Scientists Destroyed a Few Dozen Neurons in a Fish and Watched Its Social Life Fall Apart
Researchers have captured the brain activity of a fish in the moments before it swims toward a companion, and the brain starts preparing that move several seconds before the fish does anything visible.
Published in Nature Communications, the study shows this is not a split-second reaction but a coordinated process playing out across multiple brain regions. Crucially, those patterns were specific to social situations. When researchers swapped the real companion for a moving dot, the predictive brain signature did not appear.
Scientists at the Hebrew University of Jerusalem pulled this off using zebrafish, tiny, nearly transparent fish whose brains can be imaged right through their skin. One fish was gently held in place with its tail free to move, while a companion swam freely nearby behind a clear barrier, close enough to see but unable to touch. A specialized microscope tracked more than 12,000 individual brain cells at once across 44 pairs of fish, with each session running 30 minutes. By reading those brain signals alone, researchers could tell whether a fish was about to approach its companion before any movement occurred, a predictive window rarely achieved in any animal.
What the Brain Does Before the Body Moves
Across those recordings, certain neurons in the midbrain and hindbrain became quieter several seconds before the fish moved toward its companion. At the same time, a small group of cells in a region called the pallium, part of the forebrain, became more active. When the fish was about to make a non-social movement instead, the pattern flipped.
A computer decoding those brain signals in any of the three regions could identify the upcoming social move with accuracy well above chance. Fish that moved in closer sync with their companions were also more likely to be heading toward them, and showed higher overall rates of social approach.
How early the brain signal appeared depended on what the companion was doing. When the companion had been swimming steadily in one direction for a while, the focal fish’s brain pattern shifted up to 10 seconds before any movement. When the companion had only just changed direction, the signal came much later. In other words, the fish’s brain wasn’t just tracking where its companion was. It was tracking how long it had been moving that way, and adjusting accordingly.
Variability between individual fish also showed up in their brain activity. Fish with more distinct neural patterns before approach movements had higher rates of actually approaching their companions, suggesting the clarity of the brain signal and the likelihood of social behavior are linked at the individual level. About 35 percent of fish in the study showed low approach rates, consistent with individual differences in social behavior observed in freely swimming zebrafish of the same age.
Removing Key Brain Cells Changes the Whole Network
To test whether the pallial cells were truly necessary for social behavior and not just correlated with it, the team used a laser to precisely destroy a small number of those cells, averaging around 46 neurons per fish, in nine fish. Movement ability and vision remained intact afterward, and neighboring cells were unaffected. Before the ablation, most fish preferred spending time near their companions. Afterward, social preference dropped significantly. A control group that underwent a sham procedure targeting a different, uninvolved brain region showed no change in social behavior whatsoever, ruling out the laser treatment itself as the cause.
After researchers re-imaged the brains of the treated fish during social interaction, the predictive brain signal had disappeared, not just in the pallium, but also in the midbrain and hindbrain. Removing a small cluster of cells in one region had dismantled the coordinated activity pattern across the entire network. Those pallial neurons weren’t just part of the system. They were necessary for it to work.