Anatomy and Physiology of Hearing

John H. Mills
Samir S. Khariwala
Peter C. Weber

This chapter provides a brief summary of the most basic features of the anatomy and physiology of the ear. It is divided into sections on the external and middle ear, cochlea, and central nervous system (CNS). The focus is on the anatomic and physiologic bases of audition with an effort directed at functional features. Surgical anatomy, vasculature, and eustachian tube function are not discussed.

External Ear

The external ear consists of the pinna (auricle) and the external auditory canal from the meatus to the tympanic membrane (Fig. 129.1). The pinna of humans is composed mostly of cartilage and has no useful muscles. The center of the pinna, the concha, leads to the external auditory meatus, which is about 2.5 cm long. The lateral third of the canal is the cartilaginous portion. It contains cerumen-producing glands and hair follicles. The remaining medial two thirds is the bony portion, including an epithelial lining over the tympanic membrane (1).
The external ear and the head have a passive but important role in hearing because of their acoustic properties. The concha, or bowl of the auricle, has a resonance of about 5 kHz, and the irregular surface of the pinna introduces other resonances and antiresonances. These acoustic features are useful to help differentiate whether sound sources are in front of the listener or behind.
The external auditory canal (EAC) is essentially a tube that is open at one end and closed at the other; thus the EAC behaves like a quarter-wave resonator. The resonant frequency (f0) is determined by the length of the tube; the curvature of the tube is irrelevant. For a tube of 2.5 cm, the resonant frequency is approximately 3.5 kHz:f0 = Velocity of sound @ 350m/s/(4×2.5 cm)
A flat, wide-band sound measured in a sound field is changed considerably by the acoustic properties of the head and external ear. As Figure 129.2 demonstrates, a gain of about 15 dB occurs in the 3-kHz range of the human, cat, and chinchilla, and 10 dB between 2 and 5 kHz. The acoustic properties of the external ear are one of the reasons noise-induced hearing losses occur first and most prominently at the 4-kHz frequency region (boilermaker notch).
In addition to the prominence of noise-induced hearing loss in the 4-kHz region, the acoustic properties of the head and external ear have an important role in several hearing functions. In localization of sound sources, the head acts as an attenuator at frequencies at which the width of the head is greater than the wavelength of the sound. Thus at frequencies greater than 2 kHz, a head shadow effect occurs, in which interaural intensity differences of 5 to 15 dB are used to localize sound sources. At lower frequencies, at which the wavelength of the sound is larger than the width of the head, little attenuation is provided by the head. Interaural time differences (~0.6 ms for sound to travel across the head) are the salient cues for localization. The head-shadow effect is the reason right-handed hunters using rifles and shotguns have larger hearing losses in their left ears than in their right ears and vice versa. The muzzle of the gun, where the acoustic energy is greatest, is closer to the left ear, and the right ear is protected by the head-shadow effect.
The 10- to 15-dB gain provided by the external ear in the 3- to 5-kHz region is useful for improving the detection and recognition of low-energy, high-frequency sounds such as voiceless fricatives. The importance of the acoustic properties of the external ear and head is reflected in hearing-aid design and evaluations. Finally, the resonance of the external canal is approximately 8 kHz in infants and decreases to adult values after approximately 2.5 years of age. This developmental feature has several clinical implications, especially for sound-field testing and for hearing-aid design and evaluation of infants.

Middle Ear

The middle ear transmits acoustic energy from the air-filled EAC to the fluid-filled cochlea. It functions as an impedance-matching device inasmuch as it couples the low impedance of air to the high impedance of the fluid-filled cochlea. The impedance match is achieved in three ways. The first and most important factor is that the effective vibratory area of the tympanic membrane is approximately 17 to 20 times greater than the effective vibratory area of the stapes footplate (Fig. 129.3). A second factor involves the lever action of the ossicular chain. The arm of the long process of the incus is shorter, by a factor of 1.3, than the length of the manubrium and neck of the malleus. A third and minor factor is the shape of the tympanic membrane. The combined result of these three factors is a pressure gain of approximately 25 to 30 dB. The variance in published measurements of the transformer ratio is noteworthy. With the exception of studies of acoustic impedance of the ear, most data are from studies of human cadavers, with all of their shortcomings, or of animals, usually cats. In addition to its role in the transfer of power to the inner ear, the tympanic membrane protects the middle ear space from foreign material of the ear canal and maintains the air cushion that prevents insufflation of foreign material from the nasopharynx through the eustachian tube.
