Auditory nerve is also known as

Protection from excitotoxicity: “acquired” loss of auditory nerve connections to hair cells

The connection between IHCs and the AN is very sensitive to loss from the excitotoxicity generated by overrelease of the IHC transmitter, glutamate. This excitotoxicity can be induced by noise overstimulation or trauma to IHCs from hypoxia, changes in inner ear fluids or conditions that induce ROS in the hair cells. This can result in a “bursting” and loss of the unmyelinated portion of the peripheral process of the AN, normally connecting to the IHC. While there is potential for regrowth and reconnection [67] there is also potential for permanent loss. This can lead to subsequent death of the AN somata (spiral ganglion neurons—SGN) and loss of the central nervous system connection in the cochlear nucleus. A typical IHC may have connections to 10–30 AN peripheral processes (this varies depending on the position in the cochlear spiral and between species). Studies have shown that a mild to moderate level noise that produces only a temporary loss of hearing and no loss of sensory cells can still result in 20%–30% of these connections being lost, reducing the dynamic range of AN responsiveness [129,130]. Drugs that reduce excitotoxicity in the cochlea can be used to prevent or reduce the loss of IHC–AN synaptic connections. d’Aldin et al. [131] found that Piribedil, a dopamine agonist, reduced loss from a glutamate agonist. We found that the combinations of antiexcitotoxicity agents—Memantine, Piribedil, and ACEMg—when given prior to noise significantly reduced noise-induced loss of IHC-AN synaptic connections [132,133].

When there is loss of IHCs, a series of pathophysiological changes follow, including scar formation, a loss of the peripheral processes of the AN, and, over time, a substantial loss of the AN itself. This loss of AN can be related to the loss of survival/maintenance factors from hair cells and/or supporting cells. Loss of these survival/maintenance factors causes these auditory neurons to enter into the cell death cycle. Cochlear prostheses (discussed in a later section) depend on the stimulation of the AN, so large AN loss would impair function. Replacing lost survival factors can help maintain this population. Placement of GDNF, BDNF, or NT-3 into cochlear fluids following IHC loss significantly reduced subsequent loss of SGN [115,123,134–138]. Chronic electrical stimulation also reduced SGN loss [139–144]. The combination of neurotrophic factors and ES is particularly effective [145,146].

Infusion of neurotrophic factors such as BDNF, NT-3, and FGF can induce regrowth of the peripheral process of the AN that regress following IHC loss [101,138,147,148]. There is interest in inducing the regrowth of AN processes toward specific sites on the cochlear prostheses. Cochlear prostheses might have reduced thresholds if their target is closer, making them more energy efficient and reducing battery requirements, and it might also allow a greater number of channels to be used, resulting in more and better frequency separation.

Neurotrophic factors delivered into cochlear fluids can also be used to induce IHC-AN reconnection when hair cells are not lost. NT-3 delivered into cochlear fluids by poloxamer on the round window, given 1 day after a noise, induced significant reconnection of IHC-AN synaptic connections [149].

Auditory brainstem implants

There are rare cases when the auditory nerve cannot be stimulated because it has been damaged or, in even rarer cases, not formed during development (as in some cochlear malformations). In these cases, the cochlear implant is not helpful. For these subjects, auditory brainstem implants (ABIs) have been developed. These consist of the same hardware as the cochlear implant with a different electrode. The electrode is placed on the surface of the cochlear nucleus, the first station of the auditory pathway. The access to this structure is surgically complex, requiring access to the brain, thus opening the dura mater and displacing the cerebellum. This is a neurosurgical intervention with all associated risks.

A further complication is the anatomy of the cochlear nucleus. It contains three sub regions: the dorsal cochlear nucleus, the anteroventral cochlear nucleus, and the posteroventral cochlear nucleus. All these structures have a complete tonotopic representation of the cochlea. In contrast to the cochlea, they contain a high diversity of neurons, both in terms of anatomy and functional characteristics. It is still unclear which neurons can be stimulated by electric stimulation, and it remains unknown which ones are the optimal targets for such stimulation.

