lec18reading

lec18reading - Chapter Sound Reception SOUND RECEPTION IS...

Info iconThis preview shows pages 1–19. Sign up to view the full content.

View Full Document Right Arrow Icon
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 2
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 4
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 6
Background image of page 7

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 8
Background image of page 9

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 10
Background image of page 11

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 12
Background image of page 13

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 14
Background image of page 15

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 16
Background image of page 17

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 18
Background image of page 19
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Chapter Sound Reception SOUND RECEPTION IS THE OPPOSITE PROCESS from sound production: sound vibrations propagating in the medium must first be coupled to the organism, modified as necessary, and then converted into useful nerve signals. Ideally, a perfect sound receiver would have a very wide range of detectable frequencies (frequency range), a very wide range of detectable amplitudes (dynamic range), very fine fre— quency resolution (Af), very fine temporal resolution (At), very ac— curate measurement of signal amplitudes, very accurate determina— tion of the angles of a sound source in the horizontal and vertical planes, very accurate determination of the distance to the source, and. the ability to monitor very rapid changes in the time domain waveform of signals (pattern recognition). 141 142 Chapter 6 Not surprisingly, it is impossible to achieve ideal conditions for all fea- tures in a single sound receiver. Evolutionary investment in one feature in~ variably curtails perfection in another. We have already discussed the physi- cal tradeoff that must exist between having a low Af and a low At. As we discuss later, there are other serious tradeoffs that exist between the various features. In addition, the environmental conditions under which organisms live may impose severe constraints on the perfection of specific features. Some of these constraints may be direct limits on the ways in which ears work; oth— ers may arise because the signals that are most easily produced or propagated in a given medium may not be the ones most easily detected and analyzed. Clearly, some tradeoffs between sound production, propagation, and recep- tion will exist. In this chapter, we review the more common mechanisms for coupling sounds to organisms and converting them into neural signals. We then review the major taxa that use acoustic communication and contrast their success in designing sound receivers. COUPLING OF PROPAGATING SOUNDS TO ORGANISMS Ideal couplers would transfer a large fraction of incident sound energy in the propagating medium to the detector organs of the receiver organism while preserving the amplitude, phase, frequency, and directional information con- tained in the original signal. There are basically three kinds of couplers found in animals. Each of these varies in its ability to meet these requirements. Particle Detectors Particle detectors are long thin organs that move when barraged by many molecules all moving in the same direction. As we have seen, such mass movements of particles due to sound propagation occur only in the near field. Sensory cells attached to the base of such organs are alternately stretched and compressed as the oscillating sound field forces the organ back and forth. Most arthropods have a variety of hairs on their bodies that provide chemosensory and tactile information. In addition, many also have very fine hairs (called trichobothria in spiders and trichoid sensilla in insects); these hairs are particle detectors sensitive to air or water currents and to near field sounds (Figure 6.1). Some caterpillars use these hairs to detect the wing-in~ duced air currents of approaching predatory wasps (Tautz and Markl 1978). Crickets and cockroaches have long cerci at the tips of their abdomens that are covered with many fine hairs. Variation in the lengths of these hairs is corre- lated with the particular near field sound frequencies to which each hair is sensitive (Petrovskaya 1969). In other insects, thin antennae or mouth parts are used as particle detec~ tors. The courtship dance of male fruit flies produces vibrations at about 200 Hz. Females within 0.5 cm of a dancing male can detect and monitor the near field signals with a fine segment of each antenna called the arista. The plumose antennae of male mosquitoes and midges are used to detect the species—specific wing beat frequencies of females. Water boatmen have a club Sound Reception 143 (A) I (B) Figure 6.1 Particle detectors (C) (D) in arthropods. (A) Antennal ‘ arista of a fruit fly. (B) Tri— choid sensilla on a cockroach cercus. (C) Antenna of a male H mosquito. (D) Trichoboth- rium of a spider, T, and a larger sensory hair, H. (A after Burnet et a1. 1971 ; B after Autumn 1942; C after Hutchings and Lewis 1983; D after Foelix 1982.) that extends from their ear and responds to near field Vibrations generated in the water by other individuals. Note that aquatic arthropods are more often within near fields than are terrestrial forms because of the longer wavelengths for any given frequency in water. The frequency response of a particle detector depends upon its ability to oscillate in phase with the particles of the surrounding medium. Massive hairs will be unable to keep up with high frequencies or large amplitudes, and even the smallest particle detectors appear to be limited to frequencies well below 1 kHz. Insects such as male mosquitoes and female fruit flies use this constraint by adjusting the mass, hing-mg, and shape of their antennae to limit their responsiveness to the species—specific frequency band. Where they have been studied, the nerves arising from antennal particle receptors appear able to track the waveforms of conspecific sounds well as long as very high frequencies are not present. Thus they can provide considerable information to higher brain centers with only minimal distortion. Long hairs and antennae that are attached at the cuticle of the animal pro— vide a significant mechanical advantage to the sensory cells: a very small force distributed over the long axis of the hair or antenna is transmitted to the sensory cells as a very small displacement but one of great force. Thus very weak sound fields can be detected by such receivers. The enormous mechani— cal advantage of the mosquito antenna makes it one of the most sensitive au« ditory receptors known among the arthropods and similar in sensitivity to the far field response of the human ear (0 dB SPL). in water, the much smaller displacement of molecules