Mimicking the Super Hearing of a Cricket-Hunting Fly

Researchers build a silicon version of nature’s most acute sound-location system

3 min read
Mimicking the Super Hearing of a Cricket-Hunting Fly
Pivoting acoustical sensor imitates the super-acute hearing of the parasitoid fly, Ormia ochracea
Photo: N. Hall/UT Austin

Ormia ochracea is a little, yellow fly of the American south whose breeding strategy has an outsize ick factor. It deposits its larvae on the bodies of male crickets. The larvae then eat their way into their unwilling hosts, and devour them from the inside.

What is most remarkable, though, is that the female fly locates the crickets by sound, homing in on the he-cricket’s stridulations (the chirping that results from the wings rubbing together) with uncanny accuracy. The cricket’s chirp is a smear of sound across the scale from the 5 kilohertz carrier frequency to around 20 kHz. And, as anybody who has tried to evict a passionate cricket from a tent or cabin knows, the sound is maddeningly hard to pinpoint.

With an auditory apparatus—let’s call them ears—only 1.5 millimeter across, ochracea pulls off a major feat of acoustic location; a number of engineering groups are working on devices to duplicate the fly’s sensitivity.

Now, a team at the University of Texas at Austin has built a prototype replica of O. ochracea’s ear. Michael L. Kuntzman and Neal A. Hall, researchers in the school’s electrical and computer engineering department, describe the device and its performance in Applied Physics Letters.

The fly’s ears are very different from those of humans. Human ears are typically separated by about 21 to 22 centimeters—about 625 microseconds apart at the speed of sound (though of course it varies with temperature and humidity). We judge sound direction by assessing ear-to-ear differences in phase and volume. We can distinguish time differences of as little as 10 microseconds, and the phase-difference calculation is useful mainly for lower-frequency sounds—those with wavelengths longer than 21 cm. In this range, we can locate a sound source to within about 1 degree if it is dead ahead, or 15 degrees if it is off to the side.  As frequencies rise above about 1600 Hz (a very sharp G above high C), the wavelength is shorter than the ear-to-ear separation, and we fall back on using the volume difference alone to approximate the source’s position.

The fly’s ear, on the other hand, is 4.3 microseconds wide at the speed of sound, and it can distinguish phase delays much smaller than that in sounds coming in from nearly ahead or behind.

The secret is the physiology of the ochracean ear, whose centerpiece is an elastic plate that pivots on a central support. This structure responds to incoming sound waves by resonating in two distinct modes. You can picture them, albeit on an obviously much larger scale, by standing with your arms outstretched to either side. First, raise one arm while lowering the other, like the ends of a see-saw; that’s the first mode. Now move your hands up and down together, flapping like a bird; that’s the second mode.

In the fly’s ear—and in the Kuntzman-Hall device—each mode responds to a different parameter of the incoming wave. The see-saw action responds only to the x-component of the incoming pressure gradient, indicating, say, a how a high-pressure compression crest at the tips of your fingers shades into a low-pressure trough at your elbow. It tracks the changes only along the one dimension of your arm, though, and reveals nothing about the omnidirectional strength and structure of the wave. The flapping mode, on the other hand, responds only to omnidirectional pressure—the sound volume, for example, or the overall pressure on your body—and reveals nothing about the wave's direction. In the fly's ear, both modes superpose to create a composite displacement of the membrane, so the trick is to break this signal down into its see-saw and flapping components..

The fly can individually quantify the displacements of the right and left sides of its pivoting-beam auditory membrane. Then the fly's neural network subtracts the displacement of the left-side channel from the displacement of the right channel to extract the first-mode see-saw signal; this shows the incident angle of the incoming sound. At the same time, the fly's brain adds the left- and right-channel signals to yield the second-mode flapping displacement. This reveals the omnidirectional sound pressure (a clue to distance).

The UT researchers etched a spring-loaded, 1.5-mm-by-2.5-mm pivoting beam into silicon, with a lead-zirconate-titanate piezoelectric film painted on the supporting springs to sense displacement. In experiments, Kuntzman and Hall have read and analyzed the output just as the fly’s brain does. The prototype can resolve the direction of high-frequency sound sources to within 0.35 degrees for sounds in its directional “sweet spots,” and to within about 6 degrees in its less sensitive zones. (The imprecision, their paper says, is mainly due to some imperfections or asymmetries in the prototype.)

“Synthesizing the special mechanism with piezoelectric readout is a big step forward towards commercialization of the technology," said Hall, an assistant professor. There are sure to be defense applications—after all, the research is funded by the U.S. Department of Defense's Advanced Research Projects Agency (DARPA)—as well as potential for commercial products like hearing aids. 

The Conversation (0)