Acoustic Masking in Mammals

There are concerns about how anthropogenic sounds may affect hearing and communication in marine mammals. One potential effect is acoustic masking in which sounds (masking sounds, or maskers) interfere with an animal’s ability to perceive, detect, or discriminate a different sound[1]Fletcher, H., & Munson, W. A. (1937). Relation between loudness and masking. The Journal of the Acoustical Society of America, 9(1), 78–78. https://doi.org/10.1121/1.1902030.. The closer the frequency content, sound level, and duration of the potential masker to the signal of interest, the higher the probability it will interfere with an animal detecting the signal of interest. Masking is influenced by other factors as well; the reader is referred to How do Marine Mammals Hear?, and the discussions of the potential effects of masking in marine mammals and masking in fishes.

The degree of acoustic masking depends upon the animal’s ability to discriminate a sound. Detection of a signal depends primarily on some characteristics of the mammalian cochlea, which vary by species, and how it encodes acoustic signals.

The cochlea is a spiral-shaped chamber within the inner ear that transforms sound waves into nerve impulses. It is considered “the organ of hearing.” (Diagram from the Handbook for Acoustic Ecology, CD-ROM edition, B. Truax, ed., Cambridge Street Publishing, 1999. www.sfu.ca/~truax/csr.html)

The ear is the hearing organ in humans. It consists of the outer ear (pinna and auditory meatus), the middle ear (ossicles) and the inner ear (cochlea and vestibular system). Courtesy of Andrew Wright, University of Ulster.

The cochlea is a fluid-filled, spiral labyrinth that houses many structures related to hearing, including the basilar membrane and its associated hair cells. As sound moves along the cochlea, portions of the basilar membrane are displaced, and the cilia on top of the hair cells are bent, triggering the release of chemicals that result in a neural impulse that goes to the brain. The basilar membrane is tonotopically organized. Parts of the membrane have different resonance properties based on differences in the anatomy of the basilar membrane along its length. The portion of the basilar membrane that is closest to the oval window (base of the cochlea) is narrow, thick, and stiff. Therefore, it responds best to high frequencies. The membrane becomes broader, thinner, and more elastic as you move farther up the cochlea. Thus, the location near the apex of the cochlea responds best to lower frequencies.

Animation of the basilar membrane being displaced and the cilia on top of the hair cells being bent. Animation by Dr. Inna Belyantseva, NIH/NIDCD, used with permission.

The lowest sound level at any given frequency that an animal can detect in the absence of other sound is its detection threshold. Measurements of detection thresholds at multiple frequencies are plotted together to present a hearing threshold curve or audiogram, showing how the hearing sensitivity of an animal varies across its hearing range. Tubelli et al. have compiled measured audiograms (https://andrewtubelli.com/whales/, see examples in the figure below).

This image is a screen shot of the interactive graphs available on https://andrewtubelli.com/whales/. This image shows a collection of harbor porpoise audiograms that were experimentally measured. On the website created by Tubelli et al. the user can select a species to see the audiograms of individual animals within that species. Image from http://andrewtubelli.com/whales/.

 

This image is a screen shot of the interactive graphs available on https://andrewtubelli.com/whales/. This image shows a collection of bottle nose dolphin audiograms that were experimentally measured. On the website created by Tubelli et al. the user can select a species to see the audiograms of individual animals within that species. Image from http://andrewtubelli.com/whales/.

Both natural and anthropogenic sounds may be acoustic maskers, but it is challenging to define a specific measurement of how one sound masks another. Two metrics have been useful in quantifying masking: critical bandwidth and critical ratios. The next section discusses how they are measured and/or estimated, and how they may inform our understanding of masking.

Frequency selectivity is the ability of an animal to discern the frequency of one sound in the presence of sounds of different frequencies [2]Yost, W. A., & Shofner, W. P. (2009). Critical bands and critical ratios in animal psychoacoustics: An example using chinchilla data. The Journal of the Acoustical Society of America, 125(1), 315–323. https://doi.org/10.1121/1.3037232.. The ability for any sound to mask another depends on the relative levels of the sounds and their proximities of their response locations along the basilar membrane. One method for estimating frequency selectivity is the critical bandwidth (Fletcher 1940). The critical bandwidth, which has units of frequency (Hz), is determined by increasing the bandwidth of the masking sound while measuring the detection threshold of a test signal at the center frequency of the masker. As the bandwidth of the masker increases, the test signal detection threshold will also increase (meaning it is harder to detect). Eventually, the test signal detection threshold will stop increasing, even though the bandwidth of the masker continues to increase. This point is the critical bandwidth.

