Whole-body 3D kinematics of freely behaving Drosophila

To understand how a nervous system turns a simple electrical signal into a complex, coordinated movement, scientists must first be able to see that movement with absolute clarity. For decades, researchers have relied on the fruit fly, a tiny insect with a well-mapped brain, to uncover the rules of motor control. The challenge has always been that these flies are small, move with blinding speed, and often block their own bodies from view as they turn or interact with others. Traditional methods of tracking them have struggled to capture their full motion in three dimensions, leaving a gap between what the brain is sending and how the body actually responds. Without a complete picture of the fly's posture in space, it remains difficult to build accurate models of how the brain commands the muscles to create behavior.

A team of researchers has now bridged this gap by creating a system that watches flies move in full 3D without ever touching them or attaching markers to their bodies. They set up seven high-speed cameras that work together to record the insects at 800 frames per second, capturing every rapid twitch and turn. To make sense of this flood of visual data, they trained a computer program to recognize and track fifty specific points on the fly's body, from the tips of its legs to the ends of its antennae. Because the computer's initial guesses can sometimes look physically impossible—like a leg bending the wrong way—the researchers added a second step. They used a model of the fly's actual anatomy to correct these errors, ensuring that every movement the system recorded was something a real fly could physically do. This process, known as retargeting, transforms raw video into a precise, anatomically correct map of the fly's motion.

When the scientists analyzed the movements of flies running on the ground, they found something that contradicts a long-held assumption about how insects move. For a long time, it was believed that animals switch between distinct styles of running, or gaits, as they speed up or slow down. The data from this study shows that flies do not do this. Instead, they maintain a single, grounded running style across their entire range of speeds, never transitioning into a different gait. This finding suggests that the control mechanisms for their movement are more continuous and fluid than previously thought.

The system also proved powerful enough to track multiple flies at once, allowing the researchers to observe social interactions with unprecedented detail. When watching a male fly court a female, the study revealed that the male coordinates the movement of both wings to produce his courtship song. At the same time, he constantly adjusts the angle of his body, tilting his head and torso up or down to keep his eyes locked on the female's vertical position as she moves. These observations provide a rich, three-dimensional record of behavior that was previously impossible to capture. By making this large dataset and the software used to create it available to the public, the researchers have provided a new foundation for scientists to build detailed models of how the brain and body work together to generate movement.