30: The Inner Ear

• The Ear Has Three Functional Parts

• Hearing Commences with the Capture of Sound Energy by the Ear

• The Hydrodynamic and Mechanical Apparatus of the Cochlea Delivers Mechanical Stimuli to the Receptor Cells

  • The Basilar Membrane Is a Mechanical Analyzer of Sound Frequency

  • The Organ of Corti Is the Site of Mechanoelectrical Transduction in the Cochlea

• Hair Cells Transform Mechanical Energy into Neural Signals

  • Deflection of the Hair Bundle Initiates Mechanoelectrical Transduction

  • Mechanical Force Directly Opens Transduction Channels

  • Direct Mechanoelectrical Transduction Is Rapid

• The Temporal Responsiveness of Hair Cells Determines Their Sensitivity

  • Hair Cells Adapt to Sustained Stimulation

  • Hair Cells Are Tuned to Specific Stimulus Frequencies

  • Sound Energy Is Mechanically Amplified in the Cochlea

• Hair Cells Use Specialized Ribbon Synapses

• Auditory Information Flows Initially Through the Cochlear Nerve

  • Bipolar Neurons in the Spiral Ganglion Innervate Cochlear Hair Cells

  • Cochlear Nerve Fibers Encode Stimulus Frequency and Intensity

• Sensorineural Hearing Loss Is Common but Treatable

• An Overall View

Human experience is enriched by our ability to distinguish a remarkable range of sounds—from the intimacy of a whisper to the warmth of a conversation, from the complexity of a symphony to the roar of a stadium. Hearing begins when the cochlea, the snail-shaped receptor organ of the inner ear, transduces sound energy into electrical signals and forwards them to the brain. Our ability to recognize small differences in sounds stems from the cochlea's capacity to distinguish among frequency components and to inform us of both the tones present and their amplitudes.

Deafness can be devastating. For the elderly, hearing loss can result in a painful and protracted estrangement from family, friends, and colleagues. Children may lack hearing as a result of pre- or perinatal infections and especially of genetic conditions, which affect one child in a thousand. Such children are often deprived of the normal avenue to the development of speech, and thus of reading and writing as well. It is for this reason that a modern pediatric examination must include an assessment of hearing. Many children thought to be cognitively impaired are found instead to be hard of hearing, and their intellectual development resumes its normal course when this problem is corrected.

Acute hearing loss in the intermediate years exacts an enormous price for two reasons. First, hearing plays an important, but often overlooked, role in our psychological well-being. Daily conversation with family and colleagues helps to establish our social context. The abrupt loss of such social intercourse as a result of sudden deafness leaves a person distressingly lonely and may lead to depression and even suicide. Hearing also serves us in another, more subtle way. Our auditory system is a remarkably efficient early warning system, subconsciously informing us about our environment. For example, when other people enter the room or approach us we often hear them before we see them. More obviously, awareness of fire alarms and the sirens of emergency vehicles can be lifesaving. Deafness may leave a person with an ominous sense of vulnerability to unheard changes in the environment.

Hearing loss is often accompanied by another distressing symptom, tinnitus, or ringing in the ears. By interfering with concentration and disrupting sleep, tinnitus can exasperate, depress, and even madden its victims. Because on rare occasions tinnitus stems from lesions to the auditory pathways, such as acoustic neuromas, it is important in neurological diagnosis to exclude such causes. Most tinnitus, however, is idiopathic: Its cause is uncertain. Some drugs trigger the condition; antimalarial drugs related to quinine and aspirin at the high dosages used in the treatment of rheumatoid arthritis are notorious for this. Often, however, tinnitus occurs at high frequencies to which a damaged ear is no longer sensitive. In these instances tinnitus may reflect hypersensitivity in the deafferented central nervous system, a phenomenon analogous to phantom-limb pain (see Chapter 24).

Hearing depends on the remarkable properties of hair cells, the receptors of the internal ear. Hair cells receive mechanical inputs that correspond to sounds and transduce these signals into electrical responses that are forwarded to the brain for interpretation. These cells can measure motions of atomic dimensions and transduce stimuli ranging from static inputs to those at frequencies of tens of kilohertz. Hair cells also serve as mechanical amplifiers that augment our auditory sensitivity. Each of the paired cochleas contains approximately 16,000 of these cells. Deterioration of hair cells accounts for most of the hearing loss that afflicts more than 30 million Americans.

Sound consists of alternating compressions and rarefactions propagated by an elastic medium, the air, at a speed of approximately 340 m/s. As we are reminded on making the effort to shout, producing these pressure changes requires that work be done on the air by our vocal apparatus or some other sound source. Each of our ears must capture this mechanical energy, transmit it to the receptor organ, and transduce it into electrical signals suitable for neural analysis. These three tasks are the functions of, respectively, the external ear, the middle ear, and the inner ear (Figure 30–1).

The most obvious component of the human external ear is the auricle, a prominent fold of cartilage-supported skin. Much as a parabolic antenna collects electromagnetic radiation, the auricle acts as a reflector to capture sound efficiently and to focus it into the external auditory meatus, or ear canal. The external auditory meatus ends at the tympanum or eardrum, a thin diaphragm approximately 9 mm in diameter.

The external ear is not uniformly effective at capturing sound from any direction; the auricle's corrugated surface collects sounds best when they originate at different, but specific, positions with respect to the head. Our capacity to localize sounds in space, especially along the vertical axis, depends critically on the sound-gathering properties of the external ear.

The middle ear is an air-filled pouch connected to the pharynx by the Eustachian tube. Airborne sound traverses the middle ear as vibrations of three tiny ossicles, or bones: the malleus (hammer), incus (anvil), and stapes (stirrup). The base of the malleus is attached to the tympanic membrane; its other extreme makes a ligamentous connection to the incus, which is similarly connected to the stapes. The flattened termination of the stapes, the footplate, inserts in an opening—the oval window—in the bony covering of the cochlea. The first two ossicles are relics of evolution, for their antecedents served as components of the jaw in reptilian ancestors.

The inner ear, or cochlea (from the Greek cochlos, meaning snail), is a coiled structure of progressively diminishing diameter wound like a snail's shell around a conical bony core (Figure 30–1). It is approximately 9 mm across, the size of a chickpea. The cochlea is covered with a thin layer of laminar bone and embedded within the dense structure of the temporal bone. The inner and outer aspects of the cochlea's bony surface are lined with layers of connective tissue, the endosteum and periosteum.

The interior of the cochlea consists of three liquid-filled compartments termed scalae (Latin, meaning staircases). In a cross section of the cochlea at any position along its spiral course, the compartment farthest from the base is the scala vestibuli (Figure 30–2). At the basal end of this chamber is the oval window, an opening that is sealed by the footplate of the stapes. The chamber nearest the cochlear base is the scala tympani; it too has a basal aperture, the round window, which is closed by a thin, elastic diaphragm. The two chambers are separated along most of their length by the cochlear partition or duct. The scala vestibuli and scala tympani communicate with one another at the helicotrema, where the cochlear partition terminates slightly below the cochlear apex.

The third liquid-filled cavity, the scala media, lies within the cochlear partition. A pair of elastic structures separates the scala media from the other ducts (Figure 30–2). The thin Reissner's membrane, or vestibular membrane, divides the scala media from the scala vestibuli. The basilar membrane separates the scala media from the subjacent scala tympani; it is a complex structure involved in auditory transduction.