The first brainstem implants used only two electrodes for stimulation; only over time have surface arrays with larger sets of electrodes been developed. In addition, attempts were made to access the neurons that show temporally precise timing synchronized with acoustic stimuli (bushy cells). These are located deeply in the anteroventral cochlear nucleus, and require penetrating electrodes. Since this approach did not bring any improvement in performance, only surface electrodes are in clinical use currently.

The anatomy of the cochlear nucleus is complex in this respect, as shown in Fig. 11.17. While each segment of the cochlear nucleus contains a full representation of the cochlea and is tonotopically organized, it is interesting that in humans the innervating fibers do not twist towards the deeper parts. Thus here the tonotopic gradient may be perpendicular to that of a cat. This difference is of cardinal importance for the placement of the ABI. The electrode array is a two-dimensional grid containing up to 21 electrodes on a silicone carrier (Fig. 11.18). Only the caudal region of the dorsal cochlear nucleus is accessible for such surface grid. Here, in close proximity are low-frequency fibers and neurons. Thus, different electrodes of the implant do not have spatially specific access to different frequency representation, as in cochlear implants.

Only a small opening to the fourth ventricle, called the apertura lateralis or foramen Luschkae, is available to access the cochlear nucleus underneath the cerebellum. This opening allows the cerebrospinal fluid to circulate around the brain. It is located at the border between the pons cerebri and the medulla oblongata near the place where the IX nerve enters the brainstem. This needs to be identified and the brainstem implant inserted there. The surgeon has no real control of the exact position of the brainstem implant in this very tricky procedure. It is secured in place usually by a muscle fascia.

Fig. 11.19 shows the surgical view of the complicated access to the cochlear nucleus. Cranial nerve XII is located deep underneath nerve XI on panel A. Nerve VIII is deep underneath between nerve VII and nerve IX on panel B. The implant must be squeezed into the foramen Luschkae. Beyond this, the surgeon has limited control. Therefore, exact placement is one of the ongoing issues of ABIs.

The thresholds for the brainstem implants are greater than those of cochlear implants: more than 33 μA (10 nC/phase) for biphasic pulses with phase duration of 300 μs. However, they can also be substantially greater (> 666 μA, i.e., > 200 nC/phase), but still within the compliance limits of a cochlear implant processor. This higher threshold is likely due to the distance between the stimulated neurons and the implant. Another typical observation was the high variability of the thresholds between electrodes, which is not so common in cochlear implants with monopolar stimulation.

Given these factors, speech performance is significantly poorer with brainstem implants compared to cochlear implants (Fig. 11.20). Furthermore, the neurons in the cochlear nucleus show a high diversity and include inhibitory neurons. The electrical stimulation from the surface of the dorsal cochlear nucleus (DCN) cannot focus on one type of cells and is therefore assumed to have mixed effects, and generates smaller cortical responses than cochlear implants. The main factor for brainstem implant underperformance compared to that of a cochlear implant is the etiology of deafness and thus the associated pathology in the auditory system. The most common disease associated with brainstem implantation is neurofibromatosis 2 (NF2). NF2 often involves tumors in the auditory nerve that damage it. Additionally, it involves damage to the central nervous system (with other tumors occurring both in the spine and in the brain). Therefore, the cochlear nucleus is likely compromised by the disease itself and/or the expansive growth of the tumor in the auditory nerve. When subjects with head trauma and thus traumatic injury of the nerve were implanted, their best performance approached the performance observed with cochlear implants (Fig. 11.20). In addition, the type of surgery is important, where a semi-sitting position provides better outcomes than a lateral (lying) position. Nonetheless, the semi-sitting position involves more risks during surgery. Unfortunately, there is not yet a clinical consensus of all these aspects.

The stimulation strategy and the speech processor are the same as in cochlear implants, thus designed for the auditory nerve and not for the cochlear nucleus. The small number of subjects using the implants (several thousand worldwide) complicates assessment of such aspects, since each (even large) clinical center implants only a limited number of such devices each year. Despite all these issues, the brainstem implant is the most successful functional prosthesis delivering information to the central nervous system (as opposed to a “pacemaker” delivering only pacing, such as in deep brain stimulation). There is a small group of deaf subjects that can only profit from brainstem implants, including some deaf-born children with developmental abnormalities of the cochlea, and some show promising results.

