School of Danionella fish.
Something’s fishy: Danionella schooled alongside virtual counterparts when they looked and moved like real fish.
Add us as a Preferred Source on Google

Fish brains follow swimming friends to avoid foes

A subset of neurons in glassfish activate when the animals see others escape, new research shows, offering clues for how collective behaviors emerge from individual neural activity.

Swim away! Swim away! Unlike their cartoon counterparts, real fish can’t shout to warn others of danger. And yet when a predator is near, schools often manage to scatter all the same. So how do they share the message to escape?

The answer might lie in a group of visual neurons in the midbrain that respond to the actions of social partners, especially during an escape maneuver, according to a study of micro glassfish (Danionella cerebrum) published in Nature in September. The findings show how group actions can emerge from neural computations at the individual level.

Despite a wealth of descriptions and theories concerning collective behavior, to date, there is still little research on how it is implemented in the brain, says Andres Bendesky, associate professor of ecology, evolution and environmental biology at Columbia University, who was not involved in the study. This research showing the neural mechanisms underlying a collective behavior “is really a big step forward” in the field, he says.

The findings build on earlier work from the team showing that these tiny, transparent glassfish—no bigger than a fingernail—rely on vision to school and interact by copying neighbors’ movements. The new study examined how glassfish respond to danger—in the form of a dark sphere representing a looming predator—in different conditions: alone versus in a group and in light versus darkness. 

The fish escape faster together than they do alone, a phenomenon called “the many eyes” effect, the results show. And even when their direct perception of the threat is limited, they can still make a quick exit by observing when their companions react to danger. 

“Schooling is an interesting window to study a number of topics in neuroscience,” says study investigator Matthew Lovett-Barron, assistant professor of neurobiology at the University of California, San Diego. “Individuals in a group are constantly sensing and acting and deciding in response to each other.”

C

ollective behaviors have been challenging to study in the lab “for a pretty good reason,” Lovett-Barron says. “I mean, what are you going to do, put a microscope on the head of a bird and let it fly off with 1,000 other birds?” 

But technological innovations have opened up new possibilities for studying the neural activity of natural social behaviors in animals such as glassfish, he says. In this study, the team also created a virtual reality environment featuring realistic fish programmed using video game software. These virtual fish could scatter even without a looming stimulus present, enabling the researchers to see whether a live fish was truly relying on cues from peers rather than via direct threat detection.

The fish responded more naturally when their virtual counterparts moved with a lifelike burst-and-glide swimming motion rather than at a steady pace. “We were having a hard time getting our fish to interact with simple moving shapes,” Lovett-Barron says. But once the virtual fish looked and behaved like real glassfish, the live animals “loved them” and would school alongside, he says. The team then used two-photon calcium imaging to record activity across more than 3,000 neurons as the fish observed their virtual partners. 

In school: Micro glassfish (Danionella cerebrum) typically swim with a burst-and-glide motion.

In a field where this kind of neural-level study is still rare, “their setup is powerful,” says Herbert Wu, assistant professor of neuroscience at the Icahn School of Medicine at Mount Sinai, who was not involved in the work. With collective behavior, “it’s hard to find the right balance between natural behavior and controlled experimental conditions,” he says.

A subpopulation of neurons, primarily in the optic tectum (which corresponds to the superior colliculus in mammals) and thalamus, showed a heightened response to socially relevant actions such as swimming and escaping, the researchers found. These midbrain circuits, which are also found in mice and primates, may help these animals identify and respond to inherently meaningful actions within their group.

These neurons responded both when the virtual fish appeared to escape and when they disappeared at the end of the experiment. Lovett-Barron says he suspects that overlap reflects the murky freshwater streams of Myanmar, the glassfish’s native habitat, where visibility may not stretch much farther than a fish’s immediate neighbors. In this environment, escaping might not look that different from disappearing altogether. 

The results show the importance of considering behavior and neural computations in an animal’s evolutionary context, as well as experimenting with different model organisms, Lovett-Barron says. “It’s opened up many more questions than what we’ve answered.”

Sign up for our weekly newsletter.

Catch up on what you missed from our recent coverage, and get breaking news alerts.