Scientists Predict Fish Behavior from Real-Time Brain Monitoring
Scientists Predict Fish Behavior from Real-Time Brain Monitoring

Scientists Predict Fish Behavior from Real-Time Brain Monitoring

Researchers could anticipate which way an eight-day-old zebrafish will flick its tail based on brain activity 10 seconds earlier.

Abby Olena
Jan 22, 2020

ABOVE: An artistic representation of the two brain states—leading to the correct or incorrect choice of directional movement—observed in larval zebrafish during decision-making
ALIPASHA VAZIRI

Understanding the array of neural signals that occur as an organism makes a decision is a challenge. To tackle it, the authors of a study published last week (January 16) in Cell imaged large swaths of the larval zebrafish brain as the animals decided which way to move their tails to avoid an undesirable situation. Finding patterns in the data, they were then able to use imaging to predict—10 seconds in advance—the timing and direction of the fish’s movement.

“In a lot of other model systems it’s really difficult to actually . . . record something that’s happening throughout the whole brain with a high level of precision,” says Kristen Severi, a biologist at the New Jersey Institute of Technology who was not involved in the study. “When you have something like a larval zebrafish where you have access to the entire brain with single-cell resolution in a transparent vertebrate, it’s a great place to start to try to look for activity patterns that might be distributed and might be hard to connect.”

Even if an animal has learned to do something, it doesn’t execute the exact same motor responses every time, says biophysicist Alipasha Vaziri of the Rockefeller University. He adds that common approaches to studying the neural basis of decision-making may not tell the whole story. For instance, monitoring a handful of neurons and then extrapolating from their activity what’s happening brain-wide means that researchers might miss the big picture. Likewise, recording across the whole brain and then averaging results across trials risks losing details essential to understanding how the brain encodes this behavior. 

To overcome these issues, Vaziri and colleagues used a calcium-imaging technique that enabled them to monitor neural activity simultaneously from about 5,000 individual neurons throughout the brains of eight-day-old zebrafish. To study how brain regions interact to drive goal-directed behavior, they exposed zebrafish larvae to heat, which the fish do not like, and trained them to flick their tails either left or right, at which point the researchers turned the heat off.

During individual trials, the team found that the animals’ brains exhibited one pattern of activity that showed up about 10 seconds before they were going to twitch their tails to the correct side and a different pattern for the incorrect side. The researchers could predict which choice the fish would make and when in each trial with about 80 percent accuracy. This activity showed up across the brain, including in the telencephalon, habenula, and cerebellum.

“This study highlights an important point: even simple behaviors like making a goal-directed movement engage widespread activity across many brain areas,” Nuo Li, a neuroscientist at Baylor College of Medicine who was not involved in the work, writes in an email to The Scientist. “The process may require coordinated communications across different parts of the brain. We are still at the beginning stage of understanding how different brain regions communications with each other.”

The role of the cerebellum in decision-making

There was activity across the whole brain in various regions, but “we think the major contributors are so-called granular cells in the cerebellum,” says Vaziri. This finding surprised them because the cerebellum is mostly known for its control of motor function, not decision-making.

The researchers used lasers to lesion the cerebellum after training, and the fish made the correct choice less often and also took longer to decide.

Other neuroscientists question whether that means the cerebellum is involved in decision-making, or if the behavior following brain injury could simply reflect motor dysfunction.

“They take out whole chunks of the brain and the deficits are broad. Then [they] make very specific claims about the function of these brain areas in the cognitive task that they are studying. When you ablate the cerebellum, you see motor deficits, so the slow decision-making could be related to the motor function of the cerebellum,” cautions Herwig Baier, the director of the Max Planck Institute of Neurobiology in Germany who did not participate in the study. He adds that the authors are recording from a slab of nervous tissue that is in the middle of the brain. “Since the paper is about the cerebellum largely, which is on the surface on the brain, I feel a bit uneasy about the interpretation that they are really recording from the cerebellum.”

Baier says that the authors could have both confirmed the anatomy and lesioned the region much more precisely using transgenic fish lines with cerebellar neurons labeled, which would help clear up the role of that brain region.

Vaziri acknowledges that understanding the causal relationships of this interconnected circuitry is still an open question. He says his group plans to continue investigating the involvement of the cerebellum in decision-making. “We would need to not only be imaging but then doing causal interference, for instance, [with] optogenetics.”

Q. Lin et al., “Cerebellar neurodynamics predict decision timing and outcome on the single-trial level,” Celldoi:10.1016/j.cell.2019.12.018, 2020. 

Abby Olena is a freelance journalist based in Alabama. Find her on Twitter @abbyolena.