Physiological basis of cochlear transduction and sensitivity

Electrical impedance measurements of cochlear structures using the four-electrode reflection-coefficient technique

In individuals with severe-to-profound hearing loss, cochlear implants (CIs) bypass normal inner ear function by applying electrical current directly into the cochlea, thereby stimulating surviving auditory nerve fibers. Although cochlear implants are able to restore some auditory sensation, they are far from providing normal hearing. It has been estimated that up to 75% of the current injected via a CI is shunted along scala tympani and is not available to stimulate auditory neurons. The path of the injected current and the consequent population of stimulated spiral ganglion cells are dependent upon the positions of the electrode contacts within the cochlea and the impedances of cochlear structures. However, characterization of the current path remains one of the most critical, yet least understood, aspects of cochlear implantation. In particular, the impedances of cochlear structures, including the modiolus, are either unknown or based upon estimates derived from circuit models. Impedance values for many cochlear structures have never been measured. By combining the hemicochlea preparation, a cochlea cut in half along its mid-modiolar plane, and the four-electrode reflection-coefficient technique, impedances can be measured for cochlear tissues in a cochlear cross section including the modiolus. Advantages and disadvantages of the method are discussed in detail and electrical impedance measurements obtained in the gerbil hemicochlea are presented. The resistivity values for the cochlear wall in Omegacm are, 528 (range: 432-708) for scala media 3rd turn, 502 (range: 421-616) for scala tympani 3rd turn and scala vestibuli 2nd turn, 627 (range: 531-759) for scala media 2nd turn, 434 (range: 353-555) for scala tympani 2nd turn and scala vestibuli basal turn, 434 (range: 373-514) for scala media basal turn, and 590 (range: 546-643) for scala tympani basal turn. The resistivity was 455Omegacm (range: 426-487) for the modiolus.

Tissue resistivities determine the current flow in the cochlea

PURPOSE OF REVIEW

In individuals with severe to profound hearing loss, cochlear implants bypass normal inner ear function by applying electrical current directly into the cochlea, thereby stimulating cochlear nerve fibers. Stimulating discrete populations of spiral ganglion cells in cochlear implant users' ears is similar to the encoding of small acoustic frequency bands in a normal-hearing person's ear. Thus, spiral ganglion cells stimulated by an electrode convey the information contained by a small acoustic frequency band. Problems that refer to the current spread and subsequent nonselective stimulation of spiral ganglion cells in the cochlea are reviewed.

RECENT FINDINGS

Cochlear anatomy and tissue properties determine the current path in the cochlea. Current spreads largely via scala tympani and across turns. While most of the current leaves the cochlea via the modiolus, the facial canal and the round window constitute additional natural escape paths for the current from the cochlea. Moreover, degenerative processes change tissue resistivities and thus may affect current spread in the cochlea.

SUMMARY

Electrode design and coding strategies may result in more spatial stimulation of spiral ganglion cells, resulting in a better performance of the electrode-tissue interface.

Electrical resistivity measurements in the mammalian cochlea after neural degeneration

OBJECTIVES/HYPOTHESIS

In the present series of experiments, the effect of neural degeneration on the cochlear structure electrical resistivities was evaluated to test if it alters the current flow in the cochlea and if increased current levels are needed to stimulate the impaired cochlea. In cochlear implants, frequency information is encoded in part by stimulating discrete populations of spiral ganglion cells along the cochlea. However, electrical properties of the cochlear structures result in shunting of the current away from the auditory neurons. This consumes energy, makes cochlear implants less efficient, and drastically reduces battery life. Models of the electrically stimulated cochlea serve to make predictions on current paths using modified and improved cochlear implant electrodes. However, one of the model's shortcomings is that most of the values for tissue impedances are not direct measurements. They are derived from bulk impedance measurements, which are fitted to lumped-element models.

STUDY DESIGN

The four-electrode reflection-coefficient technique was used to measure resistivities in the gerbil cochlea. In vivo and in vitro (the hemicochlea) models were used. Measurements were made in normal and in deafened animals. Cochlear damage was induced by neomycin injection into the animals' middle ears. Neural degeneration was allowed to occur over 2 months before performing the measurements in the deafened animals.

RESULTS

The resistivity values in deafened animals were smaller than in the normal-hearing animals, thus altering the current flow within the cochlea.

