One frog species finds a solution to the cocktail party problem

When people attend large, noisy gatherings, hearing a conversation is often difficult. The so-called cocktail party problem also affects the green tree frog (pictured in figure 1) and other frogs in habitats teeming with the sounds of various animals and with anthropogenic sources of noise. Environmental background noise can reach volumes of 60–80 dB (relative to the standard sound-pressure-level reference), about the same amplitude as a vacuum cleaner. To be heard, tree frogs produce loud calls, often 100 dB in amplitude from 1 m away, an order of magnitude louder than the background noise.

Figure 1.

Against the cacophony, frogs and other animals may, for example, incorrectly identify species-specific call patterns or struggle to locate sound sources. In exceptionally loud environments, other species abandon acoustic communication altogether and resort to visual cues, such as waving their legs. (For more about acoustic biology, see the article by Megan McKenna, Physics Today, January 2020, page 28 .)

If your goal is to study evolutionary adaptations that improve the signal-to-noise ratio of incoming acoustic information, frogs make interesting subjects. For about 200 million years, they’ve been successfully vocalizing in noisy environments. ¹ They and other amphibians have evolved a lung-to-ear sound-transmission pathway. The inset of figure 1 shows the airborne routes by which sound waves reach the internal surface of the tree frog’s eardrum, or tympanum: directly through the eustachian tubes connected to the animal’s mouth and the more meandering path from the lungs via the glottis opening.

Researchers have known about frogs’ lung-to-ear anatomy for at least 30 years and have been studying how lungs may modulate the acoustic signals that reach the eardrum. Previous studies have shown that some frogs improve the directional sensitivity of a single eardrum at a select frequency range by inflating their lungs. ²

Rather than solving the cocktail party problem by making their own calls louder than the background noise, at least one frog species is capable of improving their communication by reducing the volume of noise around them. Norman Lee of St Olaf College in Minnesota, Mark Bee of the University of Minnesota, and their colleagues have now found that when green tree frogs inflate their lungs, incoming acoustic signals from a biologically noisy frequency band are attenuated. ³ The practice enables female tree frogs with inflated lungs to better hear the mating calls of their male suitors.

Picking up vibrations

In previous studies, researchers sought to understand if the lungs could control sound-source localization in frogs. Most experiments explored how incoming sounds may affect the directional sensitivity of a single eardrum. Lee and his colleagues also wanted to study directionality and to tease out the contributions of sound impinging on the external and internal surfaces of a frog’s ear. So they positioned a frequency-modulated acoustic stimulus at various locations around the frog and analyzed the amplitude of the resulting vibrations at the eardrum.

To measure whether having the lungs inflated affected the animal’s ear vibrations, Lee and his colleagues used laser Doppler vibrometry. The method works by shining a laser on targets—in this case the frog’s eardrum and the external body mass over the lungs—and then recording with a high-precision interferometer the frequency of the light that’s scattered from the vibrating targets. The superposition of the incident laser and a reference laser beam aimed at a photodetector are then used to calculate the Doppler shift.

The results for a single eardrum showed that the lungs contributed a small improvement to the frog’s directional hearing. But like humans and other vertebrates, frogs locate sources of sound by, among other ways, using the difference in loudness between the left and right ears (see the article by Bill Hartmann, Physics Today, November 1999, page 24 ).

When the data were analyzed from an organism-level perspective with two ears, the researchers found that the lung-inflation improvements in directionality at each eardrum canceled each other out. “That result was very striking,” says Lee. “We thought if this lung input has nothing to do with improving directionality, then what might it be?”

Lowering the volume

To learn more, Lee and his colleagues analyzed the amplitude of the eardrum’s vibrations at various frequencies and sound-incidence angles. “When we looked at the lungs’ inflated and deflated states and took the difference, those plots really showed a pattern that was striking to us,” says Lee. Figure 2a shows that difference: There’s a pronounced loss in eardrum sensitivity between 1.4 kHz and 2.2 kHz. Rather than finding that the tree frogs used inflated lungs to amplify their own calls, the researchers found that the lung inflation reduces the volume of noise from other frog species.

Figure 2.

An analysis of a data set of choruses of many species collected by the North American Amphibian Monitoring Program shows that many frogs use that frequency band for mating calls, even though the 1.4–2.2 kHz band is quite noisy. Figure 2b shows that the mating calls of green tree frogs have two local maximums in amplitude outside the range, at 834 Hz and 2730 Hz. Although the amplitude of the mating calls of green tree frogs peak on either side of the frequency range, the nearby noise makes finding a potential mate challenging.

That evolutionary pressure may have caused green tree frogs—and perhaps other frog species—to develop the noise-reducing adaptation. The inflated lungs of the tree frogs vibrate most strongly at the resonance frequency in the middle of the noisy 1.4–2.2 kHz band, which could help attenuate the noise there.

When comparing the new results with data from three other frog families that have a common 155-million-year-old ancestor, Lee and his colleagues found that all the frogs have a similar acoustic relationship: The lung resonance frequency is in a band that’s bounded by mating-call spectral peaks on either side. Whether and how a frog’s body size affects the usefulness of lung-mediated noise reduction remains to be answered.

Noise-canceling headphones

Lee and his colleagues have begun to investigate the mechanism behind the lung-inflation behavior with a generalized frog-physiology model. Amphibian ears have two sensory organs that process acoustic signals—the amphibian papilla and the basilar papilla. Each contains hair cells that mechanically transduce vibrations to electrical signals in auditory nerve fibers tuned to respond to specific frequencies. After simulating the frequency tuning of 161 individual fibers, the researchers found that the greatest reduction in eardrum sensitivity corresponds to the frequency range at which the amphibian and basilar papillae overlap.

The vibrations from the inflated lungs may decrease the response of the nerve fibers in that overlapped frequency range. Inflated lungs may also limit the possibility of two-tone rate suppression, in which sound in a particular frequency range can suppress auditory responses to sound at a lower frequency range. In tree frogs, sound energy from 1.4 kHz to 2.2 kHz can suppress the response of nerve fibers tuned to the mating calls heard at lower frequencies.

Lee and his colleagues aren’t entirely sure yet of the precise relationship between the inflated lungs and the reduction of the eardrum response at a select frequency range, but they suspect it may be similar to the destructive interference used in noise-canceling headphones. The technology uses small microphones to record external ambient noise; signal processors then emit out-of-phase sound waves that destructively interfere with the incoming noise.

The researchers hypothesize that for tree frogs, the inflated lungs act analogously to the headphones’ microphones. Then the phase of the acoustic signal is somehow modulated inside the body so that it arrives at the eardrum out of phase with the sound frequencies impinging on the frog’s body. The situation is a bit more complicated because of another sound source, which arrives at the eardrum internally via the opposite ear, but the researchers are keen to figure it out. Lee says, “We think we can get at this question with a combination of additional laser measurements and modeling.”