Page 16 - Spring 2018

Physics and the Mammalian Ear
turnal moths. These moths, prey to bats, have an ear with just two auditory sensory neurons tuned selectively to bat calls. This selective pressure has resulted in a spiraling competitive “war” between bats and moths and involve physics-based acoustic tricks (ter Hofstede and Ratcliffe, 2016; Pollack, 2017). Accord- ingly, “Evolutionarily speaking, insects have responded to se- lective pressure from bats with new evasive mechanisms, and these very responses in turn put pressure on bats to ‘improve’ their tactics” (Miller and Surlykke, 2001, p. 570).
How to Deal with the Evoked
Signal Complexity
Survival in a complex acoustic environment requires plas- ticity of the central nervous system, including learning and memory. This flexibility permits designs of the hearing pe- riphery that may be suboptimal from a physics point of view but sufficient from an evolutionary perspective. The hearing organs of vertebrates are dedicated to the sensory transduc- tion and transmission to the brain of all components that are detected in a complex acoustic signal. These include those produced as a by-product of the hearing organ’s own non- linear properties, e.g., the generation of intermodulation distortion products by the nonlinear transfer function at the level of single sensory cells (Jaramillo et al., 1993).
Why are there no filtering mechanisms in the hearing organs of vertebrates to remove processing artifacts? The biological answer to this question is that any auditory signal processed by the cochlea (external or intrinsic) will act as a selection pressure to a lesser or greater extent. Complexities in signals, such as the distortion products, are not filtered at the co- chlear level. To apply filtering, one has to know a priori the origin of a signal. Natural signals (e.g., birdsong, speech), however, are generally very broadband and their interpre- tation by the brain involves both learning and memory to make decisions about which frequencies are important and which are not. Under these circumstances, determining the difference between an internal signal generated due to dis- tortion compared with the same frequency originating in the outside world is clearly unrealistic. Moreover, distortion components tend to be more than 40 dB smaller than prima- ry signals and vertebrate ears cope with them because other advantages they offered were favored by natural selection. In addition, remarkably, mosquitoes exclusively use distortion products to hear (Simões et al., 2016).
Echoes, also, can be a bane or a boon. In reverberant envi- ronments, the brains of most animals suppress echoes so that auditory perception is dominated by the primary or leading sounds. The leading sound is enhanced and all that follows within a brief window of time is suppressed—but not in bats. Echolocation by bats relies on the timing of echoes to enable them to detect and locate prey and other sound-reflecting features of their environment (Neuweiler, 2003). The leading sound, or a signal generated by motor neurons that control the echolocation call, gates a brief window of time during which the echo responses of neurons in the auditory path- way are facilitated. Rather like the spatial representation of the visual system, where the retina is mapped onto neurons in visual brain areas, bats have echo-delay neurons mapped according to the time delays due to the distance between the bat and the target (Neuweiler, 2003; Simmons, 2017). This is true also of the incredibly sharply tuned cochlear mechani- cal resonance at the dominant frequency of the echoloca- tion calls, around 61 kHz, of P. parnellii. Because of cochlear ringing, such high-frequency resolution is accomplished at the expense of losing temporal resolution. However, neural inhibition enables a specialized subpopulation of duration- tuned neurons tuned to the bat’s 61-kHz cochlear resonance to escape the spectrotemporal trade-off (Macías et al., 2016). These remarkable neurons are the result of the inherent plas- ticity of the central nervous system and its programming. Clearly, the brains of vertebrates have evolved processing complexities that are able to provide an auditory perfor- mance that is not predictable from knowing the inherently limited capabilities of the cochlea itself.
Conclusions
Hearing organs, being sound detectors, must obey the gen- eral laws of physics. In this sense, physics imposes limits for evolution but does not direct it as such or necessarily leads to the selection of one “optimum design.” Vertebrate hearing organs arose from preexisting vestibular sensory epithelia that, over hundreds of millions of years, were slowly modi- fied in adaptation to higher frequency stimuli, first in water, then in air. The principles of evolution are not in the tool- box of physics and obey completely different laws that are derived from genetics and thus historical contingency. The resultant variety of hearing organs is concomitant with the variety of ways of life in the animal kingdom.
14 | Acoustics Today | Spring 2018