We don't always understand our emotions, but we couldn't lead normal lives without them. They steer us through life, guiding the decisions we make and the actions we take. But if they're inappropriate or stick around for too long, they can cause trouble.
Neuroscientists and psychiatrists, despite their best efforts, don't understand nearly enough about the brain activity underlying our emotions, how they make us tick and how they can make us sick.
Now, in a study scheduled to publish May 29 in Science, Stanford Medicine investigators have mapped the brainwide neuronal processing that underlies the emotional response triggered by a mildly unpleasant sensory experience. Features of this brain activity turn out to be shared by humans and mice — and, by extension, every mammal in between. (Perhaps your pet has already explained this to you.)
The findings could help unveil some of the driving forces behind numerous neuropsychiatric disorders, which are characterized in large part by troublesome emotional manifestations.
"Emotional states are fundamental to psychiatry," said Karl Deisseroth , MD, PhD, professor of bioengineering and of psychiatry and behavioral sciences, who led a collaborative team effort spanning Stanford Medicine's hospital and laboratory facilities. Sharing senior co-authorship of the study with Deisseroth are Carolyn Rodriguez , MD, PhD, professor of psychiatry and behavioral sciences; Vivek Buch , MD, assistant professor of neurosurgery; and Paul Nuyujukian , MD, PhD, assistant professor of bioengineering and of neurosurgery. The lead co-authors of the study are postdoctoral scholars Isaac Kauvar, PhD, and Ethan Richman, PhD, and MD/PhD student Tony Liu.
The study was a project of Stanford Medicine's Human Neural Circuitry research program , a multidisciplinary collaboration founded and led by Deisseroth, designed to understand the principles underlying the inner workings of the human brain in health and disease. The HNC program develops and brings together, in an inpatient medical setting, state-of-the-art methods for synchronous and ultraprecise measurement and perturbation of both human behavior and brain activity.
In this study, Deisseroth and his colleagues focused primarily on responses to negative sensory experiences. But he suspects that the brainwide activity pattern his team observed also generalizes to positive experiences. (His group is exploring those, too.)
Pulling it all together
"The mammalian lineage has made a huge evolutionary commitment to large brain size, with all its attendant costs and benefits," said Deisseroth, who is the D. H. Chen Professor and a Howard Hughes Medical Institute investigator. Even a mouse's brain (which is large compared with same-sized non-mammals) contains nearly 100 million neurons; a human brain, almost 90 billion — about 1,000 times as many.
"A bigger brain means a richer, more complex mental life," Deisseroth said. "But there are real constraints once you scale up. The human brain is so big, it takes some time for those rich and complex signals to fully propagate throughout the brain, converge and be properly integrated. Yet, to make accurate decisions, your brain has to pull together your multiple streams of sensory data, your goals, your position in space, your physiological needs and more — all at the same time. If that doesn't happen, wrong decisions will be made and wrong actions taken."
Emotions may represent states that integrate a great deal of information to guide lasting patterns of behavior, but may need a window of time with persistent communication among widely separated brain structures to accomplish that integration, Deisseroth said.
"Tuning the time scale of this communication could be an important aspect of typical brain function," Richman added. "This would be akin to the action of a piano's sustain pedal, which extends the duration of briefly played notes." Either overly shortened or overly prolonged stability of such brainwide communication patterns could contribute to neuropsychiatric disorders characterized by emotional dysfunction.
What might those emotion-enabling patterns of activity be? Because human brain activity is so complex, figuring out which observed signals are the important ones is a challenge.
Deisseroth is renowned for developing optogenetics , a sophisticated and now widespread method using a targeted light-activated protein together with pulses of light to induce select nerve cells, or groups of them, to fire or go silent at the flip of a switch. But the new study (relying on briefly hospitalized human patients) did not use optogenetics at all.
Instead, the Stanford team used a clever evolutionary trick. To determine how emotion emerges in response to experience, the researchers carried out brainwide screens of neural activity in both mice and humans — two species that emerged from the same ancestor some 70 million years ago — to search for activity patterns present in both species that could be induced by the same emotion-generating stimulus, measurable in the same way, synchronized with the same high-speed behaviors and blocked by the same interventions.
"This approach allowed us to focus our study on the key principles that were shared between mice and humans," Kauvar said.