Psychophysical experiments have established that we perceive an approximately equal increment in loudness for each tenfold increase in the amplitude of a sound stimulus. This type of relation is characteristic of many of our senses and is the basis of the Weber-Fechner law (see Chapter 21). A logarithmic scale is therefore useful in relating quantitatively sound intensity to perceived loudness. The sound-pressure level, L, of any sound may be expressed in decibels as

in which P, the magnitude of the stimulus, is the root-mean-square sound pressure (in units of pascals, abbreviated Pa, or newtons per square meter). For a sinusoidal stimulus the amplitude exceeds the root-mean-square value by a factor of √2. The reference level on this scale, 0 dB, corresponds to a root-mean-square sound pressure, P _REF, of 20 μPa. This level represents the approximate threshold of human hearing at 1 to 4 kHz, the frequency range in which our ears are most sensitive.

That sound consists of alternating changes in the local air pressure is evident when a loud noise rattles a window. The loudest sound tolerable to humans, with an intensity of approximately 120 dB, transiently alters the local atmospheric pressure by only ±0.01%. Acting on a window 1 m on each side, this oscillatory pressure nonetheless produces a force of ±14 newtons. To rattle the same window by pushing upon it, we would need to exert a comparable force, approximately ±3 lb. In contrast, a tone at the threshold level causes a fractional change in the local pressure of much less than one part in a billion.

Despite their small magnitude, sound-induced increases and decreases in air pressure push and pull effectively upon the tympanum, moving it inward and outward (Figure 30–3A, B). The subsequent motions of the ossicles are complex, depending on both the frequency and the intensity of sound. In simple terms, however, the actions of these bones may be understood as those of two interconnected levers (the malleus and incus) and a piston (the stapes). The vibration of the incus alternately drives the stapes deeper into the oval window and retracts it. The footplate of the stapes thus serves as a piston that pushes and pulls cyclically upon the liquid in the scala vestibuli. The overall effect of the middle ear is to match the impedance of the air outside the ear to that of the cochlear partition, thus ensuring the efficient transfer of sound energy from the first medium to the second.

The action of the stapes at the oval window produces changes in pressure that propagate throughout the liquid of the scala vestibuli at the speed of sound. Because the aqueous perilymph is virtually incompressible, however, the primary effect of the stapes's motion is to displace the liquid in the scala vestibuli in the one direction that is not restricted by a rigid boundary: toward the elastic cochlear partition (Figure 30–3B). The deflection of the cochlear partition downward increases the pressure in the scala tympani. The enhanced pressure displaces a liquid mass that causes outward bowing of the round window. Each cycle of a sound stimulus thus evokes a cycle of up-and-down movement of a minuscule volume of liquid in each of the inner ear's three chambers.

Because the energy associated with acoustic signals is generally quite small, compromise of the middle ear's normal structure may lead to conductive hearing loss, of which two forms are especially common. First, scar tissue caused by middle-ear infection (otitis media) can immobilize the tympanum or ossicles. Second, a proliferation of bone in the ligamentous attachments of the ossicles can deprive the ossicles of their normal freedom of motion. This chronic condition of unknown origin, termed otosclerosis, can lead to severe deafness.

A clinician may test for conductive hearing loss by the simple Rinné test. A patient is asked to compare the loudness of a vibrating tuning fork held in the air near an affected ear with that perceived when the base of the tuning fork is pressed against the patient's head just behind the auricle. If the latter stimulus is perceived to be louder, the patient's conductive pathway may be damaged but the internal ear is likely to be intact. In contrast, if bone conduction is not more efficient than airborne stimulation, the patient may have inner-ear damage, that is, sensorineural hearing loss. The diagnosis of conductive hearing loss is important because surgical intervention is highly effective: Removal of scar tissue or reconstitution of the conductive pathway with an inert prosthesis can in many instances restore excellent hearing.

The Basilar Membrane Is a Mechanical Analyzer of Sound Frequency

The mechanical properties of the basilar membrane are key to the cochlea's operation. To appreciate this, suppose that the basilar membrane had uniform dimensions and mechanical properties along its entire length, approximately 33 mm. Under these conditions a fluctuating pressure difference between the scala vestibuli and the scala tympani would move the entire basilar membrane up and down with similar excursions at all points (Figure 30–3C).

This would occur regardless of the frequency of stimulation; any pressure difference between the scala vestibuli and the scala tympani would propagate throughout those chambers within microseconds, so the basilar membrane would be subjected to similar forces at any position along its length. This simple form of basilar-membrane motion in fact occurs in the auditory organs of some reptiles.

In reality, however, the mechanical properties of the mammalian basilar membrane vary continuously along the cochlea's length. The basilar membrane at the apex of the human cochlea is more than five times as broad as at the base. Thus, although the cochlear chambers become progressively larger from the organ's apex toward its base, the basilar membrane decreases in width. Moreover, the basilar membrane is relatively thin and floppy at the apex of the cochlea but thicker and tauter toward the base. Because its mechanical properties vary along its length, the basilar membrane does not oscillate like a single string on a musical instrument; instead, it more resembles a panoply of strings that vary from the coarsest string on a bass viol to the finest string on a violin.

Stimulation with a pure tone evokes a complex and elegant movement of the basilar membrane. Over one complete cycle of a sound, each affected segment along the basilar membrane undergoes a single cycle of vibration (Figure 30–3D, E). The various parts of the membrane do not, however, oscillate in phase with one another; instead, some portions of the membrane move upward while others move downward. As first demonstrated by Georg von Békésy using stroboscopic illumination, the overall pattern of motion of the membrane is that of successive traveling waves.

Because mechanical stimulation is applied at the cochlear base as a pressure difference between scala vestibuli and scala tympani, the traveling wave progresses from the cochlear base toward the apex. Each wave reaches its maximal amplitude at a particular position appropriate for the frequency of stimulation, then declines rapidly in size as it advances toward the cochlear apex. A traveling wave ascending the basilar membrane resembles an ocean wave rolling toward the shore: As the wave nears the beach its crest grows to a maximal height, then breaks and rapidly fades away.

Although the analogy of an ocean wave gives some sense of the appearance of the basilar membrane's motion, the connection between the cochlear traveling wave and the movement of an ocean wave is only metaphorical—the physical bases of the two waves are quite distinct. An ocean wave is the result of the momentum of a wind-blown mass of water. In contrast, movement of the basilar membrane is the result of motion of the liquid masses above and below the membrane. These liquids are continuously driven up and down by the energy supplied by the stapes's piston-like movements at the oval window.

The variation in the mechanical properties of the mammalian basilar membrane explains why the membrane is tuned to different frequencies at each point along its length. In humans the membrane at the apex of the cochlea responds best to the lowest audible frequencies, down to approximately 20 Hz. At the cochlear base it responds to frequencies as great as 20 kHz. The intervening frequencies are represented along the basilar membrane in a continuous array (Figure 30–3F). In the 19th century the German physiologist Hermann von Helmholtz was the first to appreciate that the basilar membrane's operation is essentially the inverse of a piano's. The piano synthesizes a complex sound by combining the pure tones produced by numerous vibrating strings; the cochlea deconstructs a complex sound by isolating each component tone at a discrete segment of the basilar membrane.