The vibratory behavior of the ossicular chain is described in Figure 129.3. The transformer action of the tympanic membrane and ossicular chain provides for relatively efficient transfer of power to the inner ear, and the fidelity of sound transmission across the middle ear is outstanding. Distortion of sound signals does not occur in the middle ear, even for input signals with sound levels greater than 130 dB sound pressure level (SPL).
The middle ear, including the tympanic membrane, ossicular chain with supporting ligaments, and middle ear space, can be viewed as a passive mechanical system with both mass and compliant elements and therefore resonant properties. This linear system is coupled to the cochlea, which contributes a large resistance. The result is a middle ear system that is highly damped and linear and has a wide frequency response. The input–output function or transfer function of the middle ear is shown in Figure 129.4A. The ratio of the volume velocity of the stapes to sound pressure at the tympanic membrane increases in humans to approximately 800 to 900 Hz, which is the resonant frequency of the middle ear, and decreases at higher frequencies. Phase shift or time lag between movement of the tympanic membrane and the stapes generally increases with frequency (Fig. 129.4B). Although the middle ear is an impressive system in terms of frequency response, linearity, and transformer properties, considerably less than half of the power entering the middle ear actually reaches the cochlea because of the absorption of energy by the ligaments and middle ear. As shown in Figure 129.5, the human middle ear is particularly inefficient at frequencies greater than 2 kHz, especially in comparison with the ears of cats and chinchillas. It also is important to recall that a 50% loss of power is a loss of only 3 dB.
Auditory function is profoundly affected by cochlear impedance as well as the combined acoustic effects of the head, external ear, and middle ear. The combined effects of the acoustic properties of the head, external ear, and middle ear, as well the input impedance of the cochlea, have a profound effect on auditory function. For example, these factors determine the shape of the audibility curve and therefore the frequency range of human hearing (Fig. 129.6). For example, humans do not detect and recognize sounds greater than approximately 20 kHz because such high-frequency sounds are not transmitted efficiently through the middle ear to the cochlea. A second example of this sound transformation is shown in Figure 129.7, in which the spectrum of a cannon measured in a sound field is compared with the spectrum of the cannon by the time it is transformed and shaped by the acoustic properties of the external ear, head, middle ear, and input impedance of the cochlea. Low-frequency energy is not transmitted to the cochlea, and the frequency region of greatest energy concentration is 3 to 4 kHz. Thus, these acoustic properties are primarily responsible for the ability of intense low-frequency sounds (measured in a sound field) to produce high-frequency hearing losses and injuries in the basal region of the cochlea.
Two striated muscles, the tensor tympani and the stapedius, are located in the middle ear. The former attaches to the malleus and is innervated by the trigeminal nerve. The stapedius muscle attaches to the stapes and is innervated by the stapedial branch of the facial nerve. Noticeably the stapedius and tensor tympani muscles are the smallest striated muscles in the body and also have a high innervation ratio, that is, nerve fibers per muscle fiber. Although no question remains that contraction of these muscles affects sound transmission through the middle ear, the details of the effect and the extent of the influence of the middle ear muscles are still not fully understood. A number of disparate functions have been attributed to the middle ear muscles.
One function of the middle ear muscles is to protect the cochlea from loud sounds (2). When sounds louder than approximately 80 dB SPL are presented monaurally or binaurally, consensual (bilateral) reflex contraction of the stapedius muscle occurs. This contraction increases the stiffness of the ossicular chain and tympanic membrane, attenuating sounds less than approximately 2 kHz. Although the tensor tympani contracts as part of a startle response, acoustic reflex data from human subjects with neurologic involvement of cranial nerves V and VII suggest that the tensor tympani does not normally respond to intense acoustic stimulation. Laboratory and field studies of noise-induced hearing loss have shown convincingly that the stapedial reflex protects the cochlea, particularly from low-frequency (<2 kHz) sounds in excess of 90 dB. Inasmuch as the latency of the acoustic reflex is greater than 10 ms, the cochlea may be unprotected from short-duration, unanticipated impulsive sounds.
The following functions have been attributed to the middle-ear muscles. Some of these functions include providing strength and rigidity to the ossicular chain; contributing to the blood supply of the ossicular chain; reducing physiologic noise caused by chewing and vocalization; improving the signal-to-noise ratio for high-frequency signals, especially high-frequency speech sounds such as voiceless frica-tives, by means of attenuating high-level, low-frequency background noise; functioning as an automatic gain control and increasing the dynamic range of the ear; and smoothing out irregularities in the middle-ear transfer function.