If, over that vast amount of evolutionary time, a particular brain-activity pattern (ultimately determined by genes governing brain structure and function) doesn't help survival and reproduction, it will be lost, Deisseroth said, while "if a brain dynamical principle is conserved over that time, you'd better believe it could be important."
Puff, blink, squint
First the reflex, then the emotional response: You burn your hand on a stove, reflexively pull it away, then feel the pain spreading and curse. The sound of a gunshot — or a similar noise — on a dark street in a strange neighborhood late at night elicits a reflexive ducking response, then a sense of fear and caution.
Examples of emotion emerging from an unpleasant sensory input are too numerous to list. But those instances are typically tough to measure and often both difficult and dangerous to stage. For experiments, the triggering stimulus needs to be safe, reproducible and easy to deliver — and, in this case, applicable to mice and people.
For this study, the method of choice was a tool employed in every eye doctor's office. Deisseroth's team took advantage of the device an ophthalmologist uses to deliver little puffs of air to check the pressure in their patients' eyes. While not a painful experience, it certainly can be a touch unpleasant. Here, employing this aversive but medically safe stimulus permitted precision in timing, duration and intensity of the stimulus. The researchers knew exactly when each puff started and when it stopped — critical for tracking each subject's brainwide response to it.
The scientists administered multiple series of precisely timed "eye puffs" to participants, who, asked how they felt about the puffs, described them as "annoying," "unpleasant" and "uncomfortable," though certainly not painful. Repeated rapid-fire eye puffs produced an increasing feeling of annoyance that outlasted the eye puff series.
That bummed-out state of mind can be adaptive, Deisseroth noted. "Any repeated series of negative events is important to the brain, to be considered in guiding future behavior."
To record brainwide activity at high resolution, Deisseroth and his associates recruited a cohort of patients at Stanford Hospital who, because they were experiencing frequent seizures that were inadequately responsive to medications, had had electrodes surgically inserted deep into their brains so that teams of neurologists and neurosurgeons, to achieve more targeted treatment, could locate each patient's unique focus — the hyperexcitable point of origin from which seizures would spread across otherwise healthy brain tissue.
While all those electrodes had been implanted in patients' brains for purely clinical reasons, it provided a serendipitous avenue for experiments that would otherwise be difficult or impossible to perform.
"These patients typically spend about a week in a hospital bed with limited mobility, during recording from these implanted intracranial electrodes, while the treatment team waits for spontaneous seizures to occur," Liu said. During this long stretch of time, these patients were more than willing to volunteer for and participate in the investigators' innovative study.
Subjects' visible responses to randomly timed eye puffs were found to be quite consistent. Immediately in response to each puff, the subjects briefly blinked reflexively. In the seconds following each puff, subjects also exhibited additional eye squinting or rapid additional blinks. This additional post-puff eye closure was a natural response to an unpleasant stimulus (since they could not predict the timing of the next puff). It was also precisely quantifiable, offering insight into emotion-triggered behaviors immediately following a sensory stimulus.
All the while, the experimenters tracked subjects' brainwide activity. They picked up a distinctive two-phase pattern: In the first roughly 200 milliseconds after the eye puff they observed a strong but short-lived spike of activity broadcasting "news" of the eye puff throughout the brain. This was followed over the next 700 milliseconds or so by a separate, longer-lasting phase of puff-triggered brain activity more specifically localized to a subset of specific circuits across the brain associated with emotion. This pattern — which, Deisseroth noted, was discoverable thanks to the simultaneous electrical recording and behavioral technology of the team — displayed the interesting property of yielding an extended window of time for brainwide communication, which could be related to emotion.
Since the core idea of the study was to search for shared principles among humans and mice, the scientists carried out the same experiment in parallel in mice. Remarkably, the team observed a very similar two-phase pattern of brain activity in mice. Moreover, delivering a series of eight eye puffs in rapid succession to mice induced accumulating second-phase brain activity and put the mice into a generalized negative emotional state, as further evidenced by their persistently reduced willingness to engage in reward-seeking behavior. (Such persistence and generalizability are classic hallmarks of emotion.)
Gone with the squint
The researchers then used a medication, chosen to be suitable for use in both humans and mice, to further test for the importance of this persistent activity pattern. Ketamine, widely used at high doses in anesthesia, is FDA-approved at lower doses as an antidepressant. Even at these lower doses, ketamine is known to cause a phenomenon called dissociation, in which typical emotional responses to stimuli are reduced or absent.