for a given sound energy makes it challenging to produce a particle receptor response. In addi— tion, because most animal tissues have similar acoustic impedances to that of 144 Chapter 6 water, near field displacements cause most of the molecules in an organism to move in concert with the sound. When the tips and bases of particle detectors move together, there can be no flexing of the detector and thus no stimulation of the sensory cells. To get around this limitation, some organisms Whose de~ tectors receive Vibrations through water attach the tips of sensory hairs to ob- jects that have quite different masses and acoustic impedances from the sur— rounding tissues. The objects then vibrate either with a lag or with a decreased amplitude when compared to the base tissues of the hairs. The greater the dif- ferential in movement between the heavier objects and the surrounding tis~ sues, the greater the stimulation of the sensory cells. Since the objects will have their own acoustic properties, such objects may limit the response of such a system to specific frequencies. Where heavy objects such as calcium carbonate masses are used, the response will usually be limited to lower frequencies. Particle detectors are inherently directional (Figure 6.2). When the axis of particle movement is parallel to a sensory hair, the hair will not move; when the axis of particle movement is perpendicular to a hair, it will experience the greatest force upon it. Angles of incidence between 0° and 90° will create inter- mediate forces on the hair depending upon the value of the perpendicular component. Thus the amplitude of stimulation from a single hair could be used as an index of the direction of the signal source. However, there are sev- eral problems with using particle detectors to determine source direction. One is that the animal cannot know from a single hair whether variations in hair displacement are due to differing angles of incidence or to different ampli- tudes of the signal at the source: signal amplitude and signal direction are con~ founded. A second is that near fields that are dipolar or of higher—order geome- lncident wave Hair —— Vertical force on hair Figure 6.2 Components of force on a hair shaped particle detector. Incident sound will have two components of force, one perpendicular to the hair and one parallel to the hair. Only the per— V 3 pendicular component will move the Guide V hair. As the angle of incidence changes from parallel to perpendicular to the hair, the movement of the hair for a sound of a given amplitude Will increase. \ Horizontal force on hair Sensory cell nucleus Sensory nerve Sound Reception 145 try, certainly the most common source of sounds, do not necessarily produce particle movements whose axis points at the source. Even if an organism can uncouple source amplitude and directional information, it must somehow map the sound field’s geometry before extrapolating the location of the source. A common method of decoupling source angle and signal amplitude is to have multiple particle detectors. Many hairs can be spread over the body, and two or more antennae can be placed on the head. Particle detectors can be hinged or curved so that they move only along a single axis with different de— tectors aligned along different axes. Since the alignment and body curvature is given, a reasonable brain ought to be able to combine the information from such an array of detectors to provide separate estimates of source signal am- plitude and source location. Particle detectors thus meet some of our ideal conditions and not others. They are primarily limited to lower-frequency signals and have dynamic ranges and temporal tracking abilities limited by their mass and inertia. Within these limited frequency and amplitude ranges, they track slowly vary- ing waveforms well and can thus provide relatively undistorted signals to sensory cells. Particle detectors are inherently directional and thus can pro— vide information about a signal source given independent measures of signal amplitude. Their use is limited‘to near fields. Pressure Detectors Nearly all terrestrial organisms that do not utilize particle receptors for sound coupling rely on some sort of membrane (often called a tympanum). When the pressures on the two sides of a tympanum are unequal, the tympanum is bent away from the side of higher pressure. This bending of the tympanum can then be coupled to sensory cells to produce neural impulses. A pressure detector consists of a tympanum stretched over a closed cavity (Figure 6.3). Sensory cell Drum membrane Figure 6.3 Pressure detector. A thin membrane is stretched over a closed cavity. When the incident sound pressure is greater than that inside the cavity, the mem— brane (called a tympanum) is forced into the cavity When the external pressure is lower than that inside the cavity, the tympanum is forced out. Movement of the tympanum is monitored by sensory cells. 146 Chapter 6 There must be no effective opening between the cavity and the outside world and the walls of the cavity must be reflective at the sound frequencies of inter‘ est. When the typanum is at rest, the pressures of the medium inside and out- side the cavity are similar. When a propagating sound arrives at the tympa- num, the rise in pressure outside the tympanum bends it into the cavity until the compression of the enclosed medium results in an equal but opposite pressure. Similarly, when the pressure outside of the tympanum is lower than that inside, the tympanum is bent outward until the increased cavity volume results in a compensatory lowering of cavity pressure equal to that outside. Sensory cells attached to the tympanum can then record the movements of the tympanum into and out of the cavity. Pressure detectors are used in far fields. Because terrestrial organisms consist of water and solids, they must cope with the fact that much of the sound energy carried in air will be reflected at their ears. Tympana have sev- eral advantages as terrestrial sound couplers. They can be made very thin, so that their acoustic impedance is closer to that of air. Because the force, F, on a pressure detector tympanum equals the product of the sound pressure, P, and the membrane area, A, large thin membranes are very efficient ways to trap incident sound energy. The thinness of the membrane and the strong restora— tive force created by compression of the closed cavity volume also help pres- sure detectors to track changes in incident waveforms rapidly. Like a string, a stretched tympanum has a number of resonant modes at which it responds strongly to incident sounds. The frequency of the lowest mode increases with the thickness of and tension on the tympanum, and de- creases with the diameter of the tympanum. Unlike a string, the frequencies of successive modes are not harmonics of the lowest mode, but tend to be spaced at intervals much less than the lowest mode’s