Cochlea

The human cochlea is a coiled, bony tube approximately 35 mm long, divided into the scala vestibuli, scala media, and scala tympani (Fig. 129.8). The scalae vestibuli and tympani contain perilymph, an extracellular fluid-like material with a potassium concentration of 4 mEq/L and a sodium concentration of 139 mEq/L. The scala media is bounded by the Reissner membrane, the basilar membrane and osseous spiral lamina, and the lateral wall. It contains endolymph, an intracellular-like fluid with a potassium concentration of 144 mEq/L and a sodium concentration of 13 mEq/L. The scala media has a positive direct current (DC) resting potential of approximately 80 mV that decreases slightly from base to apex. This endocochlear potential is produced by the heavily vascularized stria vascularis of the lateral wall of the cochlea. The sodium–potassium–adenosine triphosphatase (Na+-K+-ATPase) pumps in a number of specialized cells of the stria vascularis contribute to this potential (3).
Acoustic energy enters the cochlea through the piston-like action of the stapes footplate on the oval window and is coupled directly to the perilymph of the scala vestibuli. The perilymph of the scala vestibuli communicates with the perilymph of the scala tympani through a small opening at the apex of the cochlea known as the helicotrema. The organ of Corti rests on the basilar membrane and osseous spiral lamina (Fig. 129.9). The basilar membrane is approximately 0.12 mm wide at the base and increases to approximately 0.5 mm at the apex. The major components of the organ of Corti are the outer and inner hair cells, supporting cells (Deiters, Hensen, Claudius), tectorial membrane, and the reticular lamina–cuticular plate complex (Fig. 129.10). Supporting cells provide structural and metabolic support for the organ of Corti. The phalangeal processes of the Deiters cells form tight cell junctions of the reticular lamina.
Outer and inner hair cells of the organ of Corti are important in transduction of mechanical (acoustic) energy into electrical (neural) energy. Outer hair cells are radically different from inner hair cells. Figure 129.11 and Table 129.1 detail these differences (4). In addition to the morphologic differences between outer and inner hair cells, neural innervation is different (Fig. 129.12). The spiral ganglion, the cell body of the auditory nerve, sends axons to the cochlear nucleus of the brainstem, whereas the dendrite projects through the osseous spiral lamina. Of the 50,000 neurons that innervate the cochlea, 90% to 95% synapse directly on inner hair cells. These are called type I neurons. Each inner hair cell is innervated by approximately 15 to 20 type I neurons. In contrast, 5% to 10% of the 50,000 neurons innervate the outer hair cells (type II neurons). Each type II neuron branches to innervate approximately 10 outer hair cells. In addition to the afferent innervation pattern of the cochlea, approximately 1,800 efferent fibers, originating from the ipsilateral and contralateral superior olivary complex, project to the cochlea (Fig. 129.13).
Transduction is initiated by displacement of the basilar membrane in response to displacement of the stapes due to acoustic energy. The displacement pattern of the basilar membrane is a traveling wave (Fig. 129.14). The basilar membrane is stiffer at the base than in the apex. The stiffness component is distributed continuously. Therefore, the traveling wave always progresses from base to apex. The maximal amplitude of basilar membrane displacement varies as a function of stimulus frequency. Traveling waves produced by high-frequency sounds (10 kHz) have maximal displacement near the base of the cochlea, whereas the waves to low-frequency sounds (125 Hz) have the maximum toward the apical region. Traveling waves generated by high-frequency sounds do not reach the apical region of the cochlea, whereas waves to low-frequency sounds can travel the entire length of the basilar membrane.
In the past, the mechanical traveling wave was considered a broadly tuned response, with finer tuning introduced subsequently by transduction, the auditory nerve, and the CNS. Data obtained with sensitive recording and detection methods, however, have shown that the traveling wave has an extremely sharply tuned response (Fig. 129.15) and that many of the remarkable frequency-selective abilities of the ear can be explained by the mechanical properties of the cochlea
The mechanism by which the sharply tuned peak is generated within the mechanical traveling wave involves an enhancement known as the cochlear modifier. This is an activity of the outer hair cells that enhances the motion of the basilar membrane at frequencies near the best frequency of the particular cochlear location. This enhancement contributes to the fine frequency-selective abilities of the ear and to the sensitivity of the ear and ability to detect extremely faint sounds. The notion of an active process in the cochlea, the cochlear amplifier, is supported by the phenomenon of otoacoustic emissions. That is, when a short-duration signal is presented to the ear, an echo emanating from the cochlea can be recorded in the external auditory meatus. Because the energy of the echo can be greater than the energy of the short-duration signal, an active process, the cochlear amplifier, is assumed. Factors that may contribute to the cochlear amplifier include motility of outer hair cells and the mechanical properties of the stereocilia and tectorial membrane.
