A school of fish can look like a single, flowing organism—dozens of bodies turning and aligning in near-perfect synchrony. But that seamless choreography emerges from many individual animals making split-second decisions together.
New research from the MCB lab of Florian Engert, published in Current Biology, (PDF) reveals that in medaka fish (Oryzias latipes), these social decisions depend on distinct movement “states” hidden within what appears to be smooth, continuous swimming. The study’s first author, Roy Harpaz, postdoctoral fellow in the Engert lab, and colleagues show that young medaka form coordinated groups early in life—and that their sensitivity to neighbors shifts depending on how they are moving at any given moment.
Schooling Starts Early
By tracking groups of medaka from larval stages through early development, the researchers found that synchronized swimming formations begin to appear around two weeks of age. Within about a month, these collective patterns stabilize and resemble adult-level schooling.
This rapid timeline positions medaka as a powerful system for studying how social behaviors develop. Because the species is genetically tractable and compatible with modern brain-imaging tools, scientists can potentially link changes in group behavior to specific neural circuits.
“We were surprised by how early robust social alignment emerges,” Harpaz says. “By just two weeks of age, you already see coordinated group behavior that stabilizes very quickly.”
Finding Structure in Continuous Motion
Unlike zebrafish and many other fish that swim in short bursts followed by gliding (“burst-and-coast” swimming), medaka move continuously, propelling themselves with steady body undulations. At first glance, their motion seems seamless.
But a closer look revealed a hidden structure. By analyzing subtle changes in swimming speed, the team showed that continuous motion naturally separates into three distinct kinematic states: acceleration, deceleration, and constant-speed swimming. These states are not just mechanical descriptions. They shape how fish process social information.
Using computational models, the researchers examined how medaka respond to nearby fish in each state. The “rules” governing social interactions shifted depending on whether a fish was accelerating, decelerating, or cruising.
“Even though medaka look like they’re swimming continuously, their behavior falls into discrete modes,” says Harpaz. “And each mode comes with a different way of responding to neighbors.”
When Are Fish Most Attuned to Others?
Perhaps the most unexpected finding was that medaka are most socially responsive during constant-speed swimming.
Fish were moderately responsive to neighbors while accelerating and least responsive while decelerating. But when maintaining a steady speed, they showed the strongest tendency to adjust their movement in response to nearby fish.
Previous work in other species suggested that behavioral responses occur mainly during active speed changes. The discovery that responsiveness peaks during steady swimming adds a new dimension to our understanding of collective behavior.
“We expected responses to happen mainly during accelerations,” Harpaz explains. “Instead, we found that the strongest sensitivity to social information occurs during prolonged constant-speed swimming.”
The researchers speculate that steady swimming may allow more efficient sensorimotor processing, enhancing a fish’s ability to integrate social cues. Future work will test whether this heightened responsiveness is specific to social signals or reflects broader sensory sensitivity.
Individual Variation, Collective Order
Notably, fish did not synchronize their movement states. One individual might be accelerating while another cruises or slows down. The timing of these transitions was largely uncorrelated across the group.
Yet collective alignment still emerged. Group-level patterns depended less on synchronized state changes and more on how frequently individuals occupied each state overall. Even with noisy, independent individual dynamics, stable schooling behavior arose.
“Collective organization doesn’t require the fish to coordinate their internal states,” says Harpaz. Even with uncorrelated state transitions, the group still aligns.”
This supports theoretical models suggesting that complex group patterns can emerge from noisy, desynchronized state-dependent rules at the individual level.
A Platform for Linking Brain and Behavior
The team adapted computational frameworks previously used in zebrafish and found that they successfully captured medaka social responses despite clear differences in movement strategy. The models revealed distinct patterns of sensitivity to neighbors—effectively different “social receptive fields”—for each kinematic state.
With medaka increasingly used for brain-wide calcium imaging and optogenetics, the study lays important groundwork for linking neural activity to collective social behavior.
“By identifying discrete movement states that shape social interactions, we now have a framework for asking how specific neural circuits control collective behavior,” adds Harpaz.
What looks like a seamless group dance, the researchers show, is built from hidden shifts in movement mode—and state-dependent computations unfolding in the brain.
(PDF)