frequency. Thus a very thin but large—diameter tympanum may respond to a much greater variety of frequen- cies within a given range than an equivalently sized string. This range makes a tympanum well—suited to respond to a wide variety of incident sound frequen- cies. In some organisms, however, this breadth of frequency response is undev- sirable. Such species have modified membrane thickness and use complex membrane shapes to constrain the resonant modes to preferred values. Sound pressure has no inherent direction. This means that a single pres~ sure detector lacks any intrinsic mechanism for determining the location of the sound source. One solution is to add a directional structure between the pressure detector and the ambient medium. Nearly all terrestrial mammals have a directional pinna that connects to the tympanum via a funnel—shaped tube (called an auditory meatus). Such a pinna can provide directionality only for wavelengths that are similar to or smaller than the size of its external opening. Thus the utility of a pinna is limited to higher frequencies. The aim“ plest use of a directional pinna is to aim it in various directions and search for a maximum in the received signal amplitude. Many mammals have elabo- rated the structure of pinnae with numerous fine folds, ridges, and protru~ sions (see Figure 6.13). Each of these ridges or folds reflects sounds of a partic- ular wavelength, or shorter. The relative phases of the separate reflections at Sound Reception 147 the tympanum, for any given wavelength, will vary depending upon the ver— tical and horizontal angle of the source. Thus the amplitude of any given fre- quency component at the tympanum will vary as a function of source loca- tion. By comparing the frequency spectrum received at one tympanum with that expected for a given type of source, the animal can often reconstruct the vertical and horizontal angle of the sender. The problem with this method is that frequency spectra of sounds are easily altered during propagation and thus may never be similar to that expected. This makes it difficult to separate the effects of location and distortion when only a single ear is used. If, how— ever, the animal uses two ears, each with a pinna, similarities in the spectra at the two ears can be used to determine the likely spectrum of the incident sound, and differences between the ears can be used to extract information about the sound source’s location. Note that this type of localization mecha- nism requires an inner ear that can perform at least a crude Fourier analysis of the received sounds. The use of a pair of pressure detectors also provides a number of other di- rectional cues. If the incident sound has a sharp onset or ending, or has some other portion clearly defined by an amplitude or frequency modulation, the difference in time of arrival of that marker at the two ears can be used to esti- mate the location of the source in at least one plane. In the simplest case, an animal with two ears can rotate its head. until the time delay is minimal. The sound source will then either be directly in front of it or directly behind it. Ro- tating a quarter—turn more will then resolve this final ambiguity. The problem with this method is thatthe sound must persist long enough to perform the head rotations. An animal could react faster by computing the angle of a sound source using the delay in arrival of a single sound at the two ears (Fig- ure 6.4). To be useful, the time delay must be long enough to be resolvable by normal nervous systems. In humans, the maximal delay of arrival at the two ears (given a typical head diameter) is about 0.5 msec. For a small terrestrial animal with only 1 cm separating its two ears, the delay is at most 0.03 msec and for a similarly sized animal in water the delay is 22% of the terrestrial value. Although some animals can resolve very fine time delays (owls are ac- curate down to 0.006 msec), this detection usually takes a very large number of brain cells. Because only larger animals can afford to dedicate so many brain cells to time-delay resolution, and because larger animals will be able to separate their ears more and create longer maximum time delays, this mecha— nism of sound source localization is most common in larger species. A second method for determining source location using two pressure dec— tectors is to compare the relative phases at the two ears. Unless an animal is facing directly toward or directly away from a sound source, the phase of an incident sound differs at the animal’s two ears. Again, the animal can rotate the head until such differences are nulled out. Maximal phase differences will occur when the wavelength of the incident sound is twice as long as the dis— tance between the two ears. At much lower frequencies, the waves are so long relative to the distance between the ears that there will be little difference in perceived pressures at the two ears. At much higher frequencies, there will be 148 Chapter 6 Figure 6.4 Determination of sound source using two ears and time delay. The time domain waveforms of signals from two ears are shown as perceived by a center in the brain that receives signals from both. The white waveform shows a signal from the left ear and the black waveform shows a signal from the right ear. Arrows show the angle of source in the horizontal plane of the head. When the sound source is along a line perpendicular to the line joining the two ears and midway between them, the sound arrives at the two ears simultaneously. When the sound is opposite one ear and on a line connecting the two ears, maximum delay is achieved. If the dis~ tance between the ears is known, the horizontal angle to the source can be computed from the measured delay. Ambiguity concerning whether the source is in front of or behind the animal can be resolved by rotating the head. many different positions at which phase differences will null. out. Thus a ter— restrial animal with ears separated by 1 cm is likely to use phase information only if it typically responds to sounds of 12—22 kHz. An animal with ears sep- arated by 10 cm will be able to use phase information for sound frequencies down to a few kHz. Finally, when a far field sound encounters an object at least as large as one-tenth of a typical wavelength, diffraction causes a distortion of the sound field around the object. Sound pressure on the side of the object nearest the source will. build up above free field levels, and interference between directly reflected and creeping sound waves will generate a complicated pattern of pressure levels at other angles around the object. As the ratio between the ob~ ject and wavelength is increased to 5, the pressure build up on the side facing the source will approach an asymptote at about 6 dB relative to the free field values. Pressures on the far side of the object vary with location, shape of the object, and the object to wavelength