The stereocilia–hair cell complex is critical to transduction. Stereocilia are bundles of actin filaments that form tubes and are inserted into the cuticular plate. They also are cross-linked between themselves. Stereocilia of inner hair cells probably do not contact the tectorial membrane, but those of outer hair cells are in direct contact. Deflection of the stereocilia by the traveling wave opens and closes nonspecific ion channels at the tips of the stereocilia, resulting in current flow (potassium) into the sensory cell. The flow of potassium ions into the sensory cell is modulated by the opening and closing of ion channels of the stereocilia. The potassium flux is caused by the endocochlear potential of +80 mV added to the negative intracellular potentials of hair cells. The resulting intracellular depolarization causes an enzyme cascade involving calcium. This ultimately leads to the release of chemical transmitters, and the subsequent activation of the afferent nerve fibers.
Although the notion of the cochlea as an active rather than a passive organ is no longer debated, specific details of the cochlear amplifier and the biologic basis of its operation are under active investigation. One point of view attributes the cochlear amplifier to the ability of hair cells to contract and lengthen in response to electrical signals, a property called somatic electromotility. A protein named prestin has been identified in outer hair cells and is considered to be the motor protein of outer hair cells and the driving force of electromotility of hair cells (5). Another point of view focuses on rapidly acting potassium and calcium ion channels presumed to be the basis of the cochlear amplifier and its regulation (6). A third approach suggests that a collection of motor proteins within a hair cell can generate oscillations that depend on the elastic properties of the cell (7). The foregoing approaches are nonlinear models that involve rapidly acting calcium channels. Specification of the biologic basis of the cochlear amplifier (nonlinearity) is important inasmuch as many forms of hearing loss involve loss of the cochlear amplifier.
The neurotransmitters of the afferent and efferent systems are the subject of intense study. In regard to the afferent system, analysis of excitatory amino acid receptor expression by the techniques of reverse transcriptase–polymerase chain reaction, in situ hybridization, and immunochemical analysis indicates that glutamate is the afferent neurotransmitter. Glutamate has been detected in both spiral ganglion cells and sensory cells (8). The principal transmitter substance of cochlear efferent fibers is acetylcholine. It is possible that the organ of Corti is mechanically modified by means of motility changes of outer hair cells under the influence of the efferent system. Acetylcholine acts on receptors to produce hyperpolarization of the cell membrane and doubling of the input conductance of the cell. The acetylcholine receptor has both muscarinic and nicotinic features. In addition to acetylcholine, γ-aminobutyric acid and several neuroactive peptides are neurotransmitters for the efferent system (9,10).

Gross Cochlear Potentials
Four gross (extracellular) potentials can be recorded in the cochlea (11)—endolymphatic (endocochlear) potential, cochlear microphonic, summating potential, and whole-nerve action potential (Fig. 129.16). Unlike the other cochlear potentials, the endolymphatic potential is not generated in response to acoustic stimulation; rather, it is a DC potential of 80 to 100 mV recorded in the scala media. It arises from the stria vascularis of the lateral wall of the cochlea. The stria vascularis is considered to be the energy source, or “battery,� of the cochlea, crucial for transduction. The nature of the energy source is related to the heavy vasculature of the stria vascularis and to the Na+-K+-adenosine triphosphatase (ATPase). This pump has been localized to several types of cochlear cells, including marginal cells of the stria vascularis, outer sulcus cells, and fibrocytes near the attachment of the Reissner membrane and in the spiral ligament. Whereas Na+-K+-ATPase must play a significant role in ion transport in the cochlea, the nature of the energy source and the details of the ion exchange remain active research issues (3).
Malfunctioning of the mechanisms involved in production of endolymph and the endolymphatic potential can produce hearing loss, sometimes called metabolic presbycusis. When the flow of endolymph through the ductus reuniens is blocked, endolymphatic pressure increases, and hydrops occurs.