The arrangement of vibration frequencies along the basilar membrane is an example of a tonotopic map. The relation between frequency and position on the basilar membrane varies smoothly and monotonically but is not linear. Instead, the logarithm of the best frequency is roughly proportional to the distance from the cochlea's apex. The frequencies from 20 Hz to 200 Hz, those between 200 Hz and 2 kHz, and those spanning 2 kHz to 20 kHz are each apportioned approximately one-third of the basilar membrane's extent.

Analysis of the response to a complex sound illustrates how the basilar membrane operates in daily life. A vowel sound in human speech, for example, ordinarily comprises three dominant frequency components termed formants. Measurement of the sound pressure outside an ear exposed to such a sound reveals a complex, seemingly chaotic signal. The movements of the tympanum and ossicles in response to a vowel sound likewise appear very complicated. The motion of the basilar membrane, however, is much simpler. Each frequency component of the stimulus establishes a traveling wave that, to a first approximation, is independent of the waves evoked by the others (Figure 30–3G). Each traveling wave reaches its peak excursion at a point on the basilar membrane appropriate for that frequency component. Moreover, the amplitude of each traveling wave is proportional, albeit in a complex way, to the intensity of the corresponding frequency component. The basilar membrane thus acts as a mechanical frequency analyzer by distributing specific stimulus energies to hair cells arrayed along its length, and in doing so begins the encoding of the frequencies and intensities in a sound.

The Organ of Corti Is the Site of Mechanoelectrical Transduction in the Cochlea

The organ of Corti, a ridge of epithelium extending along the basilar membrane, is the receptor organ of the inner ear. Each organ of Corti contains approximately 16,000 hair cells innervated by approximately 30,000 afferent nerve fibers, which carry information into the brain along the eighth cranial nerve. Like the basilar membrane itself, both the hair cells and the auditory nerve fibers are tonotopically organized: At any position along the basilar membrane the hair cells are most sensitive to a particular frequency, and these frequencies are logarithmically mapped in ascending order from the cochlea's apex to its base.

The organ of Corti includes a variety of cells, many of obscure function, but four types have obvious importance. First, there are two varieties of hair cells. The inner hair cells form a single row of approximately 3,500 cells, whereas approximately 12,000 outer hair cells lie in three rows farther from the central axis of the cochlear spiral (Figure 30–4). The outer hair cells are supported at their bases by the Deiters's (phalangeal) cells; the space between the inner and outer hair cells is delimited and mechanically supported by pillar cells.

A second epithelial ridge adjacent to the organ of Corti, but nearer the cochlea's central axis, gives rise to the tectorial membrane, a cantilevered gelatinous shelf that covers the organ of Corti (Figure 30–4). The tectorial membrane is anchored at its base by the interdental cells, which are also partly responsible for secreting this extracellular structure. The tectorial membrane's tapered distal edge forms a fragile connection with the organ of Corti. When the basilar membrane vibrates in response to a sound, the organ of Corti and the overlying tectorial membrane move with it. Because the basilar and tectorial membranes pivot about different lines of insertion, their up-and-down motion is accompanied by back-and-forth shearing motion of the upper surface of the organ of Corti and the lower surface of the tectorial membrane. This motion is detected by hair cells.

Hair cells originate from the surface ectoderm of the embryo and retain an epithelial character. Columnar or flask-shaped, a hair cell lacks both dendrites and an axon (Figure 30–5A). Around its apex the hair cell is connected to nonsensory supporting cells. A special saline solution, the endolymph, bathes the cell's apical aspect. A tight junction separates this liquid from the ordinary extracellular fluid, or perilymph, that contacts the basolateral surface of the cell. Immediately below the tight junction an intermediate junction provides a strong mechanical attachment for the hair cell.

The hair bundle, which serves as a receptor for mechanical stimuli, projects from the flattened apical surface of the hair cell. Cochlear hair bundles comprise a few hundred cylindrical processes, the stereocilia, arranged in a hexagonal array and extending several micrometers from the cell surface. Because successive stereocilia across a cell's surface vary in length, a hair bundle is beveled like the tip of a hypodermic needle (Figure 30–5B). In the mammalian cochlea the inner hair cell bundles have a roughly linear cross-sectional form. Outer hair cell bundles, in contrast, have a V shape (Figure 30–6).

Each stereocilium is a rigid cylinder whose cytoskeleton consists of a fascicle of actin filaments cross-linked by the proteins plastin (fimbrin), fascin, and espin. Cross-linking renders a stereocilium much more rigid than would be expected for a bundle of unconnected actin filaments. The core of the stereocilium is covered by a tubular sheath of plasma membrane. Although a stereocilium is of constant diameter along most of its length, it tapers over the micrometer or so just above its basal insertion. As the stereocilium narrows from approximately 0.4 μm to one-quarter that diameter, the number of actin filaments diminishes from several hundred to only a few dozen. This thin cluster of microfilaments anchors the stereocilium in the cuticular plate, a thick mesh of interlinked actin filaments lying beneath the cell membrane. Because of this tapered structure, a mechanical force applied at the stereocilium's tip causes the process to pivot around its basal insertion. Horizontal top connectors, extracellular filaments that interconnect adjacent stereocilia along each of the hair bundle's hexagonal axes, cause the bundle to move as a unit during stimulation at low frequencies. At high frequencies, the viscosity of the liquid between the stereocilia opposes their separation and again results in unitary motion of the hair bundle.

During its development every hair bundle includes at its tall edge a single true cilium, the kinocilium. This structure possesses at its core an axoneme, or array of nine paired microtubules, and sometimes an additional central pair of microtubules. The kinocilium is not essential for mechanoelectrical transduction, for in mammalian cochlear hair cells it degenerates around the time of birth.

Deflection of the Hair Bundle Initiates Mechanoelectrical Transduction

Just as in the human vestibular organs (see Chapter 40), mechanical deflection of the hair bundle is the stimulus that excites hair cells of the cochlea. A mechanical stimulus to a hair bundle elicits an electrical response, the receptor potential, by gating mechanically sensitive ion channels. In vitro, when a bundle is deflected by a probe attached near its top, the hair cell's response depends on the direction and magnitude of the stimulus.

In an unstimulated cell approximately 10% of the channels involved in stimulus transduction are open. As a result, the cell's resting potential, approximately –60 mV, is determined in part by the influx of cations through these channels. A stimulus that displaces the bundle toward its tall edge opens additional channels, thereby depolarizing the cell (Figure 30–7). In contrast, a stimulus that displaces the bundle toward its short edge shuts those transduction channels that are open at rest, thus hyperpolarizing the cell.

Hair cells respond only to stimuli parallel to the hair bundle's axis of morphological symmetry: Stimuli at right angles to the axis produce no change from the resting potential. An oblique stimulus elicits a response proportional to its vectorial projection along the axis of sensitivity.

A hair cell's receptor potential is graded. As the stimulus amplitude increases, the receptor potential grows progressively larger up to the point of saturation. The receptor potential of an inner hair cell can be as great as 25 mV in peak-to-peak magnitude. The relation between a bundle's deflection and the resulting electrical response is sigmoidal (Figure 30–7D). A displacement of only ±100 nm represents approximately 90% of the response range. During normal stimulation a hair bundle moves through an angle of ±1 degree or so, that is, by much less than the diameter of one stereocilium.