CONCLUSIONS

Resistivity changes and subsequent changes in current path should be considered in future designs of cochlear implants.

Radial current flow and source density in the basal scala tympani

Ionic movement between the scala media and scala tympani is modulated by acoustic stimulation. It underlies electrical currents in the fluids of these compartments and produces voltage gradients from which partial current-flow densities can be estimated. Radial voltage gradients were sampled in the first turn of the scala tympani of the guinea pig cochlea at known distances from the basilar membrane. Large potential gradients indicated significant current-flow densities near the organ of Corti with less flow near the lateral wall of the cochlea and over the spiral lamina. Current-source densities were highest within 100 micron of the organ of Corti. Current-source density analysis suggested that source-sink pairs can be detected from field observations in the scala tympani.

Transepithelial electrical potential of nonsensory region of gerbil utricle in vitro

Transepithelial electrical potential difference (VT) was measured across the vestibular labyrinth of the inner ear in vitro by puncturing the epithelial wall of the utricle with a glass microelectrode. A region of nonsensory cells of the utricle was isolated from the sensory regions by introducing columns of liquid Sylgard 184. Under control conditions, the VT of this region was +7.5 +/- 0.3 mV (means +/- SE), lumen positive. This potential difference was rapidly reduced by either 1 mM ouabain, 10-100 microM bumetanide, 0.5-5.0 mM Ba (in the bathing solution), or cooling, but not by the disulfonic stilbene, 4-acetamido-4'-isothiocyanostilbene-2,2'-disulfonic acid. Changes in VT due to reductions of Cl or Na or to increases of K in the bathing solution in exchange for presumably impermeant ions were observed in this region and were compared with those in a preparation in which the insulating seals were absent. The K-induced voltage change was significantly higher in the unblocked preparation, a finding consistent with a high K permeability of the sensory cells. The voltage change due to reduction of Cl was not inhibited by Cl channel blockers (9-anthracenecarboxylate and diphenylamine-2-carboxylate) in the bathing solution. These results represent the first direct demonstration that the nonsensory cells of the utricle produce a lumen-positive active-transport potential and characterize some of the properties of the cell membranes in terms of their pharmacological sensitivities and net voltage responses to changes in the bathing medium ions Na, K, and Cl.

Differential effect of ischemia on spontaneous and sinusoidal-evoked activity in semicircular canal afferents in the bullfrog

Spontaneous and sinusoidal-evoked nerve activity in semicircular canal afferent fibers of the bullfrog was evaluated prior to and following the production of ischemia of the labyrinthine arterial supply by mechanical occlusion of the vestibular artery. Neuronal spontaneous firing rates were observed to diminish by up to 100% within 10 min following the onset of ischemia. In most neurons there was a substantial increase in firing rate during the first few minutes. The sensitivity of the fibers to natural stimulation as determined by the gain in their responses to sinusoidal motion also diminished by as much as 75% over the same period. No detectable changes in the membrane potentials of the neurons were observed. The changes in excitability were closely correlated with the changes in spontaneous firing rate, but not all the neurons whose responses changed showed changes in spontaneous activity. Likewise, the relative magnitude of change varied from neuron to neuron.

Role of inner and outer hair cells in mechanical frequency selectivity of the cochlea

Resonant frequencies of inner (IHC) and outer (OHC) hair cell systems in the guinea pig cochlea were computed using data on sensory-hair stiffness obtained from in vitro organ of Corti preparations (D. Strelioff and A. Flock (1984): Hearing Res. 15, 19-28). IHC stereocilia were modelled as stiff, free-standing, uniform cylinders which rotate about their elastic attachments to the apical surfaces of IHC. OHC, with the overlying tectorial membrane (TM), were modelled as a resonant mechanical system with the TM providing mass and the three rows of OHC sensory-hair bundles providing elasticity for linear, simple harmonic motion parallel to the reticular lamina. Since computed IHC resonant frequencies increase from 128 kHz at the apex to 300 kHz at the base, it is unlikely that they contribute to frequency selectivity. In contrast, computed frequencies of the OHC-TM system are within the audio range, increasing from 1.2 kHz at the apex to 22 kHz at the base. The results of these computations support the hypothesis that the OHC-TM system contributes to mechanical frequency selectivity of the cochlea whereas IHC are passive receptors which respond to mechanical movements of the cochlear partition.