The cochlear microphonic is an alternating current (AC) voltage usually recorded within the cochlea or near the round window. It represents the potassium ion current flow through mainly the outer hair cells; that is, the electrical resistance of outer hair cells is altered by the motion of the basilar membrane. When stereocilia are bent away from the modiolus, the resistance of the hair cells decreases. The result is an increase in current flow and a small decrease in endolymphatic potential. When stereocilia are bent toward the modiolus, resistance increases and current flow decreases with an accompanying increase in the endolymphatic potential. The corresponding voltage fluctuations, the cochlear microphonic, depend on the presence of outer hair cells. Unlike neural potentials, the waveform of the cochlear microphonic mirrors the motion of the basilar membrane. The summating potential is a DC potential recorded in the cochlea in response to sound. It follows the envelope of the stimulating sound. Recordings of this DC potential can be made in the scala tympani, media, or vestibuli and in some circumstances from a gross electrode in the human ear canal. The potential can be positive or negative, and it can reverse polarity, depending on electrode location or stimulus frequency and level. The summating potential probably has several origins, but it largely reflects the DC shifts caused by stimulus-driven intracellular potentials of outer hair cells. Inner hair cells contribute to these to a lesser extent.
The whole-nerve or compound action potential arises from the all-or-none discharge of auditory nerve fibers. The compound action potential is recorded most effectively with a gross electrode placed near the round window or auditory nerve and with high-frequency signals with rapid onsets. Such signals produce synchronous neural activity, which is summed to become the compound action potential waveform. The amplitude of the compound action potential increases with stimulus intensity over a 40- to 50-dB range, whereas latency decreases as stimulus intensity is increased. At high levels, a second peak sometimes is observed that probably reflects activity of the cochlear nucleus. The compound action potential can be clinically recorded with scalp electrodes or electrodes in the external meatus or by means of a transtympanic approach in which an electrode is placed near the round window niche. The ratio of the amplitude of the summating potential to the amplitude of the compound action potential has been used as an indicator of perilymphatic fistula, but the validity of this indicator is doubtful.

Eighth Nerve Physiology
The auditory nerve has approximately 30,000 fibers in humans and approximately 50,000 in cats. Perhaps one of the most important research findings in recent years was the observation that 90% to 95% of neurons (type I, radial fibers) innervate inner hair cells, whereas 5% to 10% (type II, outer spiral fibers) innervate to the outer hair cells (Fig. 129.12). Most, if not all, recordings from auditory nerve fibers are from the larger type I fibers in contact with inner hair cells. These radial fibers have bipolar cell bodies in the spiral ganglion. Outer spiral fibers are monopolar and unmyelinated. Most recordings of single units of the auditory nerve are obtained by means of inserting a microelectrode into the auditory nerve where it exits the internal auditory meatus. The most basic measures of auditory nerve function are spontaneous rates, tuning curves, and intensity (rate-level) functions.
Most auditory nerve fibers in mammals discharge in the absence of acoustic stimulation. The nerve fibers have been classified into three categories on the basis of rate of spontaneous discharge—high (18 to 120 spikes per second), medium (0.5 to 18 spikes per second), and low (0 to 0.5 spikes per second). Fibers with high rates of spontaneous activity respond to auditory signals at lower levels than do fibers with medium or low rates of spontaneous activity. In other words, the most-sensitive fibers have the most-spontaneous activity. Fibers with high spontaneous rates have thick dendrites that tend to terminate on the side of inner hair cells facing outer hair cells. Fibers with low and medium spontaneous rates have thin dendrites that terminate on the side of the inner hair cell facing the modiolus. Ongoing studies indicate that fibers with high rates of spontaneous activity have different terminations in the auditory CNS (cochlear nucleus) than do fibers with low rates of spontaneous activity. In other words, spontaneous activity of nerve fibers is not random but is proving to be anatomically and functionally significant (12,13,14,15). The tuning curve of a single auditory nerve fiber is perhaps the most basic measure of auditory nerve function. A tone burst controlled in frequency and level is presented. The level is adjusted until a criterion change (one or two spikes per second) in firing rate is detected. Tone bursts covering a wide range of frequencies are used, and the lowest level of signal is recorded for a given frequency that produces a specific rate of discharge. The resulting isoresponse curve is called a tuning curve. Figure 129.17 shows tuning curves for six different fibers. The sharp tip of the tuning curve identifies the best, or characteristic, frequency of the fiber. Units with low characteristic frequency are fibers that innervate inner hair cells in the apical region of the cochlea, fibers with high characteristic frequency innervate inner hair cells from the basal region, and so on. Tuning curves are described according to the frequency of the tip or characteristic frequency, the high- and low-frequency side, and the tail. Fibers with a characteristic frequency less than 1 kHz are roughly V shaped. Fibers with a higher characteristic frequency have an obvious tip at the characteristic frequency and a tail that extends to the low frequencies. The high side of a tuning curve is the frequency region greater than characteristic frequency. As characteristic frequency increases, the high side of the tuning curve becomes steeper with a slope or rejection rate that can exceed 500 dB per octave. The characteristics of tuning curves of auditory nerve fibers are strikingly similar to isoamplitude curves of a mechanical traveling wave (Fig. 129.15).
