KeywordsCochlear Implant Auditory Nerve Electrode Array Cochlear Implantation Auditory Nerve Fiber
An auditory prosthesis is an implantable device used to (partially) restore the auditory function in people with a severe to profound hearing loss by electrically stimulating the auditory neural pathway. The cochlear implant, stimulating the auditory nerve from within the cochlea, is widely accepted as the standard rehabilitation device for this population. An auditory brain stem implant uses the same technology to stimulate the neurons of the cochlear nucleus in the brain stem and is used when the cochlea is not accessible (e.g., due to ossification after meningitis or severe hypoplasia) or the cause of deafness is found in the internal auditory canal (bilateral acoustic neuroma, aplasia of the auditory nerve).
Due to their reduced oral communication skills, severe to profoundly deaf people are restricted in their social functioning. Since the pioneering work of Djourno and Eyries (1957) and House in the 1970s of the last century (House and Urban 1973), it has become possible to restore some of the hearing functions through direct electrical stimulation of the auditory nerve. Currently used cochlear implants utilize electrode arrays with 12–22 contacts on a Silastic carrier that are most commonly inserted into the scala tympani through either the round window membrane or a drilled cochleostomy in its vicinity. This allows taking advantage of the tonotopic organization of the cochlea and the auditory nerve, where the high frequencies are encoded at the basal end, while the low frequencies are encoded more to the apical end (Biomedical and Life Sciences > Encyclopedia of Neuroscience > Cochlea). In this way, each electrode contact of a multichannel implant aims to stimulate a different neural population, which physiologically encodes a certain pitch as determined by its intracochlear position. With current devices, however, the wish to encode all spectral information relevant for speech understanding leads to a mismatch between their tonotopic map and the physiological one.
Components and Signal Processing of Multichannel Implants
An otolaryngologist surgically implants the internal part (the so-called receiver-stimulator package with the electrode array) under general anesthesia; the externally worn speech processor (body worn or behind the ear) is connected after several weeks of wound healing. The speech processor captures the incoming sound and, after preprocessing (typically noise cancellation and amplitude compression), encodes it into frequency-specific electrical information to be sent to the individual electrode contacts in the cochlea.
The speech coding strategies used in all current implants are extensions of the continuous interleaved sampling (CIS) strategy (Wilson et al. 1991). This strategy tries to avoid electrical interaction between neighboring electrode contacts by stimulating all contacts in a sequential mode rather than simultaneously. A digital filter bank is used to process the signal into separate frequency bands. Next, the envelope of each band, determined by rectification and low-pass filtering of the signal, is used to set the amplitude of a sequence of nonsimultaneous pulses on the implanted electrode contacts. The rate of stimulation is determined by the device brand and by the patient’s performance, but typically ranges between 400 pulses/s and 4,000 pulses/s per channel.
Both the encoded stimulation pattern and the energy are transmitted to the implanted receiver-stimulator package through an RF-link. The external and internal coils for this RF-link are kept aligned with magnets in the center of both coils. The signal is picked up by the electronics in the receiver-stimulator package, which in turn delivers electrical pulses to the auditory nerve fibers via electrode contacts in the electrode array.
Modern cochlear implants also have back telemetry, allowing to record electrically evoked compound action potentials (eCAPs) of the auditory nerve via the implanted electrode array.
To provide more insight in the fundamentals of functional electrical stimulation of the auditory nerve, computational models have been developed. This involves stimulating not only the response of a nerve fiber to an externally applied potential field but also the calculation of this potential distribution from the currents on the stimulating electrodes. This is especially intricate in the case of cochlear implants due to the complex geometry of the inner ear.
Electrical Volume Conduction in the Cochlea
An analytic solution of such a 3D volume conduction problem is restricted to geometries that are much simpler than the cochlea, and many theoretical models on the (actually three-dimensional) potential pattern set up in the cochlea by the stimulating current sources assumed an exponential decay of current from its source to the nerve fibers along the cochlea, modeled in one dimension (O’Leary et al. 1985), while other analytical approaches assume a simplified unrolled anatomy (Goldwyn et al. 2010). Suessermann and Spelman (1993) used an electrical network as a practical representation of the electro-anatomy of the cochlea.
