AI System YORU Enables Real-Time Control of Fruit Fly Courtship Behavior
Researchers have developed YORU, an AI system that detects and interrupts fruit fly courtship songs by targeting specific neurons in real time.
Scientists at Nagoya University, in collaboration with Osaka University and Tohoku University, have developed an artificial intelligence system named YORU that can detect and interrupt fruit fly courtship songs by targeting specific neurons in real time.
In laboratory experiments, male fruit flies initiate courtship by extending their wings and producing a song. YORU identifies this behavior as it begins and promptly activates a light pulse that silences the neurons responsible for the song, effectively halting the courtship display.
YORU operates by recognizing entire postures as single behaviors within individual video frames, achieving an accuracy rate between 90% and 98% across various species, including flies, ants, and zebrafish. This rapid detection allows the system to act within approximately 31 milliseconds, swiftly enough to influence ongoing behaviors.
The system employs optogenetics, a technique that uses light to control genetically modified neurons. Upon detecting the wing extension associated with courtship, YORU triggers a green light pulse that silences the targeted neurons, leading to a decrease in mating success.
YORU's precision enables it to target individual flies even in crowded environments. By integrating location data with a projector, the system directs light specifically at the intended subject without affecting others nearby.
Beyond controlling behavior, YORU assists in interpreting brain activity by correlating observed behaviors with neural patterns. This capability aids researchers in distinguishing meaningful neural signals from background activity.
While YORU demonstrates significant advancements, it has limitations. Some behaviors that unfold over multiple frames may be missed by the single-frame detection approach. Additionally, hardware constraints introduce delays that could allow fast-moving animals to evade intervention.
The development of YORU represents a significant step in neuroscience, providing a tool to study the causal relationships between neural activity and behavior in real time. Future research aims to capture more complex behaviors and reduce hardware delays to enhance the system's accuracy and applicability in dynamic group settings.