Hair cells are so sensitive that their response threshold is probably set by brownian motion; still weaker stimuli are lost in the thermal clatter of the ear's components. When observed in vitro a hair bundle exhibits brownian motion of approximately ±3 nm. However, because the auditory system averages responses over several cycles to improve its signal-to-noise ratio, the threshold of hearing corresponds to hair-bundle deflection of as little as ±0.3 nm. A stimulus of this magnitude evokes a receptor potential near 100 μV in amplitude.

The ion channels in hair cells that are involved in stimulus transduction are relatively nonselective, cation-passing pores with a conductance near 100 pS. Because small organic cations can support measurable current, and small fluorescent molecules can traverse the channel, the transduction channel's pore is about 1.3 nm in diameter. Most of the transduction current is carried by K⁺, the cation with the highest concentration in the endolymph bathing the hair bundle.

The large diameter and poor selectivity of the pore permit transduction channels to be blocked by aminoglycoside antibiotics, such as streptomycin, gentamicin, and tobramycin. When used in large doses to counter bacterial infections, these drugs have a toxic effect on hair cells; the antibiotics damage hair bundles and eventually kill hair cells. These drugs evidently insinuate themselves through transduction channels at a low rate and thus cause long-term toxic effects by interfering with protein synthesis on the mitochondrial ribosomes, which resemble prokaryotic ribosomes. Consistent with this hypothesis, human sensitivity to aminoglycosides is maternally inherited and in many instances reflects a single base change in the 12S ribosomal RNA gene of the mitochondrion.

Single-channel recordings and noise analysis suggest that each hair cell possesses only a few hundred transduction channels. Because the number of channels is about the same as the number of stereocilia in a hair bundle and because the magnitude of the receptor potential is roughly proportional to the number of stereocilia remaining in a microdissected bundle, there are probably only two active transduction channels per stereocilium. The paucity of channels along with the lack of high-affinity ligands with which to label them explains why the biochemical identity of the transduction channels remains uncertain. Genetic and physiological evidence suggests, however, that proteins of the transmembrane channel family, specifically TMC1 and TMC2, are involved in mechanoelectrical transduction.

Mechanical Force Directly Opens Transduction Channels

The mechanism for gating of transduction channels in hair cells differs fundamentally from the mechanisms used for such electrical signals as the action potential or postsynaptic potential. Many ion channels respond to changes in membrane potential or to specific ligands (see Chapters 5, 7, 9–11). In contrast, the mechanoelectrical transduction channels in the hair cell are activated by mechanical strain.

Two lines of evidence suggest that the opening and closing of transduction channels is regulated by tension in elastic structures within the hair bundle. First, a bundle is stiffer along its axis of morphological symmetry, and hence of mechanical sensitivity, than at a right angle. This observation suggests that a portion of the work done in deflecting a bundle goes into elastic elements, termed gating springs, that pull on the molecular gates of the transduction channels. Because the gating springs contribute over half of a hair bundle's stiffness, the transduction channels efficiently capture the energy supplied when a bundle is deflected. In addition, hair-bundle stiffness decreases during channel gating, a phenomenon expected if the channels are gated directly through a mechanical linkage to the hair bundle.

A second indication that transduction channels are directly controlled by gating springs is the rapidity with which hair cells respond. The response latency is so brief, only a few microseconds, that gating is more likely to be direct than to involve a second messenger (see Chapter 11). Moreover, the electrical responses of hair cells to a series of step stimuli of increasing magnitude become successively larger and faster. This behavior favors a kinetic scheme in which mechanical force controls the rate constant for channel gating. If the mechanical energy from a stimulus is stored in a spring attached to the channel's gate, the rates of channel opening and closing are determined by the probability that the stored energy of the spring exceeds the transition-state energy for channel gating.

Mechanoelectrical transduction occurs near the tips of the stereocilia, as demonstrated by three experimental techniques. First, the region where cations flow into a hair cell was inferred by measuring small differences in the extracellular potential around a stimulated hair bundle. The voltage signal is strongest at the bundle's top; cations flowing toward transduction channels converge near the stereociliary tips. Second, aminoglycoside antibiotics, which block these channels, have their greatest effect when applied from a microelectrode directed at the top of the hair bundle. Finally, Ca²⁺-sensitive fluorescent indicators initially signal Ca²⁺ entry near the apex of a deflected hair bundle. Measurements from cochlear hair cells at high temporal resolution suggest that the channels occur precisely at the stereociliary tips. When transduction current enters the channels, the ensuing change in membrane potential must propagate axially down the stereocilia before it changes the potential at the base of the cell body and thus influences the rate of transmitter release. Because the stereocilia are short, however, their cable properties do not attenuate electrical signals significantly.

The tip link is a probable component of the gating spring. A tip link is a fine molecular braid joining the distal end of one stereocilium to the side of the longest adjacent process (Figure 30–8A). The upper two-thirds of the link consists of a parallel homodimer of cadherin-23 molecules; the lower third represents a parallel homodimer of protocadherin-15 chains. The two components are joined at their tips in a Ca²⁺-sensitive manner. It is thought that each link is attached, probably at its lower end, to the molecular gates of two transduction channels. Deflection of a hair bundle toward its tall edge tenses the tip link and promotes channel opening; movement in the opposite direction slackens the link and allows the associated channels to close (Figure 30–8B).

Three experimental results suggest that the tip links are components of the gating springs. First, tip links are universal features of hair bundles and are situated at the site of transduction inferred from biophysical experiments. Second, the orientation of the links is consistent with the vectorial sensitivity of transduction. The links invariably interconnect stereocilia in a direction parallel with the hair bundle's plane of mirror symmetry. Stimulation at a right angle to the bundle's plane of symmetry, which would not be expected to alter the length of the links, elicits little or no response from a hair cell. Finally, when tip links are destroyed by exposing hair cells to Ca²⁺ chelators, transduction vanishes. As the tip links regenerate over the course of approximately 12 hours, a hair cell regains mechanosensitivity. It remains unclear whether the elasticity of gating springs resides primarily in the tip links or in the structures at their two insertions.

In the mammalian cochlea hair bundles are deflected through their linkage to the tectorial membrane. As the basilar membrane oscillates up and down, carrying the hair cells with it, a shearing motion occurs between the organ of Corti and the overlying tectorial membrane (Figure 30–9). The hair bundles of outer hair cells, whose tips are firmly attached to the tectorial membrane, are directly deflected by this movement. The hair bundles of inner hair cells, which do not contact the tectorial membrane, are deflected by movement of the liquid beneath the membrane. This mode of stimulation affords some mechanical amplification of the signals reaching hair bundles. At least for high-frequency stimuli, the movements of hair bundles are thought to be severalfold greater than that of the basilar membrane.

Direct Mechanoelectrical Transduction Is Rapid

In contrast to hair cells, many other sensory receptors, such as photoreceptors and olfactory neurons, use cyclic nucleotides or other second messengers in stimulus transduction. This strategy is advantageous in that the enzymatic pathway that generates a second messenger amplifies the signal, and feedback within the metabolic pathway readily permits adaptation and desensitization (see Chapter 11).