"Ketamine recipients are fully aware of sensory experience, but they often don't have typical emotions about that experience, even if the sensation would normally be unpleasant," Deisseroth said. "It's as if it's happening to someone or something else." This dissociative effect of ketamine wears off within an hour or so, he said.
After carefully setting up their research protocol so they could safely administer a single dose of ketamine to electrode-implanted human subjects in the hospital, and with fully informed consent, the scientists found that indeed the negative emotion caused by the repeated puffs of air (as described by the patients) was greatly inhibited.
An important part of the clinical study was the ability to directly ask participants about their experiences, Liu said.
"The air puff . . . felt entertaining," one participant said. "It felt like little whispers on my eyeballs," said another.
Consistent with this loss of their subjective sense of annoyance, the human subjects also did not show self-protective behavior — they kept their eyes open between puffs even though they were fully aware of the puffs and continued to have robust reflexive blinks. Remarkably, the same selective effect on behavior (preserving the reflexive blink while blocking self-protection with prolonged eye closure) was observed in the mice.
The team carried out a final set of definitive measurements to test their core hypothesis. If the persistent second phase of brain activity were important in the emotional response, this slower phase would be predicted to be selectively reduced by ketamine in both species, thereby effectively speeding up the brain's response. In humans and mice alike, the team found that the initial fast burst of brainwide activity was completely unaffected by ketamine. But when the scientists measured the speed at which the slower, second phase of post-eye-puff brain activity subsided, they found that ketamine sped up this decay, effectively sharpening the brain's response and restricting the puff-induced activity to a brief window of time (analogous to releasing a piano's sustain pedal to terminate the note).
"This all points to that persistent second phase of brain activity as being strongly linked to emotional state," Kauvar said.
If speeding-up of brain activity prevents formation of emotional states, this acceleration due to ketamine should also be detectable even in the eye puff's absence. As predicted, the team found that the "intrinsic time scale" — a measure of the time over which brain-activity patterns were correlated — was accelerated by ketamine even without the eye puff. In both species, intrinsic time scale rapidly recovered to its normal duration after the ketamine wore off.
Finally, the team found that ketamine also reversibly reduced synchrony across the brain in both species. "Dissociative medication may render the stabilizing phase of brain activity so ephemeral that information can't be properly integrated across the brain, including to build an emotional state," Deisseroth said.
A science of emotion based on timing?
These tunable, measurable timing properties, when pushed beyond a typical range — either in the slowed or sped-up direction — could offer clues about categorizing, quantifying and perhaps even treating neuropsychiatric disorders.
"Far too-brisk decay of that integrative brain activity (as ketamine causes) could generally prevent coordination of information flowing in from diverse regions of the brain," Deisseroth said. This could give rise to a situation in which the right hand quite literally doesn't know what the left hand is doing. "People with schizophrenia report perceptions of alien, as opposed to self-generated, control over their actions," Deisseroth said.
On the other hand, if a brain disorder causes the second wave of brain activity to decay too slowly or to accumulate excessive strength (perhaps due to differences in brain wiring or gene expression, or even related to personal experiences), this could result in hyperstabilized brain states and, consequently, persistent or untimely emotions or intrusive thoughts like those experienced by people with post-traumatic stress disorder, obsessive-compulsive disorder, depression or eating disorders. Different symptoms (and different disorders) would be expected to arise depending on the specific circuits representing this altered persistence.
Distinct from emotion in health and disease, this same quality of signal persistence could powerfully influence the fundamental speed of information processing, another property that varies substantially in the human population. "People with autism spectrum disorder are often known to have trouble keeping up with high-speed bursts of information, an ability required for language and social-information processing," Deisseroth said. Could a hyperstabilized brain state be responsible for difficulty in following rapidly changing input?
"These are fascinating possibilities, which we are now exploring," Deisseroth said. "It's amazing what an unbiased brainwide screen can reveal, especially with the right technology and across millions of years of evolution."
Stanford University's Office of Technology Licensing has filed a patent for intellectual property associated with the study.
Researchers from the Veterans Affairs Palo Alto Health Care System and Weill Cornell Medicine contributed to the work.
The study was funded by National Institutes of Health (grants P50DA042012, R01MH105461, R01MH133553 and R01NS095985), the AE Foundation and anonymous donors.
