they couldn’t they were judged unconscious and once they resumed, they were deemed reawakened—no dialogue required. Meanwhile, the researchers correlated the apparent loss and resumption of consciousness with the brain state changes apparent in the EEG.
In collaboration with Miller, who works with research animals, Brown has further validated the EEG signatures of some anesthetics by simultaneously measuring the electrical discharges (or “spikes”) of hundreds of individual neurons. Brain waves (or “rhythms”) arise when spikes of large groups of neurons are synchronous, so these measures confirmed directly from brain cells what the waves measured from outside the head seemed to indicate.
“It gives me a way of interpreting the EEG in a way that is much more neurophysiologically based,” Brown said. “When I see a dramatic alteration of spiking activity associated with whatever rhythm I’m looking at, it lends support to the idea that altering rhythms is associated with the impairing the ability of the brain regions to communicate with each other.”
In 2021 Brown and Miller’s labs showed that under propofol, neurons that spiked as many as 10 times a second during wakefulness spiked once a second or less. The brain therefore could only produce waves of very low frequency across the cortex. Wave coordination and power at higher frequencies associated with consciousness were greatly reduced. The study also showed reduced coordination between the cortex and a deeper region called the thalamus. Consciousness is not solely produced by the cortex, Brown notes, but that is where most cognitive functions take place.
Brown and Miller have also used spiking data from animals and EEG data from humans to establish ketamine’s brain wave signature. Unlike with propofol, ketamine-mediated unconsciousness included periods of high-frequency waves alternating every 4 to 10 seconds with low frequency waves. This pattern is also quite different from consciousness.
“I can make you unconscious by making your brain hyperactive in some sense, or I can make you unconscious by slowing it down,” Brown said at the time. “The more general concept is there’s a dynamic—we can’t define it precisely—which is associated with you being conscious and as soon as you move away from that dynamic by being too fast or too slow, or too discoordinated or too hypercoordinated, you can become unconscious.”
Last November Brown and Miller showed that unconsciousness under propofol is characterized not just by a broad change in brain wave patterns but by a disruption in their propagation from region to region. While awake and then under anesthesia, animals received sound and touch stimulation. As researchers have shown for decades, a region in the animals’ brains that processes raw sensory input still processed the incoming stimulus, even while under anesthesia. But the researchers also measured neural spiking, waves and synchrony in three other regions of the cortex. During wakefulness, by all measures, all four cortical regions shared a neural response to the stimulus. Under anesthesia, such activity was absent outside the sensory region.
Waves of conscious cognition
If anesthetics knock brain waves out of a pattern that enables consciousness and cognition, what does that pattern look like and how does it integrate experience? Miller has been working on that for years. In 2007 his lab published a study showing that when a new sensory stimulus focused an animal’s attention, synchrony between cortical regions was evident in fast “gamma” frequency brain waves. When they focused attention based on task rules, synchrony was evident in relatively slower “beta” waves.
Further studies by the Miller lab have shown that this dynamic—beta waves carry information about rules and intentions; gamma waves encode sensory information—also applies in tasks of working memory and making predictions. The research has also shown that beta seems to regulate gamma to control cognition. When beta waves are prominent, gamma power is suppressed. For instance, when an animal playing a memory game needs to remember a newly presented image, beta gives way so gamma can encode the image. But when the image needs to be remembered in advance of the memory test, beta takes over and prevents gamma from encoding distractions.
Miller’s research, including a paper earlier this year, has shown that beta waves arise most strongly across the cortex in its deeper layers while gamma waves have primacy in more superficial layers. Studies in his lab have also shown that these waves physically travel through specific areas of cortex (and that anesthetics radically alter those travels).
All this evidence—that beta controls gamma in spatially precise ways—led Miller to formulate a new theory of cognitive control: Spatial Computing. To selectively control just the right neurons at the right times to do the right things, the brain uses beta waves like a stencil, patterning when and where gamma waves are “allowed” to encode new information. In this way, the brain can recruit groups of neurons to represent new information within the context of a task’s rules. When you hear the combination of a lock, according to Spatial Computing, your brain’s beta waves will assign the rules (turn left, turn right, turn left again) to specific patches of your cortex and then neurons in each patch will encode the relevant number of the combination (e.g. 32, 14, 19).
Spatial Computing answers some questions about how thoughts and sensory experiences are integrated quickly and flexibly enough to produce useful cognition, Miller said. Brain waves are based on electric fields, so they can arise and spread very fast. By assigning both task rules and sensory encoding responsibilities to neurons in a patch, spatial computing explains how the cells come to represent multiple aspects of a task (a property called “mixed selectivity”). Moreover, the involvement of different wave frequencies enables task rules and sensory encoding to vary independently. If the combination changes, the brain doesn’t have to relearn the rules. The beta waves encoding the rules can stay the same even as the gamma waves encoding the new numbers vary.
Along with the answers it provides, Spatial Computing raises big questions, too, Miller acknowledges. How does the brain generate the waves that implement these dynamics? Does the brain formulate a map to manage its thoughts? Is that map therefore a map of your consciousness in some way? Answering any of those questions will require future “waves” of research and insight.