Researchers used spatial light modulators (above, with mouse for scale) to turn a laser beam into a hologram that stimulated specific neurons in the mouse brain.

Sean Quirin, James Marshel, Cephra Raja, and Karl Deisseroth/Stanford University

Laser holograms stimulate brain cells in mice to probe roots of perception and hallucination

How many neurons does it take to spark a memory, a sensation, or a movement? Neuroscientists have struggled to answer this question with relatively crude methods that don't allow them to fire up individually selected brain cells. Two teams, however, have recently adapted optogenetics—a technology for stimulating neurons with light—to precisely awaken particular cells in the visual cortex of a mouse. They showed that zapping just a few neurons could trigger the same brain activity as showing the animals a visual pattern and could make them react as if they had seen that pattern. "Essentially, they take control over the internal world of the brain," neuroscientist Thomas Knöpfel of Imperial College London says of the new experiments.

"We don't know how many cells it might take to trigger a more elaborate thought, sensory experience, or emotion in a person," says Karl Deisseroth, a neuroscientist and psychiatrist at Stanford University in Palo Alto, California, who led one of the new studies, published online this week in Science, "but it's likely to be a surprisingly small number, given what we're seeing in the mouse."

That observation might help explain why disordered states—hallucinations, unwanted thoughts, and harmful actions—arise so readily in the brain, Deisseroth says. And single-neuron optogenetics may someday point researchers toward highly targeted ways of stamping out these states and treating symptoms of brain diseases.

Neuroscientists have spent decades watching how mice behave when parts of their brains are stimulated with electrodes or, more recently, with optogenetics, which involves introducing a gene for one of several light-sensitive proteins called opsins into neurons. In most experiments, researchers awaken opsin-bearing neurons of a specific cell type with a pulse of diffuse blue-green light. But Deisseroth's group and others have been targeting optogenetics more precisely with a red light–sensitive opsin and the sharp, penetrating beam of a near-infrared laser.

"Imagine every neuron in the brain like a key on the piano," says Rafael Yuste, a neuroscientist at Columbia University who has pioneered such experiments. "You can literally choose which neurons to turn on."

Replaying perception

In a new study, researchers used light to precisely activate cells in a mouse's visual cortex, re-creating the brain activity involved in seeing specific patterns.

“Lick” signal “Don’t lick” signal Neurons tuned to vertical bars Calcium imaging Vertical bar response recalled Subset of tuned neurons stimulated with light Near-infrared laser RecordingMice saw one of two on-screen patterns while a microscope captured which neurons were “tuned” to respond to one pattern or the other. PlaybackStimulating some of those cells with light reactivated thebrain’s response to a pattern and made mice act as ifthey were seeing it again—with nothing on the screen.
L. CARRILLO-REID ET AL., CELL (2019), J. MARSHEL ET AL, Science (2019) ADAPTED BY V. ALTOUNIAN/SCIENCE

In the two new studies, Deisseroth's and Yuste's groups targeted predefined sets of cells by sculpting the laser beam into a hologram with a device called a spatial light modulator. Along with an opsin gene, they injected the gene for a molecule that fluoresces when neurons fire, allowing them to discern what cells were active. They showed the mice a pattern of drifting parallel lines on a screen and trained them to lick at a water spout when those lines were in one of two orientations (horizontal or vertical), but not in the other. They identified the cells "tuned" to fire preferentially for either the horizontal or vertical pattern.

Yuste's group, which published its experiments last month in Cell, found that stimulating as few as two particularly well-connected neurons made the mouse more likely to lick when the vertical bars on screen were hard to discern. In some trials, the stimulation even prompted the animals to lick when there was nothing on the screen.

The results, Yuste says, support the long-standing theory that ensembles of co-activated neurons—not individual cells—form the basic building blocks of our perceptions and memories. That's still a controversial suggestion, says Michael Brecht, a neuroscientist at Humboldt University in Berlin. It's also possible that individual neurons "just do their thing and contribute incrementally" to brain function, he says—that cells don't have to form these defined groups in order to collectively represent experiences. But future studies of precisely triggered neurons may yet resolve the role of ensembles, Brecht notes.

Deisseroth's group, meanwhile, activated larger sets of vertically or horizontally tuned neurons than in the Cell study, and evaluated whether mice could distinguish between the two possible perceptions. Using a newly discovered gene from a single-celled marine organism that produces a highly sensitive opsin, they found that zapping sets of roughly 10 to 20 cells that were tuned to one visual pattern or the other improved a mouse's ability to distinguish increasingly dim on-screen bars. Eventually, this stimulation alone prompted accurate "lick" or "don't lick" decisions.

It's impossible to know whether the mice really "saw" the absent bars, but both the behavioral tests and imaging suggest "the brain is doing what it does during natural perception," Deisseroth says.

"It's probably a bit too early" to claim that optogenetic stimulation can fully recreate real vision, which is much more complex than simple moving bars, says Valentina Emiliani, a physicist at a vision institute affiliated with CNRS, the French national research agency in Paris. Still, she says, it's exciting that hitting a few neurons can call up an entire pattern of brain activity related to vision.

The Deisseroth and Yuste labs now plan to use single-neuron optogenetics to find neurons underlying more complex behavior—including symptoms of brain disease. Yuste has launched experiments in mice that aim to reverse symptoms of schizophrenia and Alzheimer's disease by stimulating ensembles of neurons that don't activate as strongly in the diseased mice as healthy ones.