What is the advantage of transduction without the intervention of a second messenger? The answer probably lies in its speed. Hair cells operate much more quickly than do other sensory receptor cells of the vertebrate nervous system, and indeed more quickly than neurons themselves. To deal with the frequencies of biologically relevant sounds, transduction by hair cells must be rapid. Given the behavior of sound in air and the dimensions of sound-emitting and sound-absorbing organs such as vocal cords and eardrums, optimal auditory communication occurs in the frequency range of 10 Hz to 100 kHz. Much higher frequencies propagate poorly through air; much lower frequencies are inefficiently produced and captured by animals of moderate size.

Locating sound sources, one of the most important functions of hearing, sets even more stringent limits on the speed of transduction. A sound from a source directly to one side of a person reaches the nearer ear somewhat sooner than the farther. Although this delay is at most 700 μs, a human observer can locate sound sources on the basis of much smaller delays, about 10 μs. For this to occur, hair cells must be capable of detecting acoustic waveforms with microsecond-level resolution.

The ability of hair cells to discriminate high frequencies of stimulation implies that transduction channels are gated very rapidly. Even in animals sensitive to relatively low frequencies, the response to a stimulus of moderate intensity has a time constant of only 80 μs at room temperature. For mammals to be able to respond to frequencies greater than 100 kHz, the hair cells evidently display gating rates that are an order of magnitude greater.

The mechanical sensitivity of hair cells is not constant; responsiveness varies in such a way that a given cell best detects behaviorally relevant stimuli. When it is appropriate that low-frequency inputs be disregarded, hair cells possess a unique mechanism of adaptation that acts as a high-pass filter. In addition, many hair cells in auditory systems display electrical resonance that tunes them to specific frequencies of stimulation. Finally, hair cells employ mechanical amplification that enhances and further tunes their mechanosensitivity.

Hair Cells Adapt to Sustained Stimulation

Despite the precision with which a hair bundle grows, it cannot develop in such a way that the sensitive transduction apparatus is perfectly poised at its position of greatest mechanosensitivity. Some mechanism must compensate for developmental irregularities, as well as for environmental changes, by adjusting the gating springs so that transduction channels are active at the bundle's resting position. An adaptation process that continuously resets the hair bundle's range of mechanical sensitivity does just that. Because of adaptation, a hair cell can maintain a high sensitivity to transient stimuli while rejecting static inputs a million times as large.

Adaptation manifests itself as a progressive decrease in the receptor potential during protracted deflection of the hair bundle (Figure 30–10). The process is not one of desensitization, for the sensitivity of the receptor persists. However, with prolonged stimulation the sensitivity shifts from that of the bundle's resting point to approximately 80% of its deflected position. Adaptation occurs on a time scale three orders of magnitude slower than mechanoelectrical transduction: The time constant of adaptation is approximately 20 ms when endolymph bathes the hair bundle. The rate and extent of adaptation increase with increasing concentration of Ca²⁺ in the liquid contacting the apical cell surface.

How does adaptation occur? Because the mechanical force exerted by a hair bundle changes as adaptation proceeds, the process evidently involves an adjustment in the tension borne by the gating springs. It appears likely that the structure anchoring the upper end of each tip link, the insertional plaque, is repositioned during adaptation by an active molecular motor. Hair bundles contain at least five isoforms of myosin, the motor molecule associated with motility along actin filaments (see Chapter 34). Immunohistochemical studies indicate that myosin-1c occurs in clusters at insertional plaques and near the stereociliary tips, and site-directed mutagenesis implicates this isozyme in adaptation. Several dozen such myosin molecules associated with each tip link are thought to maintain tension by ascending cytoskeletal actin filaments and pulling the link's insertion with them (Figure 30–11).

When a stimulus step increases the tension in a gating spring, the associated transduction channel opens, permitting an influx of cations. As Ca²⁺ ions accumulate in the stereociliary cytoplasm, they bind to the calmodulin light chains that adorn the neck region of myosin-1c. The activation of calmodulin in turn reduces the upward force of the myosin-1c molecule, thereby shortening the gating spring. When the spring reaches its resting tension, closure of the channel reduces the Ca²⁺ influx to its original level, restoring a balance between the upward force of myosin and the downward tension in the spring.

Hair Cells Are Tuned to Specific Stimulus Frequencies

As a result of the tonotopic arrangement of the mammalian basilar membrane, every cochlear hair cell is most sensitive to stimulation at a specific frequency, termed its characteristic, natural, or best frequency. On average, the characteristic frequencies of adjacent inner hair cells differ by approximately 0.2%; adjacent piano strings, in comparison, are tuned to frequencies some 6% apart.

The sensitivity of a cochlear hair cell extends within a limited range above and below its characteristic frequency. This follows from the fact that the traveling wave evoked even by a pure sinusoidal stimulus spreads somewhat along the basilar membrane. The traveling wave of a stimulus tone with a pitch lower than the characteristic frequency of a particular hair cell passes that cell and peaks somewhat farther up the cochlear spiral. A higher-pitched tone causes a traveling wave that crests below the cell. Nevertheless, in either instance the basilar membrane undergoes some motion at the hair cell's site, so that the cell responds to the stimulus.

The frequency sensitivity of a hair cell may be displayed as a tuning curve. To construct a tuning curve, an experimenter stimulates the ear with pure tones at numerous frequencies below, at, and above the cell's characteristic frequency. The intensity of stimulation is adjusted for each frequency until the response reaches a predefined criterion magnitude. An investigator might, for example, ask what stimulus intensity is necessary at each frequency to produce a receptor potential 1 mV in peak-to-peak magnitude. The tuning curve is then a graph of sound level, presented logarithmically in decibels sound-pressure level, against stimulus frequency.

The tuning curve for an inner hair cell is typically V-shaped (Figure 30–12). The curve's tip represents the cell's characteristic frequency, the frequency that produces the criterion response for the lowest- intensity stimulus. Sounds of greater or lesser frequencies require higher intensities to excite the cell to the criterion response. As a consequence of the traveling wave's shape, the slope of a tuning curve is far steeper on its high-frequency flank than on its low-frequency flank.

Hair cells must contend with acoustic stimuli that have a very low energy content. If the stimulus consists of a periodic signal, such as the sinusoidal pressure of a pure tone, a detection system can increase the signal-to-noise ratio by enhancing selectively the response to a relevant frequency. At least two cellular mechanisms are known to accomplish this task, thereby supplementing the tuning accomplished by the basilar membrane.

First, the mechanical properties of a hair bundle help tune it to a particular frequency, in the same way a tuning fork's resonant frequency depends on the mechanical properties of its tines. These properties include the bundle's flexibility and mass. The flexible elements that restore a bundle to its resting upright position are the gating springs and the actin-filled rootlets at the base of the stereocilia. Because the bundle moves through a viscous medium, the mass relevant to the bundle's tuning includes that of a volume of water dragged along by the moving bundle. Viscosity also heavily dampens the motion.

In the auditory organs of many animals the lengths of the hair bundles vary systematically along the tonotopic axis. Hair cells that respond to low-frequency acoustic stimuli have the longest bundles, whereas those that respond to the highest-frequency signals bear the shortest bundles. In the human cochlea, for example, an inner hair cell with a characteristic frequency of 20 kHz bears a 4 μm hair bundle. At the opposite extreme, a cell sensitive to a 20 Hz stimulus has a bundle more than 7 μm high.