Responses of gerbil and guinea pig auditory nerve fibers to low-frequency sinusoids

The characteristics of time-locked auditory nerve fiber responses to 50 Hz acoustic sinusoids were studied in gerbils and guinea pigs. Whereas the time-locked responses of all guinea pig fibers produced single-peaked period histograms, those of the gerbil produced distorted, multiple-peaked response histograms, especially fibers with characteristic frequencies (CFs) between 2 and 10 kHz. Although the shapes of the period histograms vary with stimulus intensity, the phases of the fundamental components are essentially invariant over the range of stimulus intensities used. In contrast to the phase of the cochlear microphonic produced by the 50 Hz stimulus, which was constant along the length of the cochlea in both species, the phase of the neural responses depends on the fiber CF in each of the two species. In guinea pigs, the phase of the neural responses relative to the acoustic stimulus decreases with the fiber CF from a phase lead of 90 degrees for fibers with CFs below 300 Hz to a phase lag of nearly 60 degrees for fibers with CFs greater than 3 kHz. In gerbils, the response phase also decreases with increasing CF below 2 kHz and above 10 kHz but undergoes an abrupt 160 degrees phase increase between those frequencies.

Some electrical circuit properties of the organ of Corti. I. Analysis without reactive elements

A simplified network model of the organ of Corti is analyzed with the assumption of parametric excitation via resistance changes in the hair cells' apical membrane. Pertinent network variables (intracellular resting and receptor potentials, cellular input resistance, extracellular potentials) depend on the ratios of basal (perilymphatic face) and apical (endolymphatic face) receptor cell resistances, denoted as shape factors. In the Appendix two methods are suggested for the computation of shape factors; both are based on the geometrical properties of hair cells. Various electrical quantities computed on the basis of shape factors are consistent with recent recordings from third turn inner and outer hair cells (Dallos et al. (1982): Science 218, 582-584). The model provides a plausible explanation for the experimentally observed discrepancy between inner and outer hair cell resting and receptor potentials. One potentially significant result of the analysis is the demonstration that since shape factors for outer hair cells are probably longitudinally graded, so must be all cellular electrical characteristics. Another interesting finding is that electrical interaction among neighboring hair cells is unlikely. A large-signal analysis of the circuit demonstrates that even in the absence of a non-linear input, the parametrically excited circuit itself generates pronounced distortion. The most significant consequence of this nonlinearity is a response asymmetry in which the depolarizing phase is greater than the hyperpolarizing one. Thus the circuit nonlinearity may, at least in part, account for the large positive d.c. response seen in both types of receptor cell (Dallos et al. (1982): Science 218, 582-584; Russell and Sellick (1978): J. Physiol. Lond. 284, 261-290).

Fine structure of the sensory epithelium of guinea-pig organ of Corti: subsurface cisternae and lamellar bodies in the outer hair cells

The fine structure of subsurface cisternae and lamellar bodies in the outer hair cells of the guinea-pig organ of Corti was studied with thin sections and freeze-fracture replicas. Subsurface cisternae in the outer hair cells consist of multilayers along the lateral plasma membrane of the cell. The outermost layer is a flattened cistern in the upper part of the supranuclear region, but comprises a series of tubules in the lower part. Deeper layers are fenestrated cisternae in which disc-like areas are found in the upper part of the supranuclear region. Lamellar bodies consist of concentric layers of fenestrated cisternae and are located in the apical cytoplasm beneath the cuticular plate. They are continuous with the subsurface cisternae. In the supranuclear cytoplasm, bulges of the subsurface cisternae and the lamellar bodies are found. Dilated cisternae are also present. Some dilated cisternae contain many small vesicles, which display acid phosphatase activity. The dilated cisternae are considered as forms of the bulges undergoing transformation into multivesicular bodies. The possible role of the lamellar bodies, and the origin and fate of the subsurface cisternae are discussed.

Intracellular recordings from cochlear outer hair cells

Intracellular recordings were made from outer hair cells in the third turn of the guinea pig cochlea, and the electrical characteristics of the cells were compared to those of inner hair cells, supporting cells, and extracellular spaces from the same recording region. Outer hair cells have higher membrane potentials than do inner hair cells, but they produce smaller a-c receptor potentials. The frequency response characteristics of both types of hair cells are probably not significantly different. In the frequency region where tuning is optimal, both cell types produce depolarizing d-c receptor potentials, but outer hair cells also generate hyperpolarizing responses at low frequencies.