Injury or damage to sensory cells, including stereocilia, can alter the shape of tuning curves dramatically (Fig. 129.18). The lower right portion of the figure shows that when outer hair cells are destroyed, the tuning curve of auditory nerve fibers from normal inner hair cells is changed in several ways. The sensitive tip region is missing; that is, the threshold of the fiber is elevated by approximately 40 to 45 dB. The high-frequency side no longer has a steep slope, and the low-frequency side becomes slightly more sensitive, or hypersensitive. The characteristic frequency of the fiber appears to be much lower in frequency, and the band width of the fiber appears broader. The upper left portion of Figure 129.18 shows the consequences of partial injury to the stereocilia of outer hair cells. A threshold shift of approximately 30 dB occurs, but a short, sharply tuned tip remains, and the low-frequency tail is again hypersensitive. Irregularities in this tuning curve may explain monaural diplacusis; that is, a tone in one ear (800 Hz) has two pitches, for example, one at 800 Hz and a second at approximately 2.8 kHz.
The upper left portion of Figure 129.18 shows a tuning curve in which stereocilia of inner hair cells are damaged or in disarray, whereas most of the stereocilia of outer hair cells appear normal or nearly so. The threshold of the unit is elevated approximately 30 dB, but the tuning curve is approximately normal. The lower left portion of the figure shows responses to signals in a narrow range of frequencies only at sound levels greater than 90 dB SPL. In this case, sensory cells are present, but stereocilia of inner hair cells are destroyed, and those of outer hair cells are destroyed or in disarray. Thus normal neural activity, including sensitivity (detection of faint sounds) and frequency-resolving power, depends on intact outer hair cells and normal stereocilia.
Although thresholds of auditory nerve fibers are related to the rate of spontaneous discharge, most afferent nerve fibers (60%) have high spontaneous rates and thresholds within 20 dB greater than the thresholds for the animal. The remaining low-spontaneous fibers have thresholds that cover approximately 60 dB. The dynamic range of most auditory nerve fibers is approximately 30 dB from threshold to saturation (Fig. 129.19), although some low-spontaneous fibers have a much wider dynamic range. Given the dynamic range of human hearing (0 dB SPL to ≥100 dB SPL), the auditory system must have neurons the thresholds of which cover a wide range and have firing rates that also cover a wide range of intensities. The ability of the human ear to respond appropriately to sounds over a 120-dB range (10,12) is remarkable. One way is with low-spontaneous fibers; another is recruitment of fibers of characteristic frequency.
One of the most common features of sensorineural hearing loss is recruitment of loudness. Figure 129.20 gives an explanation. It is assumed that loudness depends on the total activity of the auditory nerve. As Figure 129.20A shows, the number of fibers activated increases slowly as intensity is increased, and only the tips of tuning curves are activated. As the intensity increases further, the tails of the tuning curves are encountered, and the number of fibers activated increases rapidly. In the case of sensorineural hearing loss, the tips of the tuning curves are missing, and the fibers are not activated until the level of the signal is sufficient to reach the tails of the tuning curves. Abruptly, many fibers then are abruptly activated simultaneously.

Nonlinear Properties of the Ear
Some of the outstanding features of the middle ear transformer are its linear properties, but the outstanding features of the cochlea and auditory nerve are the nonlinear characteristics. Perhaps the most studied nonlinearities are combination tones, described herein in relation to cochlear emissions, and two-tone rate suppression, as recorded in auditory nerve fibers.