ratio but are at most 8—10 dB lower than those on the near side. Further increase in the ratio above 5—10 leads to a sound shadow on the far side with differences between near and far sound pressures up to 25 dB. The diffractive distortion of a sound field around an animal with two ears can thus be used to determine the location of a sound source. As before, a Sound Reception 149 simple rotation of the head until sound pressures are similar on both sides is one way to locate a line to or from the source. Because diffractive fields around heads and bodies are highly dependent upon the wavelengths of the incident sounds as well as the sizes and shapes of the organisms, computation of a sound source's location from diffractive intensity differences is much more complicated than comparing time delays. In addition, the constraints on maximum intensity differences, particularly at intermediate object-to- wavelength ratios, limit the accuracy. One way mammals get around these limits is through the use of pinnae. As noted earlier, these organs limit the angle of acceptance of impinging sounds and, by altering the frequency spec- tra as a function of incident angle, provide information about source loca- tions. Because a pinna and its attached auditory meatus are both tapered or— gans, they focus the sound energy captured over the large opening of the pinna onto the small opening at the end of the meatus. This process can result in a frequency—dependent amplification of the incident sound wave by as much as 30 dB. If the tuning of the pinnae and meatus are set to maximally amplify frequencies at intermediate object—to—wavelength ratios, much greater ranges of intensity difference between the two ears and thus more accurate es- timates of source location can be obtained. Pressure—Differential Detectors When a propagating sound can reach both sides of a tympanum, the sound receiver is called a pressure-differential detector (also called a pressure~dif— ference or pressure-gradient detector). Such an ear samples the sound field at two different locations, and the two samples are conducted to opposite sides of the tympanum. As long as the two samples are out of phase when they reach the tympanum, (a function of how far they have traveled and their rela— tive phases when sampled), the tympanum will be bent toward the sample with lowest pressure (Figure 6.5). Like particle detectors, pressure—differential detectors are intrinsically di— rectional. Quantitatively, the magnitude of the force, F, exerted by an incident sound on a pressure-differential membrane is F = (2‘)": AP AL cos 0)/l, where A is the surface area of the membrane, P is the incident sound pressure, it is the wavelength of the incident sound, AL is the extra distance the incident Membrane Tube Figure 6.5 Pressure-differential detector. A closed tube samples the sound field at two different locations. The tympanic membrane can be locat— ed at any point within this tube or at either end. The pressure difference at two sampling points displaces the Pressure difference membrane. : 150 Chapter 6 waves must travel to get to one side of the membrane when compared to other side (this equation assumes that AL S 2.), and 9 is the incident angle of the sound relative to its perpendicular incidence on the membrane (Michelsen and Nocke 1974). The cosine term is maximal when sound hits the tympanum perpendicularly (one sample arriving sooner at the near side of the tympa~ num), and is zero when the sound propagation direction is parallel to the tympanum (since both samples arrive having traveled the same distance). This term generates the directionality of the detector. Because the forces on their tympana are inversely related to wavelength, animals with pressure-dih- ferential detectors have difficulty detecting lower frequencies. One way to get around this problem is to increase AL sufficiently to compensate for the other— wise low force generated by low-frequency sounds. Small animals such as in“ sects, frogs, and birds use this trick by means of a tube that connects to one side of their pressure-differential membranes and that routs the other end to a location as far from the membrane as is anatomically feasible. In principle, the right combination of AL, 2L, and 6 could produce a net force on the tympanum twice what would be experienced by a pressure de~ tector at the same location. However, most animals that use pressure-differen- tial detectors do so because they are too small relative to the wavelengths of relevant sounds to use either time delays or intensity for sound—source local~ ization. This means that the ratio AL/it is very small, and thus the force expev rienced by a pressure—differential detector in a far field is often much smaller than that for an equivalent pressure detector. This cost in far field sensitivity is presumably compensated by the opportunity to have accurate directional» ity even when useful wavelengths are much larger than the animals detecting them. It is also the case that a pressure—differential detector, which responds most strongly when the spatial gradient of pressure is highest, will be much more sensitive in a near field than will an equivalent pressure detector. Thus losses in sensitivity of an ear that is a pressure—differential detector relative to a pressure detector at a considerable distance from a sound source become relative gains when the sender and receiver are close. In principle, the two new ceivers should show similar sensitivities at higher frequencies. However, [ricer tion imposed by the walls of the narrow tubes connecting the internal side of the pressure—differential tympanum to a secondary opening will attenuate the signal amplitude, and this attenuation will increase with frequency. Thus in many species with openings on the internal side of the tympanum, the ear works as a pressure~differential detector only at lower frequencies. At higher frequencies, it then acts as if the openings were absent and thus as a pressure detector. This means that the ear loses its intrinsic directionality at these higher frequencies. However, these are usually the frequencies at which. other cues such as intensity differences are most available. Pressure-differential ears can incorporate most of the refinements and ex— ternal modifications used for pressure detectors. The directionality and ampli» fication of a single ear can be improved by imposing an external cone and. an auditory meatus between the tympanum and the ambient sound fields as, for example, occurs with the feathered ear funnels of birds. As long as the ear can ——.