Longitudinal distribution of cochlear potentials and the K+ concentration in the endolymph after acoustic trauma

Guinea pigs were exposed to 142 dB third-octave band of noise control at 1 kHz for 1 h. At different times after exposure the endocochlear potential (EP), the anoxic negative endocochlear potential (-EP), the concentration of K+ (K+e) and microphonic potentials were recorded in scala media in four cochlear turns. The remaining hair cells were counted in each animal. Immediately after the exposure, the EP and K+e decreased evenly in all four cochlear turns and gradually returned to normal physiological values in 5-20 days. When measured 20 days after the exposure, essentially normal EP and K+e values were observed, with an apicalwards decline, which was similar to that found along the cochlea in nonexposed animals. Abnormal increased EP was observed in some animals 20 days after the exposure in the first and second turns. In contrast to positive EP and K+e values, the anoxic negative EP attained less negative values in the second turn of exposed animals, i.e., in the turn where the narrow band noise exerted the major destructive effect. An almost normal distribution of hair cells and most negative EP values were found in the fourth turn. The distribution of persistent hair cells correlated positively with the values of the anoxic negative EP and amplitudes of the microphonic potentials. It is assumed that, in addition to the difference in K+ concentration between endolymph and perilymph, the anoxic negative EP is dependent upon the functional state of the organ of Corti.

The Davis theory: a review, and implications of recent electrophysiological evidence

The Davis theory of mechano-electrical transduction asserts that the endocochlear potential and the hair cell resting potential summate to provide a driving force for current flow through the hair cell. However, while a variety of agents which depress the endocochlear potential simultaneously depress auditory nerve sensitivity and reduce the cochlear microphonic, recent reports suggest that the hair cells may be depolarised without such effects ensuing. The relevant literature is reviewed.

Recovery of the endocochlear potential and the K+ concentrations in the cochlear fluids after acoustic trauma

Intense noise stimulation (142 dB, 1/3-octave-band noise centred at 1 kHz for 1 h) causes damage mainly in the second turn of the cochlea. Several hours (3-5) after the noise exposure, the endocochlear potential (EP) was found to be very low (5.7 +/- 6.0 mV). Similarly, the K+ concentration in the endolymph (Ke+) had decreased to low values (18.9 +/- 9.5 mM). The return of EP and Ke+ to normal values took 5-20 days. In contrast to the Ke+ changes, the perilymph K+ concentration (Kp+) increased slightly after the noise exposure to 4.5 +/- 1.7 mM and returned to normal values one day after the exposure. Differences were found in the time course of the EP, Ke+ and Kp+ changes after the arrest of ventilat ion when animals with acoustic trauma were com,ared with normal healthy individuals. The anoxic EP in noise-exposed animals never decreased to values more negative than -20 mV. The results imply that the inner ear mechanisms maintaining positive EP, Ke+ and Kp+ are severely damaged after acoustic trauma and that their function is restored in 5-20 days. With respect to some parameters (decrease of the EP during anoxia, the value of anoxic negative EP, EP overshoot after reventilation) the inner ear mechanisms are, however, still abnormal.

Evidence for intracochlear impedance changes following ethacrynic acid administration

The effects of intra-arterial 30-, 40-, and 50-mg/kg doses of ethacrynic acid upon cochlear function in guinea pigs were studied for periods of three to five hours. Cochlear potentials recorded in the first turn included the endocochlear potential, whole nerve response, cochlear microphonics, and summating potentials in scala media, scala tympani, and scala vestibuli. Evidence of organ of Corti damage at 50 mg/kg was found in addition to electrical impedance changes in the cochlear membranes at all dose levels.

Reduction of the endocochlear potential by the new "loop" diuretic, bumetanide

The effect of bumetanide upon the endocochlear potential (EP) was examined in 46 guinea pigs. The EP was reduced with dosages of 5 mg/kg or more. The maximum depression of the EP (reduction to -30 to -40 mV) was obtained at a dosage of 30 mg/kg. The recovery of the potential was incomplete at any dosage within three hours and the response pattern of the EP to bumetanide was similar to that to ethacrynic acid. The present experiments revealed that bumetanide, by weight, has a stronger ototoxic potency than the other "loop" diuretics--furosemide and ethacrynic acid. However, the diuretic effect of 1 mg bumetanide is equivalent to 40 to 60 mg furosemide or ethacrynic acid. Therefore, the relative ototoxic potency of bumetanide is much smaller suggesting that from a clinical standpoint bumetanide is much safer than the other "loop" diuretics.

