How Do Dragonflies Catch Prey in Midair With Such High Accuracy?

16 neurons encode prey direction across a full 360° — and dragonflies catch over 95% of what they pursue, completing the entire hunt in roughly 200–500 milliseconds.

Dragonflies intercept moving prey in midair at a success rate exceeding 95%, using a network of just 16 target-selective descending neurons. That number is striking on its own. What makes it stranger is the mechanism underneath it: those neurons don’t track where prey currently is — they encode where it will be, generating a population vector that guides the wings to the future intercept point.

The whole sequence — visual lock, neural calculation, flight adjustment, capture — runs in roughly 200 to 500 milliseconds. The prey rarely detects the approach at all.

How the 16-Neuron Targeting System Actually Works

The 16 neurons are target-selective descending neurons — a closed-loop visual-motor system connecting the eyes to the wing control centers. During a hunt, dragonflies keep the prey image fixed on a specialized dorsal region of the compound eye. The descending neurons then produce a population vector encoding prey direction, which the wing system reads in real time to steer toward the intercept point rather than the prey’s current location.

All four wings adjust independently. The system continuously updates as the prey changes direction. Researchers tested this across a full 360° of simulated prey movement and found the directional encoding remained accurate throughout.

What “Predictive Interception” Means in Biological Terms

Simple reactive pursuit — following where something is — would fail at the speeds involved. Instead, dragonfly neural circuitry supports interception trajectories, meaning the flight path aims at where prey will arrive, not where it started. The PNAS study framing this describes the neurons as giving wing motor centers an accurate and reliable directional signal, rather than a moment-to-moment chase command.

UC Davis research adds a concrete speed detail: regardless of how fast a bead moved in experimental trials, dragonflies flew approximately one meter per second faster than the target throughout the pursuit. That consistent speed buffer is part of what makes the interception geometry work.

What a 95% Success Rate Reveals About Neural Efficiency

Aerial predators across the animal kingdom have been studied for targeting performance. Dragonflies consistently rank at the top for capture success — not because of size or raw speed, but because 16 neurons execute a reliable directional system with very low error rates. The head housing this system is smaller than a grape.

The efficiency here isn’t about brute processing power. It’s about a minimal circuit solving a geometric problem — one that scales across variable prey speeds, directions, and distances without requiring more neurons to handle edge cases.