——_—_——————W Sound Reception 151 perform some sort of Fourier analysis, directional information can be ex- tracted by examining frequency spectra. In addition, most animals with pres- sure—differential detectors have paired ears that provide the same directional cues as for pressure detectors. These animals also have an extension tube con- necting one tympanum to that of the opposite ear. This arrangement ensures that the interior sides of the two tympana experience the same signal. Con- trasts between the two ears are thus more clearly attributable to differences in location relative to the sound source. In summary, pressure-differential ears are less effective at low frequencies than are pressure detectors. They are also less sensitive than pressure detec— tors in far fields, except at higher frequencies, but respond better than pres— sure detectors in near fields. Pressure detectors have no single ear directional- ity. Animals that are large relative to the wavelengths of their sounds can get around this problem by using two pressure detectors to determine the loca- tion of a sound source. This arrangement will not work for small animals. Be- cause pressure—differential detectors can provide quite accurate indications of the location of a source, even with only a single ear, they are widely used by small animals to guarantee directionality. Directionality and amplification in both kinds of ears can be increased by adding a pinna or funnel between the incident sounds and the tympana. Both systems provide the advantages of large membranes in capturing incident sound energy and providing good tracking of incident waveforms. MODIFICATION OF COUPLED SOUND VIBRATIONS As we noted on pages 22—24, the energy carried by a sound wave in a given medium equals the product of the pressure wave and the distance over which the wave moves the medium particles. Sound traveling in a low—impedance medium such as air is represented by low pressure and large particle move— ments of medium molecules. A sound of the same total energy traveling in a medium of high impedance, such as water, is represented by high pressure and short movements of medium molecules. When sound traveling in one medium strikes the boundary of a medium with a very different impedance, most of the energy will be reflected. Since terrestrial animals are largely made of water and solids, they have much higher acoustic impedances than the sur- rounding air. This makes it difficult for them to trap enough energy in inci~ dent sound waves to stimulate their ears. The compensatory adaptation that we find in the ears of organisms is a stepwise matching of the impedances of the medium to that of the organism to permit better sound-wave capture. Put another way, a good ear trades off pressure for molecular displacement with as little loss as possible as one moves from one medium to another. One solution for a terrestrial organism is to mount a horn exterior to the tympanum. Horns are tubes with large diameters at the air interface and small diameters inside the animal’s body. Just as a horn can be used to in— crease sound output by a signaler (see Chapter 4, pages 90—91), a horn on the ear can be use to increase sound capture by a receiver. Birds and mammals 152 Chapter 6 typically have funnel—shaped tubes connecting their tympana to the outside world. These tubes allow for a gradual change in acoustic impedance as the sound enters the animal. The pinnae of mammals and dish-shaped plumage of some birds further enhance this impedance matching by extending the fun- nel beyond the skull. Once the sound reaches the tympanum, reflection of sound waves can be further reduced by making the tympanum thin and light, and by placing an air-filled cavity behind it. These strategies are, of course, common for pressure detectors and pressure—differential detectors. The combination of a thin, light membrane and an air cavity makes the average acoustic impedance of the tympanum intermediate between that of air and that of the fluid—filled cells of the inner ear. This arrangement thus reduces the losses due to reflection at the tympanum and provides the first body part that is made to move by the im- pinging sound. The final step of connecting the vibrating tympanum to the inner ear usually involves an additional impedance—matching step. In terres- trial vertebrates, the tympanum is connected by a narrow rod or chain of ar- ticulated rods to a thin membrane (oval window) in the wall of the inner ear. When the size of the oval window is much smaller than that of the tympau num, all of the energy collected by the large tympanum can be concentrated on this smaller window, providing greater force per unit area. This concenw trated force better matches the higher pressure required to transmit sounds in the fluid-filled inner ear. ANALYSIS OF COUPLED SOUND VIBRATIONS Once the sound has been coupled to sensory cells of the organism, various neural configurations can be used to characterize and decompose the sound into its components. Two leVels of processing generally occur. Peripheral pro- cessing occurs at the level of the sound reception organs, and is followed by central processing at higher centers such as the brain. Peripheral Frequency Analysis There are three common methods by which animals decompose complex waveforms into frequency spectra. The simplest is to fine~tune different groups of receptors cells to different frequency bands. If the cells tuned to a specific band vary their nerve impulse rate as a function of the intensity of stimulation, higher-order centers in the brain can compare the rates of neural firing from different groups and reconstruct the frequency spectrum. Because it is difficult and expensive to build widely different kinds of receptors in t0 the same ear, this type of mechanism is limited to narrow frequency ranges. A second method utilizes sensory cells that are sensitive to a very wide range (“if frequencies. If a given cell fires whenever the signal waveform exceeds a cer- tain threshold, it will be able to track the signal periodicity with its impulse . rate. This method is sometimes called the telephone principle or phase-lo ck- ing. The frequencies of simple sinusoids would thus be mapped onto nerve firing rates and identified by feeding these impulses to higher—order cells, W Sound Rebeptian 153 which only trigger when stimulated at certain impulse rates. Varying the sen- sitivity thresholds of these higher—order cells or using cells which only re— spond to changes in impulse frequency would facilitate decomposition of complex waveforms. However, even with very sophisticated processing, this method is limited at the front end by the inability of the sensory cells to fire impulses at high enough rates. In practice, therefore, this method is limited to frequencies less than 1—1.4 kHz. Sensory nerves of mosquitoes, which need to track sounds of only 400 HZ, Show phase—locking between sound wave peaks and nerve impulses. The final method of frequency analysis relies on the coupling of incoming sounds to a substrate of spatially variable acoustic properties. If the coupling is properly designed, the location of maximal displacement along such a sub- strate varies with the frequency of the incoming sound. Sensory cells