The effects of ethacrynic acid upon the potassium concentration in guinea pig cochlear fluids

After i.v. injection of 50 mg/kg ethacrynic acid (EA), potassium concentration in the endolymph (Ke+) measured with K+-specific microelectrodes decreases by 10 mM at the most and endocochlear potential falls to negative values. Potassium concentration in the perilymph (Kp+) generally does not change, but sometimes a transient decrease in Kp+ level of about 0.5 mM was observed, presumably due to the electrogenic effect of the time-related decrease of the endocochlear potential. When anoxia is induced approximately 120 min after EA administration Ke+ slowly decreases. The decrease in Ke+ 50 min after the arrest of ventilation is smaller when compared with the Ke+ anoxic decrease without preceding EA administration. The endocochlear potential, which falls to negative values during anoxia after EA administration, does not return to the zero level as in the case when only anoxia is applied. Similarly, during anoxia, which follows EA administration, the perilymphatic Ke+ concentration increases more slowly than in the case when only anoxia is introduced. It is assumed from the results that EA abolishes activity of the positive electrogenic K+ pump and reduces the passive permeability of the walls of the cochlear duct to the potassium ions.

The generation of resting membrane potentials in an inner ear hair cell system

1. The macula sacculi in the mudpuppy is an inner ear sensory area accessible for intracellular recordings in vitro and in vivo. 2. The resting potentials recorded in vitro can be explained by the electrodiffusion theory assuming a uniform ionic selective in the membranes of the neuroepithelial cells. 3. The resting potentials recorded in vivo are significantly larger than predicted by the electrodiffusion theory, probably because of an electrogenic metabolic process present in the neuroepithelial cells. 4. An equivalent circuit is proposed to explain the resting electrogenesis in the neuroepithelial cells present in the sensory area.

Frequency response of the lateral-line organ of Xenopus laevis

The stimulus response relation of the epidermal lateral-line organ of Xenopus laevis was studied by recording activity of single afferent nerve fibres in isolated preparations. Linear frequency response analysis over a frequency range of 0.1--100Hz was performed under steady-state conditions, using small amplitude, sinusoidal water displacements produced by a glass sphere at a short distance from the skin. Period histograms of afferent nerve activity were computed, and amplitude, phase and mean activity of the response were determined by means of Fourier analysis. A standardization procedure at the start of each experiment made scaling of the frequency responses of different preparations unnecessary. The results show that for small stimulus amplitudes the response of the lateral-line organ over the whole range of frequencies studied can adequately be described as a modulation of the spontaneous activity. The amplitude of the response is proportional to the stimulus amplitude, and the phase of the response is independent of stimulus amplitude. The lateral-line organ of Xenopus laevis can thus be regarded as a linear system for stimuli which produce modulation of the spontaneous activity. The frequency response demonstrates unequivocally that the lateral-line organ of Xenopus laevis functions as a water velocity detector. For frequencies of stimulation from 0.1--20Hz the gain increases with a slope of 7.5 dB/oct, and up to 5Hz the response is almost in phase with the water velocity. The extent to which the different transmission steps between stimulus and response will contribute to the frequency response is discussed.

Noxious effects upon cochlear metabolism

The influence of various toxic substances and of drugs with ototoxic side effects upon energy generation, energy utilization, and membrane processes of the cochlea were studied. None of the drugs tested interfered with energy generation to as great an extent as did anoxia or cyanide and 2,4-dinitrophenol. Ouabain produced a pronounced interference with energy utilization of the stria vascularis. The "loop" diuretics ethacrynic acid and furosemide produced a reduction of energy utilization of a lesser degree than did ouabain. The "loop" diuretics do not seem to exert their toxic action upon strial Na+K+-ATPase, but may act by interfering with strial adenylate cyclase. Aminoglycoside antibiotics and diuretic and nondiuretic mercurials seem to exert their primary noxious action upon cochlear function by interfering with membrane processes of the structures bounding the cochlear duct.