A) As the bandwidth of the masking sound increases, the ability of the animal to detect the tonal signal increases until it reaches an inflection point, which is called the critical bandwidth. B) Three examples where the detection threshold of the signal of interest (blue) was measured as the bandwidth of the masking sound (yellow) increased. Figure by DOSITS.

Sounds within the critical bandwidth are able to mask other sounds within that bandwidth. Sounds at frequencies higher or lower than a critical bandwidth do not mask sounds within that critical bandwidth. One hypothesis is that the mammalian auditory system can be approximated as a set of overlapping bandpass filters , each with a width equal to the corresponding critical bandwidth.

Estimating critical bandwidths can be difficult, requiring trained animals held in captivity to respond when they detect very quiet sounds; it can be particularly difficult for marine mammals that are usually tested in the water in the presence of ambient sound. For marine mammals, physiological and behavioral measures of critical bandwidths have been completed under water for species such as bottlenose dolphin[3]Au, W. W. L., & Moore, P. W. B. (1990). Critical ratio and critical bandwidth for the Atlantic bottlenose dolphin. The Journal of the Acoustical Society of America, 88(3), 1635–1638. https://doi.org/10.1121/1.400323., harbor porpoise[4]Popov, V. V., Supin, A. Ya., Wang, D., & Wang, K. (2006). Nonconstant quality of auditory filters in the porpoises, Phocoena phocoena and Neophocaena phocaenoides (Cetacea, Phocoenidae). The Journal of the Acoustical Society of America, 119(5), 3173–3180. https://doi.org/10.1121/1.2184290., California sea lion[5]Southall, B. L., Schusterman, R. J., & Kastak, D. (2003). Auditory masking in three pinnipeds: Aerial critical ratios and direct critical bandwidth measurements. The Journal of the Acoustical Society of America, 114(3), 1660–1666. https://doi.org/10.1121/1.1587733., harbor seal[6]Turnbull, S. D., & Terhune, J. M. (1990). White noise and pure tone masking of pure tone thresholds of a harbour seal listening in air and underwater. Canadian Journal of Zoology, 68(10), 2090–2097. https://doi.org/10.1139/z90-291.[7]Southall, B. L., Schusterman, R. J., & Kastak, D. (2003). Auditory masking in three pinnipeds: Aerial critical ratios and direct critical bandwidth measurements. The Journal of the Acoustical Society of America, 114(3), 1660–1666. https://doi.org/10.1121/1.1587733., and northern elephant seal[8]Southall, B. L., Schusterman, R. J., & Kastak, D. (2003). Auditory masking in three pinnipeds: Aerial critical ratios and direct critical bandwidth measurements. The Journal of the Acoustical Society of America, 114(3), 1660–1666. https://doi.org/10.1121/1.1587733..

Another way to measure frequency selectivity is based on the concept of the critical ratio at a frequency. The critical ratio is a dimensionless unit that is the decibel difference between the sound level of a pure tone signal of interest and the power spectral density level of a masking sound when the pure tone can just be detected. The critical ratio represents the signal-to-noise (SNR) ratio, in decibels, that just allows a signal of interest to be heard in the presence of a masking sound. If the level of the masking sound increases beyond the critical ratio, the signal of interest is masked.

Critical ratio is the SNR that allows a signal of interest (blue) to be detected amidst a masking sound (yellow). Figure by DOSITS.

Using the information that the SNR is equal to 0 dB at the detection threshold, Fletcher (1940) put forward the equal-power assumption that the signal power equaled the noise power within the critical bandwidth at the detection threshold. He assumed that the critical bandwidth was a rectangular bandpass filter, in which case the critical bandwidth (CBW, in Hz) could be estimated from the critical ratio (CR, in dB):

CBW = 10CR/10

Since critical ratios have been measured for many animals, this would be a convenient estimate of frequency selectivity; however, estimates of critical bandwidth from critical ratios do not agree with direct measurements of critical bandwidth[9]Yost, W. A., & Shofner, W. P. (2009). Critical bands and critical ratios in animal psychoacoustics: An example using chinchilla data. The Journal of the Acoustical Society of America, 125(1), 315–323. https://doi.org/10.1121/1.3037232..

Research has shown that some mammal species are inefficient at detecting a tonal signal among masking sounds because they use wideband processing[10]Yost, W. A., & Shofner, W. P. (2009). Critical bands and critical ratios in animal psychoacoustics: An example using chinchilla data. The Journal of the Acoustical Society of America, 125(1), 315–323. https://doi.org/10.1121/1.3037232.. That is, they either have wider critical bandwidths or they use several narrowband critical bandwidths together to detect signals. Yost and Shofner (2009) studied the chinchilla, an animal often used to study mammalian hearing in general, and hypothesized that wideband processing may have evolved to avoid missing important signals, such as those from a predator. That is, the animals may trade frequency selectivity for broadband sensitivity so that they can process efficiently many different signals at one time. Yost and Shofner (2009) recommended that critical ratios not be used to estimate critical bandwidths. Care must be taken when extrapolating measurements of signal detection since the signal of interest, the masking sound, and other background sounds will influence an animal’s detection ability.