are then attached to this substrate on one side and to a less frequency-sensitive base on the other. When a sound with a given frequency is coupled to the substrate, only sensory cells in a given location are stimulated. Different frequencies dis— place different parts of the substrate and thus different groups of cells. This method is called place principle analysis and the set of spatial correlations between nerve cells and frequencies is called a tonotopic map. Locusts and mole crickets trap sounds with a tympanum that has a very carefully designed shape and variable pattern of thickness (Michelsen and Nocke 1974; Michelsen 1979, 1983; Michelsen and Larson 1985). When a com— plex sound hits such a membrane, standing waves are set up by reflections from the membrane boundaries. Whether a particular place on the mem— brane is set into motion or not depends upon which, if any, of the resonant frequencies of the membrane are present in the sound. When resonant fre- quencies are present, the amount of movement at an antinode will depend on the amplitude of that frequency component in the signal. By attaching clusters of receptor cells at different locations on the tympanum, the presence and intensity of several discrete frequency bands within the complex signal can be determined. More complex methods using the place principle occur in katydids, crick» ets, cicadas, and terrestrial vertebrates. In all of these forms, the sensory cells are not placed against the tympanum, but are located in a separate auditory organ. The motions of the tympanum are then coupled to the auditory organ by some thin and responsive structure (a ligament in cicadas, membranes in crickets and katydids, and middle ear bones in terrestrial vertebrates). This placement has the advantage that the tympanal design can be dedicated to faithful sound capture, whereas the auditory organ can be given properties that facilitate efficient frequency analysis. When the two functions must be ac— complished in the same organ (as in locusts), neither can be made very effi- cient. Taxa that have a separate auditory organ differ in the way that sensory cells are arranged. Some organs have spatially distinct clusters of sensory cells (usually correlated with sensitivies to discrete bands of frequencies) and some have orderly rows of sensory cells (usually correlated with continuous tono— topic mappings of frequency onto location). 154 Chapter 6 In both insects and vertebrates that use the place principle for frequency separation, signal intensity is coded by the firing rate of the sensory neurons. As a rule, each sensory cell has a fixed range of intensities over which it can vary its firing rate. Different cells have different minimum or threshold inten- sities that must be attained before they will fire. Increasing the sound intensi- ties above this threshold increases the firing rate up to a limit, after which the cell impulse rate approaches an asymptote. In most organisms, the rate of sen- sory nerve firing tends to rise with the logarithm of the sound intensity or some similar power function. This means that the decibel (dB) scale is a more accurate representation of what the ear perceives than is a linear scale. It also means that each cell can hear a Wider range of intensities than if impulse rates were linearly related to intensity. Typical insect ear cells have dynamic ranges of 20—30 dB; vertebrate sensory cells have dynamic ranges of 40—50 dB. Where there are enough sensory cells to have several committed to each frequency band, cells Within that band are divided into several sensitivity sets, each dif- fering from the next by 20—30 dB. This arrangement permits dynamic ranges of 100 dB or more. Central Processing of Sound Signals The neuronal stimulation produced in ears is usually conveyed to the re- ceiver’s central nervous system for processing. Much of the lower part of the central nervous system is used to enhance contrasts and thus improve resolu- tion. This enhancement is often accomplished by using one kind of sensory cell to excite a higher-order neuron and other sensory cells with slightly dif— ferent sensitivities to inhibit the same neuron. Maximal stimulation of the higher—order neuron will only occur when the stimulus is closest to the op ti- mal stimulus for the excitatory cell and will drop off quickly as the stimulus is varied away from that optimum. When the criterion for similarity is fre— quency, such a process greatly improves the bandwidth (Af) at the level of the central nervous system when compared to that at the level of the ear. Similar arrangements in animal central nervous systems increase the resolution of signal amplitudes and the measurement of time delays between successive events such as pulses or modulations (At). This improved resolution of both frequencies and amplitudes in the central nervous system facilitates the use of frequency spectra for recognition of sounds and spatial localization of sound sources. It is interesting that the tonotopic organization of ears is often pre- served at successively higher levels in the brain. Although the intensity and temporal properties of the highest level brain cells may become very complex, cells with similar optimal frequency responses are still grouped together spa— tially. One reason for this continued segregation of cells with different optimal frequencies is the crucial role of frequency spectra in providing important in- formation to the receiver. In various arthropods, frogs, and reptiles, (but not birds or mammals), pe- ripheral tuning of the ear is used to reduce stimulation by sound frequencies other than the bands present in species-specific signals. Within these bands, frequency-spectrum information may be used to augment species recognition W 7” Sound Reception 155 or to discriminate between different conspecific signals. In bullfrogs (Rana catesibiana), a croak must contain energy at 200 Hz and 1400 Hz to evoke a re— sponse, but must not contain significant energy at 500—600 Hz (Capranica 1965). Adult males produce croaks with the appropriate spectra; younger males tend to have croaks with significant energy in the 500—600 Hz range and are ignored by females and adult males. Although peripheral tuning is widespread in some groups, it is often insufficient by itself to establish species identity (Capranica 1992). For example, even with strong peripheral filtering, a frog or arthropod is still likely to hear sounds of a number of other species. Additional species specificity relies on recognition of temporal patterning in the signals. Central nervous systems of frogs and arthropods invariably have phasic neurons (called chirp or pulse coders) that respond to the onset of am- plitude modulations and then quickly stop firing. Higher—order cells that re- ceive inputs from such phasic neurons can use the delay between these short bursts of stimulation to monitor the pulse or modulation rates in