Fruit flies use memory to track odors

When bats use echolocation to find an object, they don’t point their sonar beam directly at the target, where the intensity of the signal bouncing back would be the strongest. Instead, they aim slightly off axis, so the returning beam contains sharper signal differences. The research team that observed this in 2010 predicted that the same strategy would apply to scent tracking. 

That prediction was correct, a paper published in July in Nature shows. When fruit flies catch a whiff of apple cider vinegar, they zigzag along the edge of the odor plume, where the concentration difference is sharpest, rather than traveling through the middle, where a stronger concentration is likely to hold steady. 

“The edge of the plume is potentially where some of the most information might be stored,” says Marie Suver, assistant professor of biological sciences at Vanderbilt University, who was not involved in the work. “Whereas if you’re in the middle of the plume, you’ll be getting more packets of odor, but it’s not as stark of a concentration gradient as at the edge.”

Keeping tabs on a plume is also more complex than researchers previously thought. When flies and other insects first encounter an odor, they surge upwind and cast side to side when they lose the trail—a behavior that seemed to be a simple reflex, says Matthieu Louis, associate professor of molecular, cellular and developmental biology at the University of California, Santa Barbara, who was not involved in the study. 

“It was supposed to be a memoryless system,” says study investigator Vanessa Ruta, professor and head of the Laboratory of Neurophysiology and Behavior at Rockefeller University. “Basically, all the animal needed to know was the exact sensory experience and information it had at that one moment, and nothing about its prior history would be relevant.”

Instead, flies can store a memory about the angle their path must take to return to the scent. Neurons in the central complex, the navigational hub of the fly brain, contribute to the memory, imaging experiments show. 

T

So Ruta and her team created a virtual-reality setup that provided those details. In the paradigm, a tethered fruit fly walks on a foam ball about 6 millimeters across that functions as a tiny treadmill. As the fly turns itself in different directions, a tube delivering a stream of air rotates around it, simulating the wind. The researchers created odor plumes with customizable geometries and concentrations by adjusting the amount of apple cider vinegar vapor added to the air stream.

“Stimulus control with odors is a nightmare. I really hate it. But they cleverly designed their system to control the timing of odor concentration dynamics with high precision,” says Matt Smear, associate professor of neuroscience and psychology at the University of Oregon, who was not involved in the work. 

Almost immediately, the team noticed that the flies stuck to the edge of the plume, Ruta says. “It was a very striking behavior, and very robust.” 

The flies tracked the plume’s edge irrespective of how the odor concentration changed as the fly kept walking. The behavior persisted when the plume’s trajectory tilted away from the wind direction or even ran perpendicular to it. “That was one of the wildest things for me,” Suver says. To pull this off, the flies must remember the angle to travel when returning to the plume, modeling experiments showed. 

The memory is indeed directional and not positional: When tracking a jumping plume that shifts 20 millimeters away whenever the flies leave it, the flies walked past the plume’s old location and kept trekking in the direction that they expected would lead them to their goal. The researchers could even rewrite an entry-angle memory by delivering a whiff of vinegar when the flies spontaneously walked in the direction of the new plume. 

A population of neurons in the central complex called FC2 signals the direction flies should walk just before they turn back toward the plume, imaging experiments showed. Silencing the FC2 neurons impaired the edge-tracking behavior.

“I’m convinced that they’re using a memory to go back to where they last encountered the odor,” Smear says. 

Follow-up work should explore when edge tracking fails and which strategies take over, says Tobias Ackels, group leader at the University of Bonn, who was not involved in the study. “What happens when the signal gets noisy?” Other work could examine if flying, which adds a vertical component to plumes, requires a different tracking approach than walking, and if flies handle non-food-related and aversive odors in the same ways, Suver says. 

It will also be interesting to see if these findings hold up in a freely moving fly, Louis says. The study is “a very solid basis on which to build.”