In summary, the degree of acoustic masking depends on the animal’s ability to discriminate a sound. Detection of a signal depends primarily on the characteristics of the cochlea and how it encodes signals. Investigations into hearing sensitivity have focused on very fine detection of pure tones in laboratory settings, which is not a situation likely to be encountered in nature and certainly not conditions that have driven the evolutionary development of hearing. Therefore, while critical bandwidths and critical ratios may inform the degree of masking of very specific, tonal signals, other processing mechanisms have likely evolved to detect highly variable, biologically important signals. Furthermore, other parameters also influence the degree of masking and the reader is referred to How do Marine Mammals Hear? and the discussions of the potential effects of masking in marine mammals and masking in fishes for more information.

 

Additional Links on DOSITS

Additional Resources

References

  • Au, W. W. L., & Moore, P. W. B. (1990). Critical ratio and critical bandwidth for the Atlantic bottlenose dolphin. The Journal of the Acoustical Society of America, 88(3), 1635–1638. https://doi.org/10.1121/1.400323
  • Erbe, C., Reichmuth, C., Cunningham, K., Lucke, K., & Dooling, R. (2016). Communication masking in marine mammals: A review and research strategy. Marine Pollution Bulletin, 103(1–2), 15–38. https://doi.org/10.1016/j.marpolbul.2015.12.007.
  • Fletcher, H. (1940). Auditory Patterns. Reviews of Modern Physics, 12(1), 47–65. https://doi.org/10.1103/RevModPhys.12.47.
  • Popov, V. V., Supin, A. Ya., Wang, D., & Wang, K. (2006). Nonconstant quality of auditory filters in the porpoises, Phocoena phocoena and Neophocaena phocaenoides (Cetacea, Phocoenidae). The Journal of the Acoustical Society of America, 119(5), 3173–3180. https://doi.org/10.1121/1.2184290
  • Southall, B. L., Schusterman, R. J., & Kastak, D. (2003). Auditory masking in three pinnipeds: Aerial critical ratios and direct critical bandwidth measurements. The Journal of the Acoustical Society of America, 114(3), 1660–1666. https://doi.org/10.1121/1.1587733
  • Turnbull, S. D., & Terhune, J. M. (1990). White noise and pure tone masking of pure tone thresholds of a harbour seal listening in air and underwater. Canadian Journal of Zoology, 68(10), 2090–2097. https://doi.org/10.1139/z90-291
  • Yost, W. A., & Shofner, W. P. (2009). Critical bands and critical ratios in animal psychoacoustics: An example using chinchilla data. The Journal of the Acoustical Society of America, 125(1), 315–323. https://doi.org/10.1121/1.3037232

Cited References

Cited References
1 Fletcher, H., & Munson, W. A. (1937). Relation between loudness and masking. The Journal of the Acoustical Society of America, 9(1), 78–78. https://doi.org/10.1121/1.1902030.
2, 9, 10 Yost, W. A., & Shofner, W. P. (2009). Critical bands and critical ratios in animal psychoacoustics: An example using chinchilla data. The Journal of the Acoustical Society of America, 125(1), 315–323. https://doi.org/10.1121/1.3037232.
3 Au, W. W. L., & Moore, P. W. B. (1990). Critical ratio and critical bandwidth for the Atlantic bottlenose dolphin. The Journal of the Acoustical Society of America, 88(3), 1635–1638. https://doi.org/10.1121/1.400323.
4 Popov, V. V., Supin, A. Ya., Wang, D., & Wang, K. (2006). Nonconstant quality of auditory filters in the porpoises, Phocoena phocoena and Neophocaena phocaenoides (Cetacea, Phocoenidae). The Journal of the Acoustical Society of America, 119(5), 3173–3180. https://doi.org/10.1121/1.2184290.
5, 7, 8 Southall, B. L., Schusterman, R. J., & Kastak, D. (2003). Auditory masking in three pinnipeds: Aerial critical ratios and direct critical bandwidth measurements. The Journal of the Acoustical Society of America, 114(3), 1660–1666. https://doi.org/10.1121/1.1587733.
6 Turnbull, S. D., & Terhune, J. M. (1990). White noise and pure tone masking of pure tone thresholds of a harbour seal listening in air and underwater. Canadian Journal of Zoology, 68(10), 2090–2097. https://doi.org/10.1139/z90-291.