the received sounds and thus identify gross species—specific patterning. Additional levels of cells combining and comparing responses of lower levels are then used to improve discrimination between more complex temporal patterns. Another major acoustic task undertaken by the central nervous system is the determination of the location of a sound. As we have seen, angular loca- tion of a sound source requires accurate resolution of time delays, signal am- plitudes, and / or determination of frequency spectra. Even with particle or pressure—differential detectors, a pair of ears is required to untangle the con— founding effects of signal amplitude at the source, source location, and distor- tion during propagation. This means that the central nervous system must at— tempt to improve the resolution of temporal, amplitude, and frequency information from each ear, and then combine the outputs of the two ears to determine source locations. Because the frequency and temporal properties of the sensory cells in the ear are often linked, the interpretation of time delays at the two ears must include knowledge of which frequencies were involved. This is another reason why tonotopic structure of the ear is usually preserved at higher levels in the central nervous system. It also makes the combining of information from the two ears to determine source location more complicated than if delays from all frequency stimuli could be pooled. Finally, the binaural auditory signals are often combined with information from Visual, olfactory, or other modalities to maximize identification of source locations. The usual mechanism is to create a neural map for each modality and then to compare these maps. In a neural map,‘cells are stimulated only if a sound source is at a particular angular location, and adjacent cells represent adjacent angular loca— tions. Such maps appear to be the rule in birds and mammals. TAXONOMIC CONTRASTS OF EAR DESIGN In the remainder of this chapter, we compare the tradeoffs that have been made as ear design has evolved for different animal groups. Arthropods are generally small and this fact has had major consequences for the appropriate 156 Chapter 6 ear design. Fish live in water and this leads to an entirely different set of prob- lems when compared to terrestrial vertebrates. In addition to the ecological constraints imposed upon ear evolution, the phylogenetic history of a group sets the starting point on which subsequent selection can work. The effects of earlier historical conditions are particularly clear in the evolution of verte- brate ears. Insect Ears Most species of insects (except flies and beetles) have small fan-shaped clus- ters of sensory cells in the tibial joint of their legs called subgenual organs. These have no clear coupling devices other than attachment to the interior membranes of the tibia. Where they have been studied, subgenual organs ap- pear to provide the owner with a reasonable sensitivity to substrate~borne vi— brations. For some species, this ability is probably used to warn of the ap- proach of predators. For others, such as insects stridulating on plants (see pages 115—116), fiddler crabs (genus Uca) tapping the mud, or water striders (Hemiptera: Gerridae) perturbing the water’s surface (Wilcox 1988), sub- strate-borne vibrations are a major modality for conspecific communication. The fact that cells in the central nervous systems of several orthopterans re»- ceive input from both subgenual and ear sensory neurons suggests that these animals may combine or compare sound vibrations received Simultaneously by the two routes (Kalmring et a1. 1985). Subgenual organs respond only to frequencies below a few kHz, although the organs in some katydids may be sensitive to airborne sounds up to 4~w5 kHz. This probably reflects the better propagation of lower frequencies in solid substrates. It may also reflect limitations on the kinds of mechanisms available for generating signal vibrations. Water striders communicate by means of sur— face waves that they generate by pumping their forelegs (Figure 6.6). This type Figure 6.6 Water strider sending signals on water’s surface. Male Limnoporus notabilis producing territorial signal to repel conspecifics. Rings of surface perturbation can be seen radiating away from moving Strider. (Photo- graph courtesy of Stim Wilcox.) Sound Reception 15 7 of motion limits the frequencies that they can easily produce to 25—100 HZ. The organs used to receive these signals are apparently able to discriminate be- tween different frequencies within i 1.5 HZ (Wilcox 1988). Tuning of subgen- ual organs is thought to be performed either by adjustment of the moveable ends of the sensory cells or by higher order responses to phase—locked sensory neurons. It is likely that most of the more sophisticated ears in insects have evolved from such subgenal organs. A number of other insects have ears that are designed solely to detect the echolocation calls of bats. These calls are emitted in flight; the resulting echoes indicate to the bat the location, direction of motion, size, and shape of possible prey targets (see Chapter 26). To obtain good echoes from small targets, most bats use frequencies of from 15—125 kHz for echolocation calls. Some taxa of moths (Lepidoptera) have bat-sensitive tympana located on the thorax or first abdominal segments. In noctuids, each ear contains a single pair of sensory cells that are directly attached to the tympanum. In geometrids, there are four receptor cells for each ear. The ears of both groups respond to frequencies in the 15—125 kHz range. Although the internal sides of the tympana are con— nected to large tracheal air sacs that contact the corresponding sacs of the op— posite ear, the very high frequencies used by the bats prohibit any pressure- differential effect and cause each ear to act as a pressure detector. Despite the small number of cells per ear, the moth’s ears have very large dynamic ranges: in noctuids, signal intensities over a 40—50 dB range can be discrimi— nated. In part, this range is achieved by one cell responding at lower intensi- ties, and the second beginning to respond only when the first has reached sat- uration responses. Higher—order cells use this intensity information to determine whether the moth has time to fly out of the bat’s path or instead should dive for the ground. Binaural comparisons in the moth’s central ner- vous system determine which direction the moth should fly to evade the bat. Bat—detector ears are also known in lacewings (Neuroptera), hawk-moths (Lepidoptera), preying mantises (Mantodea), and some beetles (Coleoptera). For insects that use airborne sounds for sexual advertisement, the audi- tory tasks are considerably different from those for evading predatory bats. Receivers are more likely to be able to detect the advertisements of a potential mate if the latter uses loud signals with low to moderate