Simulating the Auditory Nerve Fiber Responses
Colombo and Parkins (1987) were the first to develop a cable model of the mammalian auditory nerve neuron based on the classical work on amphibian nerve fibers of Frankenhæuser and Huxley (1964). In order to fine tune the model to represent physiological data obtained from single auditory nerve fiber experiments in squirrel monkeys, they had to adapt the modeled nerve fiber’s anatomy significantly. Motz and Rattay (1986) used a single-node model with the Hodgkin and Huxley model of unmyelinated squid giant axon membrane (http://www.springerreference.com/docs/html/chapterdbid/348192.html) to investigate the time structure of the response of the (myelinated!) auditory nerve to electrical stimuli. The gSEF model (Frijns et al. 1995) is a nonlinear cable model, which represents essential mammalian nerve fiber properties, including spike duration and conduction velocity, refractory behavior, and repetitive firing, better than previous models and can deal with arbitrary stimulus wave forms. It is based upon voltage clamp measurements in rat and cat motor nerve fibers at mammalian body temperature performed by Schwarz and Eikhof (1987). The gSEF model and its variants have, in conjunction with electrical volume conduction models, been used not only to predict which (intact or degenerated) fibers are excited by specific patterns of electrical stimulation (Smit et al. 2010; Frijns et al. 2009a, 2011) or to explain the results obtained with psychophysical experiments (Carlyon et al. 2010; Snel-Bongers et al. 2013) but also to calculate the eCAP produced on the basis of predicted single fiber action potentials (Briaire and Frijns 2005; Westen et al. 2011).
The abovementioned neural models have in common that they are deterministic in the way they treat the neural membrane responses. If the focus of research is more on the effect of high stimulation rates or on repetitive near-threshold stimulation, stochastic models come into play. Most models of this type are single-node threshold models (Bruce et al. 1999), while cable models (Rubinstein et al. 1999; Imennov and Rubinstein 2009), although computationally very intensive and requiring supercomputers, can give insight in more complex stimulation patterns.
Integrated Use of Volume Conduction and Neural Models: State of the Art
While in the early days of cochlear implantation all insights in the mechanisms underlying their function had to come either from clinical practice and associated psychophysics or from animal experiments, nowadays sophisticated computational models exist, which integrate a model of electrical volume conduction in the cochlea with active neural models. Such models can not only be used to explain effects of current and future electrode designs and stimulation schemes but are also able to predict the of anatomical variations (Frijns et al. 2009b), species differences (Frijns et al. 2001), and the effects of neural degeneration (Briaire and Frijns 2006; Snel-Bongers et al. 2013) (Fig. 3).
- Biomedical and Life Sciences > Encyclopedia of Computational Neuroscience > Hodgkin and Huxley (http://www.springerreference.com/docs/html/chapterdbid/348192.html)
- Carlyon RP, Macherey O, Frijns JHM, Axon PR, Kalkman RK, Boyle P, Baguley DM, Briggs J, Deeks JM, Briaire JJ, Barreau X, Dauman R (2010) Pitch comparisons between electrical stimulation of a cochlear implant and acoustic stimuli presented to a normal-hearing contralateral ear. J Assoc Res Otolaryngol 11(4):625–640PubMedCentralPubMedCrossRefGoogle Scholar
- Frankenhæuser B, Huxley AF (1964) The action potential in the myelinated nerve fiber of Xenopus laevis as computed on the basis of voltage clamp data. J Physiol (Lond) 171:302–315Google Scholar
- Whiten D (2007) Electro-anatomical models of the cochlear implant. PhD thesis, Massachusetts Institute of Technology, Cambridge, MAGoogle Scholar
- Biomedical and Life Sciences > Encyclopedia of Neuroscience > Cochlea http://www.springerreference.com/docs/html/chapterdbid/114643.html; http://www.springerreference.com/docs/html/chapterdbid/34Sig2.html
- NIH factsheet cochlear implants: http://report.nih.gov/nihfactsheets/ViewFactSheet.aspx?csid=83