A possible key to developing reliable navigation in autonomous vehicles may have been found through a window into the fruit fly’s brain. With the aid of live calcium imaging and two- photon microscopy, Northwestern University researchers recorded brain activity in neurons located at the base of the flies’ antennae that process external temperature signals. The team put Drosophila melanogaster (fruit flies) into little chambers and then adjusted the chambers’ temperature using hot or cold tiles. The researchers found that the insects consistently navigated away from spaces that reached a certain temperature or that varied too much from the 25 °C (77 °F) that the flies seemed to favor. When it comes down to it, it is an effective means of heating and cooling: Avoid an unwanted temperature altogether by moving away from it. Calcium indicators flow between the neural cells at the base of fruit flies’ antennae in response to temperature. Courtesy of Marco Gallio/Northwestern University. Marco Gallio, an associate professor of neurobiology in the Weinberg College of Arts and Sciences at Northwestern, said his team focused its research on G-CaMP, a green fluorescent protein that grows brighter when it binds with calcium. The flow of calcium between cells, along with the fluorescence that marks the flow’s occurrence, is an indicator of brain activity. The team recorded the behavior of Drosophila while the flies’ brains were also monitored. Gallio said that to make sure the researchers’ simulation of fly behavior was accurate, they specifically observed neural activity as temperature changes were initiated, to get a clear indication of what was driving increased communication within the brain. In the published study, the researchers said that the flies made a U-turn when faced with a drastically changed temperature in the compartments and stayed within a boundary governed by the 25 °C temperature level. This hot/cold “barrier” worked as well as any street sign. A thermal imaging system was used inside the chamber to track changes in the flies’ behavior and in their surroundings. The live calcium imaging and two-photon microscopy showed that the temperature of Drosophila receptor neurons responded to temperature changes of less than one degree. The flies were observed to turn away from the heat when they approached it, and they also seemed to effectively judge very slight temperature changes between their antennae, demonstrating their steering mechanism in real time. These findings challenge traditional assumptions about the way such small creatures interact with the environment. And the study may even be a precursor to learning how sensors could be designed to function in autonomous vehicles in the future, the researchers said. To test this theory, they designed a vehicle at fly scale. “These were in silico models, in the computer only,” Gallio said. The researchers used an algorithm that provided the vehicle with specifications that enabled it to match the avoidance behavior displayed by the insects when they approached a perceived threat. The study revealed that some of the models performed well enough — albeit guided by the lessons of machine learning, as opposed to living memory — that the models warranted further study. How would this avoidance capability potentially function in a real-world scenario? Gallio has thought of that: “IR thermal imaging can detect living things.” While the connection to the autonomous vehicle trend is an intriguing one, he said his team would keep the window open in its future work of studying the brain. “We are a neuroscience lab,” Gallio said. “We discovered [that] flies use complex flexible information processing to compute navigational decisions. We now want to understand how their brain does the math, computes decisions, and learns.”