frequencies. The only way an insect, which is small relative to such wavelengths, can detect moderate frequencies and still extract directional information about the sender is to use a pressure-differential ear. To ensure species specificity, the sender will surely incorporate some temporal patterning in its call. Wide bandwidth signals preserve temporal patterns during propagation better than do narrow bandwidth ones (Rohmer 1992), and directional and distance in- formation about a sender is most easily extracted from a wide bandwidth sig- nal. Sexual advertisement thus tends to favor sounds with wide bandwidths and ears that compare frequency spectra of these wide bandwidth signals. Such ears require more sensory cells and sophisticated couplings than is the case for simple bat detectors. Note that this does not preclude a sexually re— ceptive ear from also acting as a bat detector: ultrasonic sensitivity and bat 158 Chapter 6 avoidance responses have recently been found in crickets, locusts, and katy~ dids, all of which also respond to conspecific sexual calls (Hoy 1992). Most cicadas (Homoptera: Cicadidae) and grasshoppers (Orthoptera: Acrididae) produce sexual advertisement sounds in the range of 1—9 kHz, Both exhibit paired tympana on the anterior part of the abdomen (Figure 6,7). In male cicadas, there is a single large air sac that fills a large part of the ab- domen between the two ears. In females, the sac is smaller, but apparently still provides an acoustic connection between the two ears. Acridid grasshop— pers have several tracheal sacs and fat deposits connecting the two ears. For sounds below 10 kHz, these couple the two ears acoustically. Thus in both groups, acoustical connections between their ears allow each ear to act as a pressure—differential detector with intrinsic directionality and without resort‘ ing to very high frequencies. Although some grasshoppers are sensitive to fre« quencies above 10 kHz, the acoustic coupling between the ears becomes less effective at higher frequencies and thus the ears then become pressure detec» (A) p (B) Air sacs -—-— Tyinpanum I , Auditory * \ capsule Muller's organ Tympanum (C) (D) Spiracle Tracheal tube —-— —— Anterior tympanum Posterior tympanum Figure 6.7 Ears and associated air spaces in insects. (A) Cross section of the body of a locust (Orthoptera: Acrididae). Central air sacs connect the internal sides of the tympanal cavities. Muller’s organ contains sensory cells and is attached at several points to the internal side of the tympanum. (B) Cross section of a male cicada (Homoptera: Cicadidae). A single large air sac fills the abdominal cavity. Tympana are located inside grooves in the body. Sensory cells are located in an auditory cap- sule attached to the body wall and connected to the tympanum by a thin apodeme filament. (C) Cross section of a katydid (Orthoptera: Tettigoniidae). Funnel—shaped tracheal tubes connect the internal side of each tympanurn (two on each leg) to a large open spiracle on the body wall. Sensory cells are located on a thin membrane between tympana inside the leg. (D) Cross section of a cricket (Orthoptera: Gryl— lidae). The anatomy is similar to that of the katydid except that spiracles can be closed and tracheal tubes touch in mid-body, allowing for acoustical exchange. Sound Reception 159 tors. The two taxa differ in their mechanisms of frequency analysis. Grasshop- pers attach clusters of sensory cells directly to the tympanum and use the in- trinsic modes of vibration by the receptor-tympanal complex to provide fre— quency analysis. Locusts have about 70 sensory cells organized into four clusters collectively called Muller’s organ: subzones of the membrane with at- tached clusters have resonances of 3.5—4.1 kHz, 5.5—6.5 kHz, and 16—19 kHz (Michelsen and Nocke 1974; Stephen and Bennet—Clark 1982). Cicadas have a separate auditory organ containing up to 1000 sensory cells; this organ is at— tached to the tympanum by a thin rod (apodeme). Although not much is known about how these structures function, most species appear to have moderate peripheral tuning to the species—specific male advertisement fre— quency of their own species. Males are less narrowly tuned than females. In female cicadas, the air sac is small; it connects the two ears and thus permits pressure—differential directionality. Removal of the sac lowers female sensitiv— ity to sounds by 10—15 dB. In males, the sac is very large because of its func- tion as a resonant chamber to aid sound production. It also appears to res~ onate when exposed to nearby sounds at the species—specific frequency and thus acts as a sound reception coupling device. Removal of the sac in males reduces sensitivity to the species—specific frequency by 25 dB. In katydids (Orthoptera: Tettigoniidae) and crickets (Orthoptera: Grylli— dae), the tympana are located in the tibia of the front legs. Most species have a tympanum on each side of each foreleg, but some crickets may have tym— pana on one side only. Inside the leg is a tapered tracheal tube with its small opening behind the tympanum and its large opening at a spiracle on the ani— mal’s thorax (Figure 6.7). These tubes have two functions. First, they consti— tute amplification horns that increase the intensity of sound at the membrane by 10—30 dB. Without such amplification, the sound pressure on the back side of the tympanum would be much lower because of the longer distance the sound has traveled in a small tube. This construction generates maximal dis— placement of the tympanum when the exterior and interior sounds are out of phase. The second function is to make the tympanum a pressuredifferential detector with concomitant directional properties, at least at lower frequencies. In crickets, the situation is even more complex because the two tracheal tubes in the thorax are acoustically coupled. This means that the sound reaching any given tympanum arrives from four sources: directly outside the tympa— num, through a spiracle on the same side of the body, through a spiracle on the opposite side of the body, and through the tympanum on the other front leg. Michelsen (1979) reports that the sounds arriving from the opposite tra— cheal horn have only 35% of the energy directly striking a tympanum, and that arriving from the other front leg’s tympanum is less than 10—20% of the direct signal. The availability of pressure-differential detection allows these insects to use lower frequencies and still have the benefits of directionality. Perhaps be- cause of their connections between ears across the body, crickets have partic— ularly availed themselves of this opportunity and their sexual advertisement calls are in the range of 2—8 kHz. Crickets have a second call, called the ...
View Full Document

Page1 / 19

lec18reading - Chapter Sound Reception SOUND RECEPTION IS...

This preview shows document pages 1 - 19. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online