Two-tone rate suppression is the reduction in firing rate produced by one tone when a second tone is introduced. Figure 129.21 shows a tuning curve with a suppression area outlined above the characteristic frequency of the nerve fiber and an area below the characteristic frequency of the fiber. Tones presented in the dotted or suppression areas in the figure reduce the firing rate caused by the probe tone. Both the excitor and suppressor tones are presented simultaneously, and because little or no time lag is associated with this phenomenon nor is any evidence available that it is neurally produced, the effect is called suppression rather than inhibition. Two-tone suppression in single units is reflected in the compound action potential. Figure 129.21 (right) shows tuning curves of the compound action potential with suppression areas shown in the dotted areas. In this case, the amplitude of the compound action potential is altered by the suppressing signal, whereas in the single-unit case (left), the firing rate of a neuron is reduced by an arbitrary amount (20%). The single-unit and compound action potential suppression areas are similar. Inasmuch as two-tone suppression can be observed in the DC intracellular response of inner hair cells, it is probable that two-tone suppression originates in the active nature of cochlear mechanics and before the inner hair cells.In the presence of sensorineural hearing loss caused by exposure to noise or to ototoxic drugs, two-tone rate suppression is severely affected, if at all measurable. Two-tone rate suppression appears normal or nearly so in cases of cochlear hearing loss in which the sensory cells, including stereocilia, are normal or nearly so, but the stria vascularis is affected. The latter scenario leads to presbycusis (16).
Otoacoustic emissions (OAEs) are sounds that are detected in the ear canal when the tympanum receives vibrations transmitted through the middle ear from the cochlea. OAEs provide support for the notion that the cochlea is not just a passive receiver of acoustic energy but can also generate or amplify sounds. Several different types of OAEs are found (17). Spontaneous OAEs occur in the absence of acoustic stimulation and are typically highly stable pure tones of -10 to 30 dB SPL, which are found in 30% to 40% of healthy young ears (18,19). The precise frequency of a spontaneous OAE does not imply an origin at a precise place in the cochlea, but only a particular coincidence of travel time and reflection from an ill-defined region of high outer cell activity. Spontaneous OAEs can be recorded over long periods with only minor but seemingly systematic variations in frequency and amplitude.A second class of OAEs are produced after exposure to an acoustic signal. Transient-evoked OAEs (TEOAE) are made via a probe placed in the ear canal. The oscillatory sound pressure waveform seen in TEOAE responses actually corresponds to the motion of the eardrum resulting from pressure fluctuations generated within the cochlea (Fig. 129.22). Although stimulatory clicks excite the entire cochlea, TEOAE responses can be used to give frequency-specific information about the cochlea through splitting of the responses into different frequency bands. TEOAEs are highly sensitive to cochlear pathology in frequency-specific manner. Frequencies at which hearing thresholds exceed 20 to 30 dB hearing loss (HL) are typically absent in the TEOAE response (20,21). Because of their sensitivity to cochlear dysfunction, TEOAEs have found widespread application in newborn hearing screening programs (22).
Distortion-product OAEs also are used widely in clinical situations. The TEOAE and DPOAE techniques complement each other. DPOAEs offer a wider frequency range of observation with less sensitivity to minor and subclinical conditions in adults. When two primary tones, F1 and F2, are presented to the cochlea, several distortion products are produced. The most prominent of all these intermodulation distortion products is the cubic distortion tone, 2F1-F2. Measurement of DPOAEs at multiple stimulus levels can establish the OAE “growth rate.� Healthy ears tend to exhibit a DPOAE growth rate of 1 dB of OAE per 1 dB of stimulus or less. Ears with some impairment show steeper growth. Single DPOAE results can be misleading, and results must be averaged across a range of frequencies. The DPOAE is easily recordable in patients with a normal middle ear system (23).Auditory Central Nervous System
The ascending and descending auditory pathways are described briefly herein in relation to auditory evoked potentials. Schematics of the afferent and efferent pathways are shown in Figs. 129.23 and 129.13, respectively. These diagrams oversimplify the system but provide a rough introduction to the auditory CNS and its complexity. All eighth-nerve afferent fibers stop at the level of the cochlear nucleus. Five major cell types are found within the cochlear nucleus, each with distinct cell morphologic and physiologic features, such as response to stimulus onset, stimulus offset, and frequency modulation. From the cochlear nucleus, most fibers cross the brainstem to the contralateral superior olivary complex; a much smaller number of fibers run to the ipsilateral superior olivary complex.
The superior olivary complex is considered the first center in the ascending auditory system, where inputs from both ears converge. Auditory nuclei above the superior olivary complex can be excitatory or inhibitory with inputs from each ear. Stimulation of the contralateral ear typically is excitatory to cell bodies of the auditory CNS, whereas stimulation of the ipsilateral ear is inhibitory. As shown in Figure 129.13, the medial superior olivary complex is the origin of the crossed efferent fibers that terminate on outer hair cells, whereas the lateral superior olivary complex is the origin for the uncrossed efferent fibers that terminate on inner hair cells. Although many functions have been attributed to the efferent auditory system, especially protecting the cochlea from loud sounds, the functions of the system are unknown; those that have been proposed are easily debated
The inferior colliculus is a complex nucleus with at least 18 major cell types and at least five areas of specialization. It is involved in probably all forms of auditory behavior, including differential sensitivity for frequency and intensity, loudness, and binaural hearing. The inferior colliculus is clearly more than a relay center. The medial geniculate body of the thalamus sends projections to the auditory cortex, but its specific functions are unknown.