Computer modeling shows that the tuning-fork mechanism helps to tune cochlear hair cells with freestanding hair bundles, those that are not attached to a tectorial membrane. In humans these cells are the inner hair cells, the receptors that provide most of the information conveyed by the cochlear nerve. The length of the stereocilia may also affect the tuning of cells whose hair bundles are inserted into a tectorial membrane, for in these hair cells too there is an inverse relation between bundle length and characteristic frequency.

The second mechanism that tunes individual hair cells to specific frequencies is electrical in nature. In many fishes, amphibians, and reptiles including birds the membrane potential of each hair cell resonates at a particular frequency in response to an injected current pulse (Figure 30–13A). When a cell is stimulated with sounds of various frequencies but constant amplitude, it responds over a broad range of frequencies. However, when current is injected the cell responds most strongly to stimulation at the particular frequency at which the cell's membrane potential resonates. Whether electrical resonance contributes to frequency tuning in the ears of mammals, including humans, remains uncertain.

The basis of electrical resonance has been determined by voltage-clamp recordings from isolated hair cells. The depolarizing phase of an oscillation is driven by current carried into the cell through voltage-gated Ca²⁺ channels, whereas the repolarizing component results primarily from outward current through Ca²⁺-sensitive K⁺ channels (Figure 30–13B). Several factors establish the frequency and sharpness of the resonance, including the membrane capacitance, the numbers of Ca²⁺ and K⁺ channels, and the time course of Ca²⁺ removal. In addition, variation in the K⁺ channels along the cochlea is a factor in the differences in frequency selectivity of the hair cells. Alternative splicing of the mRNA encoding cochlear K⁺ channels generates several channel isoforms that differ in their kinetics and their sensitivities to Ca²⁺ and voltage. Moreover, expression of the channel's auxiliary β subunit, which also regulates gating kinetics, displays a gradient along the tonotopic axis. How hair cells become tuned to their characteristic frequencies during development remains to be determined.

Sound Energy Is Mechanically Amplified in the Cochlea

The inner ear faces an important obstacle to efficient operation: A large portion of the energy in an acoustic stimulus goes into overcoming the damping effects of cochlear liquids on basilar-membrane motion rather than into excitation of hair cells. The sensitivity of the cochlea is too great, and auditory frequency selectivity too sharp, to result solely from the inner ear's passive mechanical properties. The cochlea must therefore possess some means of actively amplifying sound energy.

One indication that amplification occurs in the cochlea comes from measurements of the basilar membrane's movements with sensitive laser interferometers. In a preparation stimulated with low-intensity sound the motion of the membrane at any point is highly sensitive to frequency. As the sound intensity is increased, however, the membrane's sensitivity declines precipitously and its tuning becomes less sharp: The sensitivity of basilar-membrane motion to stimulation at 80 dB is less than 1% that for 10 dB excitation. The sensitivity predicted in modeling studies of a passive cochlea corresponds to that observed with high-intensity stimuli. This result implies that the motion of the basilar membrane is augmented more than 100-fold during low-intensity stimulation but that amplification diminishes progressively as the stimulus grows in strength.

In addition to this circumstantial evidence, experimental observations support the idea that the cochlea contains a mechanical amplifier. When a normal human ear is stimulated with a click, that ear emits one to several measurable pulses of sound. Each pulse includes sound in a restricted frequency band. High-frequency sounds are emitted with the shortest latency, approximately 5 ms, whereas low-frequency emissions occur after a delay as great as 20 ms (Figure 30–14A). These so-called evoked otoacoustic emissions are not simply echoes; they represent the emission of mechanical energy by the cochlea, triggered by acoustic stimulation.

A still more compelling manifestation of the cochlea's active amplification is spontaneous otoacoustic emission. When a suitably sensitive microphone is used to measure sound pressure in the ear canals of subjects in a quiet environment, at least 70% of normal human ears continuously emit one or more pure tones (Figure 30–14B). Although these sounds are generally too faint to be directly audible by others, physicians have reported hearing sounds emanating from the ears of newborns! The ears of adults, too, occasionally emit audible sounds. The active process in the cochlea ordinarily serves to counter the viscous damping effects of cochlear fluids on the basilar membrane. However, if this cochlear amplifier is overly active the ear emits sound, just as a public-address system howls when its gain is excessive.

What is the source of evoked and spontaneous otoacoustic emissions, and presumably of cochlear amplification as well? Several lines of evidence implicate outer hair cells as the elements that enhance cochlear sensitivity and frequency selectivity and hence as the energy sources for amplification. The afferent nerve fibers that extensively innervate the inner hair cells make only minimal contacts with the outer hair cells. Instead, the outer hair cells receive an extensive efferent innervation that, when activated, decreases cochlear sensitivity and frequency discrimination. Pharmacological ablation of outer hair cells with selectively ototoxic drugs degrades the ear's responsiveness still more profoundly.

When stimulated electrically, an isolated outer hair cell displays the unique phenomenon of electromotility: The cell body shortens when depolarized and elongates when hyperpolarized (Figure 30–15). This response can occur at frequencies exceeding 80 kHz, an attractive feature for a process postulated to assist high-frequency hearing.

The energy for these movements is drawn from the experimentally imposed electrical field rather than from hydrolysis of an energy-rich substrate such as adenosine triphosphate (ATP). Movement occurs when changes in the electric field across the membrane reorient molecules of the protein prestin. The concerted movement of several million of these molecules changes the membrane's area and thus the cell's length. When an outer hair cell transduces mechanical stimulation of its hair bundle into receptor potentials, cochlear amplification might then occur as voltage-induced movement of the cell body augments basilar-membrane motion. Consistent with this hypothesis, mutation of certain amino-acid residues required for the voltage sensitivity of prestin abolishes the active process in mice.

Because sharp tuning, high sensitivity, and otoacoustic emissions are also observed in animals that lack outer hair cells, electromotility cannot be the only form of mechanical amplification by hair cells. In addition to detecting stimuli, hair bundles are also mechanically active and contribute to amplification. Hair bundles can make spontaneous back-and-forth movements that might underlie spontaneous otoacoustic emissions. Under experimental conditions bundles can exert force against stimulus probes, performing mechanical work and thereby amplifying the input. In vitro experiments indicate that active hair-bundle motility contributes to the cochlear active process even in the mammalian ear.

Several features of auditory responsiveness suggest that cochlear hair cells operate on the verge of an instability termed the Hopf bifurcation. This phenomenon explains hair cells' amplification and frequency selectivity, their nonlinear sensitivity to stimulus intensity, and their capacity to become unstable and spontaneously emit sound. The fact that active hair-bundle motility demonstrates a Hopf bifurcation in vitro provides further evidence that this mechanism contributes to cochlear amplification.

Although active hair-bundle movements have been demonstrated at sound frequencies as high as a few kilohertz, it remains uncertain whether bundles can generate forces at the very high frequencies at which sharp frequency selectivity and otoacoustic emissions are observed in the mammalian cochlea. Active hair-bundle motility and electromotility may function synergistically, with the former serving metaphorically as a tuner and preamplifier and the latter as a power amplifier. Alternatively, hair-bundle motility may operate at relatively low frequencies but be superseded by electromotility at the highest frequencies.