Loss of Auditory Nerve Connections and Auditory Nerve

The connection between inner hair cells and the auditory nerve is very sensitive to loss from the excitotoxicity generated by over-release of the inner hair cell transmitter, glutamate. This excitotoxicity can be induced by noise overstimulation or trauma to inner hair cells from hypoxia, changes in inner ear fluids or conditions that induce ROS in the hair cells. This can result in a 'bursting' and loss of the unmyelinated portion of the peripheral process of the auditory nerve, normally connecting to the inner hair cell. While there is potential for regrowth and reconnection [28] there is also potential for permanent loss leading to subsequent death of the auditory nerve somata (spiral ganglion neurons) and loss of the central nervous system connection in the cochlear nucleus. A typical inner hair cell may have connections to 10–30 auditory nerve peripheral processes (this varies depending on the position in the cochlear spiral and between species). Studies have shown that a mild to moderate level noise that produces only a temporary loss of hearing and no loss of sensory cells can still result in 20–30% of these connections being lost, reducing the dynamic range of auditory nerve responsiveness [130,131]. Drugs shown to reduce excitotoxicity in the cochlea [28] or in other regions could provide a means to reduce this loss.

When there is loss of inner hair cells, a series of pathophysiological changes follow, including scar formation, a loss of the peripheral processes of the auditory nerve, and, over time, a substantial loss of the auditory nerve itself. This loss of auditory nerve is related to the loss of survival/maintenance factors, including the deafferentation-associated loss of neural activity and the loss of neurotrophic factors that had been released by hair cells and other cochlear elements lost during the scar process. Loss of these survival/maintenance factors causes these auditory neurons to enter into the cell death cycle.

Cochlear prostheses depend on direct stimulation of the auditory nerve, so auditory nerve survival is of major importance. The auditory nerve loss can be blocked or reduced either by blocking the cell death pathway or by replacing the survival/maintenance factors. Thus, activity can be replaced by direct cochlear electrical stimulation E1 with a cochlear prosthesis, and auditory nerve survival is enhanced [132–137]. Stimulation may serve as a survival factor through activation of voltage-gated ion channels and/or through upregulation of autocrine factors, including neurotrophic factors. Neurotrophic factors can be infused into the cochlear fluids, to replace those that are lost and this is even more effective than electrical stimulation in enhancing auditory nerve survival following deafness [66,138–141]. The combination of electrical stimulation and application of neurotrophic factors is more effective than when either is applied singly [142,143].

There is also interest in inducing regrowth of peripheral processes of the auditory nerve. Cochlear prostheses might have reduced thresholds if their target is closer, making them more energy efficient and reducing battery requirements. If regrowth can be directed toward specific sites on the cochlear prostheses, this might also allow a greater number of channels to be used, resulting in more and better frequency separation. Studies have now shown that infusion of neurotrophic factors BDNF, NT-3, and FGF can induce regrowth of the peripheral process of the auditory nerve that regress following inner hair cell loss [52,144–146]. Thus, neurotrophic factors can serve not only to enhance survival of the auditory nerve but also to restore its peripheral processes, although different factors may be necessary for most effective treatment(s). Directed regrowth remains a challenge.

Speech processing

If sufficient spiral ganglia and the auditory nerve are intact, hearing may be augmented by providing electrical stimulation to the cochlea. Exactly reproducing the original auditory processing of speech is impossible. However, brain plasticity enables certain patients to decode new electrical stimulation patterns over time as speech when frequency-place coding (the location of basilar membrane vibration) and temporal coding (synchronization or phase-locking of ganglia to the period of a pressure wave) are at least grossly preserved. Adult patients who had language skills before deafness can associate new stimulation patterns with their memories of sounded speech. Pediatric patients younger than the age of 2 years can learn spoken language at a normal or near-normal rate because their brains quickly adapt to electrical stimulation. But older children who did not have aural stimulation during their early years have a much more difficult time acquiring speech and oral language skills. What would normally be auditory areas of their brains have been reallocated for different tasks before cochlear implantation.

