Fish Brains Copy Neighbors’ Escape Moves to Flee Predators, Study Finds

Fish brains copy their neighbors‘ escape moves to flee predators, a study published in Nature in September shows — researchers watched more than 3,000 neurons in micro glassfish fire as the animals tracked companions’ swimming to dodge threats they couldn’t see themselves, pinpointing midbrain visual neurons behind collective escape.

The work addresses a problem neuroscientists have circled for years: collective behavior is everywhere in nature, but there is still little research on how it is implemented in the brain. As one outside expert put it, this study moves the field from describing the swarm to explaining it.

A fish the size of a fingernail

The subjects are Danionella cerebrum, micro glassfish no bigger than a fingernail. They’re transparent, which matters: the same optical clarity that lets hobbyists see the fish’s skeleton lets researchers image its brain while it behaves. Earlier work from the same team showed the fish rely on vision to school and interact by copying neighbors’ movements.

The new study pushed further. Researchers presented the glassfish with a dark sphere looming overhead — a stand-in for a predator — and tracked how the fish responded alone versus in a group, and in light versus darkness.

Two findings stand out. Fish escape faster together than alone, an effect ecologists call “the many eyes” effect. And even when a fish’s own view of the threat is blocked, it still makes a fast exit — by watching when its companions bolt.

Virtual fish built with game software

To prove the fish were reading social cues rather than sensing the threat directly, the team built a virtual reality environment. The virtual fish were programmed using video game software, and live glassfish schooled alongside them.

When the animations looked wrong

Not everything worked on the first pass. “We were having a hard time getting our fish to interact with simple moving shapes,” Matthew Lovett-Barron, assistant professor of neurobiology at the University of California, San Diego, said. But once the virtual fish moved with a lifelike burst-and-glide swimming pattern, the fish “loved them.”

Even with no looming stimulus present, realistic virtual fish could trigger a scatter — confirming the live fish were swimming off purely on the actions of their peers. That had the makings of a clean experimental readout.

With the VR rig dialed in, the team ran two-photon calcium imaging across more than 3,000 neurons while fish watched their virtual partners. The recording revealed a group of visual neurons in the midbrain that respond to the actions of social partners, especially during escape maneuvers. That region, per the study, is where the social signal appears to be computed.

“A really big step forward”

The reaction from researchers outside the study points at why the result matters beyond fish.

Andres Bendesky, associate professor of ecology, evolution and environmental biology at Columbia University, who was not involved in the study, said it shows the neural mechanisms underlying a collective behavior, calling that “really a big step forward” for a field with little research so far on how collective behavior is implemented in the brain.

For decades, descriptions and theories of collective behavior have outpaced the biology underneath them. The gap has been practical, not conceptual: recording single-neuron activity in a freely interacting animal is difficult. Lovett-Barron framed the constraint plainly. “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?” he said.

The glassfish workaround — tiny animals, virtual companions, a fixed imaging window — is a template other labs can copy. “Individuals in a group are constantly sensing and acting and deciding in response to each other,” Lovett-Barron said, adding that schooling offers “an interesting window” onto that loop.

Why AI and robotics engineers are watching

There’s a second audience for this paper, and it isn’t biologists.

Until then, the mechanism is a correlation with a very compelling animated fish.

“Schooling is an interesting window to study a number of topics in neuroscience,” Lovett-Barron said.

Photo of author

Technology Desk

Technology Desk at Archyde.

Surrey PCC Warns of AI Voice Scams Ahead of Christmas

Russian Representatives Present Document Demanding NATO Withdraw to 1997 Positions; Eastern Leaders Reject