In addition to being sensory receptors, hair cells also form synapses with sensory neurons. The basolateral membrane of each cell contains several presynaptic active zones at which chemical neurotransmitter is released. An active zone is characterized by four prominent morphological features (Figure 30–16).

A presynaptic dense body or synaptic ribbon lies in the cytoplasm adjacent to the release site. This fibrillar, osmiophilic structure may be spherical, ovoidal, or flattened, and usually measures a few hundred nanometers across. The dense body resembles the synaptic ribbon of a photoreceptor cell and represents a specialized elaboration of the smaller presynaptic densities found at the neuromuscular junction and at central nervous system synapses. In addition to molecular components shared with conventional synapses, ribbon synapses contain large amounts of the protein ribeye.

The presynaptic ribbon is surrounded by clear synaptic vesicles, each 35 nm to 40 nm in diameter, which are attached to the dense body by tenuous filaments. Between the dense body and the presynaptic plasma membrane lies a striking presynaptic density that comprises several short rows of fuzzy material. Within the plasmalemma rows of large particles are aligned with the strips of presynaptic density. These particles are thought to include the Ca²⁺ channels involved in the release of transmitter as well as the K⁺ channels that participate in electrical resonance.

Studies of nonmammalian experimental models show that, as with most other synapses (see Chapter 12), the release of transmitter by hair cells is evoked by presynaptic depolarization and requires Ca²⁺ from the extracellular medium. Hair cells lack synaptotagmins 1 and 2, however, and the role of those proteins as rapid Ca²⁺ sensors has probably been assumed by the protein otoferlin, which also promotes the replenishment of synaptic vesicles. Postsynaptic recordings indicate that the release of the hair cell's synaptic transmitter is quantal in nature, resembling that of the neuromuscular junction. Although glutamate is the principal afferent neurotransmitter, other substances are released as well.

The presynaptic apparatus of hair cells has several unusual features that underlie the signaling abilities of these cells. Hair cells release synaptic transmitter continuously at rest. The rate of transmitter release can then be modulated upward or downward, depending on whether the hair cell is respectively depolarized or hyperpolarized. Consistent with this observation, the Ca²⁺ channels of hair cells are activated at the resting potential, providing a steady leak of Ca²⁺ that evokes transmitter release from unstimulated cells. Another unusual feature of the hair cell's synapses is that, like those of photoreceptors, they must be able to release neurotransmitter reliably in response to a threshold receptor potential of only 100 μV or so. This feature, too, results from the fact that the presynaptic Ca²⁺ channels are activated at the resting potential.

Most hair cells receive inputs from neurons in the brain stem at large boutons on the basolateral cell surface. These efferent terminals contain numerous clear synaptic vesicles about 50 nm in diameter as well as a smaller number of larger, dense-core vesicles. The principal transmitter at these synapses is acetylcholine (ACh); calcitonin gene–related peptide (CGRP) also occurs in efferent terminals and may be co-released with ACh. ACh binds nicotinic ionotropic receptors consisting of α9 and α10 subunits. These receptor-channels have a substantial permeability to Ca²⁺ as well as Na⁺ and K⁺. The Ca²⁺ that enters through these channels activates small-conductance Ca²⁺-sensitive K⁺ channels (SK channels), whose opening leads to a protracted hyperpolarization. The cytoplasm of a hair cell immediately beneath each efferent terminal holds a single cisterna of smooth endoplasmic reticulum. This structure may be involved in the reuptake of the Ca²⁺ that enters in response to efferent stimulation.

The best-understood role of efferent input in the cochlea is its effect on hair cells that employ electrical resonance for frequency tuning. Stimulation of the efferent nerve fibers hyperpolarizes a hair cell. The associated increase in membrane conductance perturbs the critically tuned resonance circuit in the cell's membrane, thus decreasing both the sharpness of frequency selectivity and the gain of electrical amplification. In the mammalian cochlea, where efferent fibers contact outer hair cells, the efferent system desensitizes the cochlea by turning down the active process.

Bipolar Neurons in the Spiral Ganglion Innervate Cochlear Hair Cells

Information flows from cochlear hair cells to primary sensory neurons whose cell bodies lie in the cochlear ganglion. The central processes of these bipolar neurons form the cochlear division of the vestibulocochlear nerve (cranial nerve VIII). Because this ganglion follows a spiral course within the bony core of the cochlea, it is also called the spiral ganglion. Approximately 30,000 ganglion cells innervate the hair cells of each inner ear.

The afferent pathways from the human cochlea reflect the functional distinction between inner and outer hair cells. At least 90% of the spiral ganglion cells terminate on inner hair cells (Figure 30–17). Each axon contacts only a single inner hair cell, but each cell directs its output to several nerve fibers, on average nearly 10. This arrangement has three important consequences.

First, the neural information from which hearing arises originates almost entirely at inner hair cells. Second, because the output of each inner hair cell is sampled by many afferent nerve fibers, the information from one receptor is encoded independently in several parallel channels. Third, at any point along the cochlear spiral, or at any position within the spiral ganglion, each ganglion cell responds best to stimulation at the characteristic frequency of the presynaptic hair cell. The tonotopic organization of the auditory neural pathways thus begins at the earliest possible site, immediately postsynaptic to inner hair cells.

Relatively few cochlear ganglion cells contact outer hair cells, and each such neuron extends branching terminals to numerous outer hair cells. Although the ganglion cells that receive input from outer hair cells are known to project into the central nervous system, these neurons are so few that it is not certain whether their projections contribute significantly to the analysis of sound.

The patterns of efferent and afferent connections of cochlear hair cells are complementary. Mature inner hair cells do not receive efferent input; just beneath these cells, however, are extensive axo-axonic synaptic contacts between efferent axon terminals and the endings of afferent nerve fibers. In contrast, efferent nerves have extensive connections with outer hair cells on their basolateral surfaces. Each outer hair cell receives input from several large efferent terminals, which fill the space between the cell's base and the associated Deiters's cell.

Cochlear Nerve Fibers Encode Stimulus Frequency and Intensity

The acoustic sensitivity of axons in the cochlear nerve mirrors the connection pattern of the spiral ganglion cells. Each axon is most responsive to a characteristic frequency. Stimuli of lower or higher frequency also evoke responses but only when presented at greater intensities. An axon's responsiveness may be characterized by a tuning curve, which is V-shaped like the curves for basilar-membrane motion and hair-cell sensitivity (Figure 30–12). The tuning curves for nerve fibers with different characteristic frequencies resemble one another but are shifted along the abscissa.

The relation between sound-pressure level and firing rate in each fiber of the cochlear nerve is approximately linear. Because of the dependence of level on sound pressure, this relation implies that sound pressure is logarithmically encoded by neuronal activity. Very loud sounds saturate a neuron's response. Because an action potential and the subsequent refractory period last almost 1 ms, the greatest sustainable firing rate is somewhat above 500 spikes per second.

Even among nerve fibers with the same characteristic frequency, the threshold of responsiveness varies from axon to axon. The most sensitive fibers, whose response thresholds extend down to approximately 0 dB, characteristically have high rates of spontaneous activity and produce saturating responses for stimulation at moderate intensities, approximately 40 dB. At the opposite extreme, the least sensitive afferent fibers have less spontaneous activity and much higher thresholds but respond in graded fashion to intensities in excess of 100 dB. The activity patterns of most fibers range between these extremes.