In many systems, cochlear implant processing is based on the continuous interleaved sampling (CIS) processing strategy. For this processing strategy, acoustic sound is first filtered to preemphasize speech frequencies less than 1.2 kHz. The prefiltered speech is bandpass filtered into 4–22 channels. Each frequency range is then rectified and lowpass filtered (at a relatively high cutoff frequency, such as 200–400 Hz), to obtain a speech envelope. Each envelope signal is compressed with a nonlinear mapping function in order to map the wide dynamic range of sound into the narrow dynamic range of electrically evoked hearing. The extracted envelope is used to modulate a biphasic pulse train. Each biphasic pulse channel output is directed to a single intracochlear electrode, with low- to high-frequency channels assigned to apical-to-basal electrodes. This mimics the order of frequency mapping in the normal cochlea. The pulse trains are interleaved in time, so that pulses across channels are not simultaneous and channel interaction is reduced (Fig. 15.9).

Fig. 15.9 simplifies how each envelope signal was originally mapped to an electrode channel. A single pulse train was sent from the transmitter coil to the receiver coil. Four channels of envelope data were interleaved onto this pulse train, with each channel using pulse amplitude modulation (PAM). In PAM, the amplitudes of regularly spaced pulses are proportional to the corresponding sample values of a continuous envelope signal, e(t). For each channel, the PAM signal, s(t), is mathematically equivalent to the convolution of eδ(t), the instantaneously sampled version of e(t), and the pulse, h(t):

(15.2)s(t)=eδ(t)*h(t).

At the receiver end, the pulse train was demultiplexed to obtain separate channels of envelope data.

Various processing strategies preserve formants, which are groups of overtone pitches or vocal tract resonances, in spoken words. For any word, there is no unique signature involving specific frequencies. Rather, a formant pattern encompassing the relationship between two dominant magnitude peaks allows speech to be flexible. This flexibility enables normal speech variations, such as speech patterns from people in northern versus southern US states, to be accurately recognized. When normal speech is bandpass filtered and transformed into four amplitude-modulated sine waves, normal-hearing listeners can understand about 90% of words in simple sentences (Fig. 15.10) (Dorman & Wilson, 2004; Wilson & Dorman, 2009).

Combinations: Prostheses with Microchannels

Studies now show that combining electrical stimulation and chemical delivery is more effective in enhancing auditory nerve survival following deafness than either applied by itself (Kanzaki et al., 2002; Shepherd et al., 2005). Moreover, more patients with some residual hearing have been shown to benefit from cochlear implants and more patients with surviving hair cells are now considered candidates for cochlear implants. These subjects would benefit from protection of these remaining hair cells from the trauma of cochlear prosthesis insertion. There is also the potential of inducing regrowth of peripheral processes toward the stimulation sites, which could lower thresholds and enhance selectivity and separation. Therefore recent efforts have been made by cochlear implant manufacturers and research groups to develop cochlear prostheses capable of both electrical stimulation of the auditory nerve and delivery of pharmaceuticals into the cochlear fluids (Altschuler et al., 2005). These have been successfully applied in animal studies (Shepherd et al., 2005), and clinical application is beginning.

Damage to the Periphery

Sensorineural deafness caused peripherally can result from serious damage to the hair cells or to the auditory nerve. Clearly, if the neurons of the hearing nerve are damaged, their peripheral processes cannot be driven, and stimulation from sites in the cochlea will not work. In those cases, central prostheses have been used experimentally (Shannon et al., 1993; McCreery et al., 1997; Lim et al., 2008; McCreery, 2008).

Damage to the hair cells can result from a number of causes. Pyman et al. cite eleven root causes in people over six years of age, and seven causes in people less than six years of age (Pyman et al., 1990). Their population was 65 people in the former case, and 29 in the latter. Large numbers of subjects had unknown causes of deafness, but there were cases of meningitis, otosclerosis, and trauma that caused the problems (Pyman et al., 1990). In another study, Hinojosa and Marion analyzed 65 ears and found six causes of congenital deafness in 19 subjects, and nine causes of acquired deafness in 46 subjects. In the latter population, otosclerosis caused the greatest damage, followed closely by bacterial infections (Hinojosa and Marion, 1983).