The auditory cortex is located in the sylvian fissure of the temporal lobe; many secondary auditory areas are clustered around the primary area. In each area, the cells are tonotopically organized in a columnar manner, each column having a special attribute. The cells in one column can have different tuning at a similar characteristic frequency, whereas another column can be associated with intensity encoding, another with providing inhibitory responses to stimulation of one ear and excitatory responses of the other ear, and so on. As is common for thalamic connections with the cortex, nuclei within the medial geniculate body that send fibers to the auditory cortex also receive fibers from the same area of the cortex. Bilateral lesions of the temporal lobe have been shown to produce wide-ranging effects (cortical deafness, in which several auditory behaviors are severely affected, including speech discrimination, localization of sound, temporal processing of information, and the detection of faint, short-duration signals) (25). Another important feature of the auditory system is its tonotopic nature. From the basilar membrane to the auditory cortex, the system is organized spatially with respect to frequency. Each place on the basilar membrane responds best to a specific frequency—high-frequency sounds are localized to the base, and low-frequency sounds, to the apex. The tonotopic organization of the cochlea is preserved at the cochlear nucleus. Figure 129.24 shows that as an electrode penetrates the cochlear nucleus, fibers with different characteristic frequencies are contacted, and the characteristic frequencies form an orderly progression. Similar data exist at all nuclei of the auditory CNS, including the auditory cortex
The most obvious clinical application of basic information on the auditory CNS involves interpretation of evoked potentials. The auditory brainstem response (ABR) is one component of auditory evoked potentials. The existence of the ABR was first reported by Sohmer and Feinmesser in 1967 (26). The ABR is recorded from electrodes attached to various positions on the head. The ABR consists of a series of seven waves occurring within about 10 milliseconds after stimulus onset. The convention in the United States is to label wave peaks with Roman numerals. It is generally accepted that the ABR is generated by the auditory nerve and subsequent fiber tracts and nuclei within the auditory brainstem pathways. It is widely believed that each wave is generated as follows: wave I and II are the eighth nerve, III is cochlear nucleus, IV is superior olive/lateral lemniscus, and V is lateral leminiscus/inferior colliculus.
The ABR is generated by a click stimulus because it yields the clearest response. The ABR is used clinically both in the estimation of auditory sensitivity and in otoneurologic assessment. In this way, it can be used to detect lesions along the auditory nerve and brainstem pathways. The study can be performed regardless of state of wakefulness, and the result is unaffected by most medications. As a result, children are often tested while under sedation or during sleep.
The field of clinical objective audiometry has recently gained an additional technique in the auditory evoked response battery. The auditory steady-state response (ASSR) promises to be a valuable study in the workup of auditory dysfunction. Unlike ABRs, which are obtained through the use of transient stimuli, ASSRs are evoked by using sustained continuous tones. The tones are frequency specific because the continuous tones do not have spectral distortion problems as do brief tone bursts or click (27). Of note, ASSR also can be performed regardless of the state of wakefulness.
There are several advantages of ASSR over ABR. First, ASSR is a better technique for evaluating hearing aid performance because hearing aids and cochlear implants process continuous stimuli with less signal distortion than transient stimuli. Furthermore, ASSR can provide threshold information in a frequency-specific manner at intensity levels of 120 dB or greater (28,29). This allows differentiation of severe and profound hearing loss, which cannot be accomplished with ABR. This characteristic of ASSR may allow it to be used in assessing pediatric patients for cochlear implant candidacy (30). Last, ASSR has been shown to be more time efficient by determining more thresholds in a shorter time compared with ABR (31). Future research and clinical use are likely to solidify the status of ASSR in the audiologic armamentarium.
The neuroanatomic features of the system are complicated. Processing of neural information probably involves both parallel and serial processing. The former is anatomically described by a single fiber with ramifications to many target areas. Serial processing involves a fiber going to one target, which in turn goes to another target, and so forth. In the auditory CNS, both serial and parallel processing are involved. Because the auditory CNS is a highly redundant, complicated, and extremely powerful system, interpretation of evoked-potential data, and of other CNS neural data, is not straightforward.