The afferent neurons of lowest sensitivity contact the surface of an inner hair cell nearest the axis of the cochlear spiral. The most sensitive afferent neurons contact the hair cell's opposite side. The multiple innervation of each inner hair cell is therefore not wholly redundant. Instead, the output from a given hair cell is directed into several parallel channels of differing sensitivity and dynamic range.

The firing pattern of fibers in the eighth cranial nerve has both phasic and tonic components. Brisk firing occurs at the onset of a tone presented for a few seconds. As adaptation occurs, however, the firing rate declines to a plateau level over a few tens of milliseconds. When stimulation ceases there is usually a transitory cessation of activity with a similar time course before resumption of the spontaneous firing rate (Figure 30–18).

When a periodic stimulus such as a pure tone is presented, the firing pattern of a cochlear nerve fiber encodes information about the periodicity of the stimulus. For example, a relatively low-frequency tone at a moderate intensity might produce one spike in a nerve fiber during each cycle of stimulation. The phase of firing is also stereotyped. Each action potential might occur, for example, during the compressive phase of the stimulus. As the stimulation frequency rises, the stimuli eventually become so rapid that the nerve fiber can no longer produce action potentials on a cycle-by-cycle basis. Up to a frequency in excess of 4 kHz, however, phase-locking persists; a fiber may produce an action potential only every few cycles of the stimulus, but its firing continues to occur at a particular phase in the stimulus cycle.

Periodicity in neuronal firing enhances the information about the stimulus frequency. Any pure tone of sufficient intensity evokes firing in numerous cochlear nerve fibers. Those fibers whose characteristic frequency coincides with the frequency of the stimulus respond at the lowest stimulus level, but respond still more briskly for stimuli of moderate intensity. Other nerve fibers with characteristic frequencies close to the stimulus also respond, although somewhat less vigorously. Regardless of their characteristic frequencies, however, all the responsive fibers display phase locking: Each tends to fire during a particular part of the stimulus cycle.

The central nervous system can therefore gain information about stimulus frequency in two ways. First, there is a place code; the fibers are arrayed in a tonotopic map in which the position is related to characteristic frequency. Second, there is a frequency code; the fibers fire at a rate that signals the frequency of the stimulus. Frequency coding is of particular importance when a sound is loud enough to saturate the neuronal firing rate. Although fibers with different characteristic frequencies respond to such a stimulus, each provides information about the stimulus frequency in its firing pattern.

Whether mild or profound, most deafness falls into the category of sensorineural hearing loss, often misnamed "nerve deafness." Although hearing loss can result from damage to the eighth cranial nerve, for example from an acoustic neuroma (see Chapter 45), deafness stems primarily from the loss of cochlear hair cells.

The 16,000 hair cells in each human cochlea are not replaced by cell division but must last a lifetime. However, in amphibians and birds supporting cells can be experimentally induced to divide and their progeny to produce new hair cells. In the zebrafish some hair cell populations are regenerated continually by the activity of stem cells. Researchers have recently succeeded in replenishing mammalian hair cells in vitro. Until we understand how hair cells can be restored to the organ of Corti, however, we must cope with hearing loss, which is becoming more prevalent because of our aging population and increasingly noisy environment.

The last few decades have brought remarkable advances in our ability to treat deafness. For the majority of patients who have significant residual hearing, hearing aids can amplify sounds to a level sufficient to activate the surviving hair cells. A modern aid is custom-tailored to compensate for each individual's hearing loss, so that the device most amplifies sounds at frequencies to which the wearer is least sensitive while providing little or no enhancement to those that can still be heard well. To the credit of our society, the stigma formerly associated with wearing a hearing aid is dissipating rapidly; using such an aid is now regarded as unremarkable as wearing eyeglasses.

When most or all of a person's cochlear hair cells have degenerated, no amount of amplification can assist hearing. However, hearing can be restored by bypassing the damaged cochlea with a cochlear prosthesis. A user wears a compact unit that picks up sounds, decomposes them into their frequency components, and forwards electronic signals representing these constituents along separate wires to small antennas situated just behind the auricle. The signals are then transmitted transdermally to receiving antennas implanted in the temporal bone. From there, fine wires bear the signals to appropriate electrodes implanted at various positions along the scala tympani. Activation of the electrodes excites action potentials in nearby axons (Figure 30–19).

The cochlear prosthesis takes advantage of the tonotopic representation of stimulus frequency along the cochlea. Because the axons innervating each segment of the cochlea are concerned with a specific, narrow range of frequencies, each electrode in a prosthesis can excite a cluster of nerve fibers that are sensitive to similar frequencies. The stimulated neurons then forward their outputs along the intact eighth nerve to the central nervous system, where these signals are interpreted as a sound of the frequency represented at that position on the basilar membrane. An array of approximately 20 electrodes can mimic a complex sound by appropriately stimulating several clusters of neurons.

Cochlear prostheses have now been implanted in more than 100,000 patients worldwide. Their effectiveness varies widely from person to person. In the best outcome an individual can understand speech nearly as well as a normally hearing person and can even conduct telephone conversations. At the other extreme are patients who derive little benefit from prostheses, presumably because of extensive degeneration of the nerve fibers near the electrode array. Most patients find their prostheses of great value. Even if hearing is not completely restored, the devices help in lip reading and alert patients to noises in the environment.

The other way to overcome deafness relies not on high technology but on the efforts of generations of deaf individuals and their teachers. Sign languages have probably existed for as long as humans have spoken, and perhaps even longer. Many such languages represent attempts to translate spoken language into a system of hand signs. Signed English, for example, provides an effective means of communication that largely obeys the rules of English speech.

However, sign languages that diverge more radically from spoken languages are far more effective. Freed from the constraint of mirroring English, American Sign Language, or ASL, has become an elegant and eloquent language in its own right. Linguists now recognize American Sign Language as a distinct language whose range of expressiveness generally matches—and sometimes exceeds—that of spoken English.

Hearing, a key sense in human communication, begins with capture of sound by the ear. Mechanical energy captured by the outer ear flows through the middle ear to the cochlea, where it causes the elastic basilar membrane to oscillate. Aligned along the basilar membrane in a tonotopic array, 16,000 hair cells detect the frequency components of a stimulus and transduce them into receptor potentials that cause sensory neurons to fire.

As the population ages and as society becomes more concerned about hearing loss, physicians will increasingly confront patients and families who are experiencing the social difficulties associated with deafness. This subject is at present politically charged. On the one hand, the rapid technical improvement in cochlear prostheses leads their developers to advocate use of the devices whenever practical, including for young children. Many members of the deaf community, on the other hand, believe that widespread implantation of cochlear prostheses, particularly in children, will foster a generation of individuals whose ability to communicate is dependent on technological support of as yet unproved durability. The extensive use of prostheses might also diminish the use of American Sign Language and thus reverse the deaf community's remarkable recent advances.

Although this debate will not soon subside, it is worthwhile to emphasize the most positive aspect of the issue. A few decades ago there were no widely effective ways of coping with profound deafness; now there are two. Moreover, these solutions are not mutually exclusive; with a cochlear prosthesis a deaf individual can benefit from bilingualism in spoken English and American Sign Language.