Damage to the hair cells from loud sounds requires special mention, since the popularity of painful audio systems in automobiles, and as portable sources of entertainment, is increasing. Hair cells can be damaged by intense sounds (Popper and Fay, 1991); chronic exposure to loud sounds should be avoided, despite the relatively small numbers cited by Hinojosa and Pyman (Hinojosa and Marion, 1983; Pyman et al., 1990).

4.3.1.2 Low-power analog processor for cochlear implants

Cochlear processors used in deeply deft patient implants transform sound into spatial electrode stimulation patterns for the auditory nerves. These processors can be implemented using either an analog preprocessing (Figure 4.1b) or only a digital signal processing (Figure 4.1a). However, in this application, the analog signal preprocessing has shown to be more power-efficient than the digital processing in more than an order of magnitude [33] by compressing incoming data and using low-resolution and low-speed ADC at later stages of computation. The analog circuits also ensure robustness, that is, rejection to power supply noise, thermal noise, temperature variation, cross talk, and immunity to process variation and artifacts.

In Figure 4.5, is depicted the architecture of an analog cochlear implant processor described in Ref. [33] that runs a low-power algorithm termed as asynchronous interleaved sampling (AIS), originally presented in Ref. [34]. This AIS-based processor, conceived for next generation of fully implantable cochlear prostheses, will allow deaf patients not only speech hearing but also music hearing by providing fine-phase-timing encoding for auditory nerve stimulation.

The AIS algorithm allows detecting the input channel (neuron) with the highest intensity to perform channel sampling at a higher rate than in the lower-intensity channels. This asynchronous sampling approach prevents simultaneous channel stimulation and spectral smearing in the output stimulation patterns. The working principle is based on an array of neural capacitors that are charged by halfway rectified current outputs (envelope detection). The first charged capacitor that crosses a predefined threshold wins the race among all channels, and a pulse (spike) is fired to signal event detection and also to reset all other capacitors. Then, a negative current is applied to the firing neuron capacitor to inhibit subsequent firing in the same channel during the neuron refractory period. The adaptive sampling rate, which also determines the output stimulation rate, considers the time and spectral signal content of the signal in each channel to avoid power consumption during quite periods and to sample at higher rates the high-intensity predominant channels.

The AIS-based processor in Figure 4.5 works as follows [33,35]. The microphone detects sound and outputs a signal with a wide DR of 75 dB. Subsequently, the microphone signal is preamplified and then compressed using an AGC amplifier, which reduces the internal DR to 55 dB. The AGC circuit is implemented by varying gm in a transconductance-resistance variable gain amplifier. The compressed signal is fed to a full processing chain of 16 parallel channels. Each channel performs band-pass filtering in the suitable frequency range using two cascaded blocks of second-order OTA Gm–C circuits. Next, an envelope detector halfway rectifies the compressed signal and outputs peaks (spikes) that are fed to the AIS processor implementing the algorithms introduced above. The compressed signal is used also as the input of the logarithmic ADC to extract the logarithm of the spectral energy in the corresponding spectrum band. A dual-slope ADC circuit is implemented using a proportional-to-absolute temperature (PTAT) circuit at the input, which converts the output current from the envelope detector into a logarithmic voltage at the input of a wide-linear-range transconductor circuit followed by a standard comparator. The PTAT circuit allows for the temperature and offset compensation required to ensure analog circuits robustness. The spike produced by the enveloped detector also enables the tristate buffer of the winning channel connected to a common bus to output the 7-bit digitized envelope amplitude. This output provides in a single event both the amplitude information and the fine phase timing required for generating the electrode stimulation pattern of the auditory nerves.

The analog preprocessing approach adopted for this cochlear implant allowed reducing the power consumption of the whole processor to 357 μW using a 1.5 μm CMOS process compared to the 5 mW estimated by the authors for the digital signal processing approach.

Is auditory nerve same as acoustic nerve?

Where are the auditory nerves?

What is the name and number of the auditory nerve?

What is the vestibulocochlear nerve?