Current Cochlear Implant Systems and Processing Strategies
There are currently three FDA approved cochlear implant systems available today in the United States (Clarion, Nucleus and MED-EL). Competition among these manufacturers is fierce. Generally, the results of this competition have been good, encouraging all three companies to invest significantly in research and development and to strive for improvements in functioning and packaging. A less desirable outcome of this competition is the resulting increase in marketing. It can be difficult for prospective cochlear implant candidates and their families to sort through all of the information available today and to separate advertising claims from proven facts. The process of becoming an informed consumer is made even more difficult because the terminology used to describe the ear and the individual cochlear implant systems typically is foreign to the average person. The goal of this section is to describe in general terms how cochlear implants produced by each of the three major manufacturers are similar and then to highlight some of the features that distinguish these three different multichannel cochlear implant systems.
- The Nucleus Cochlear Implant Systems
- The Clarion Cochlear Implant Systems
- The MED-EL Cochlear Implant
- Special Electrode Arrays
- Device Selection
The Nucleus Cochlear Implant Systems
The Nucleus 22-channel cochlear implant was the first cochlear implant to receive FDA approval for use in adults and children and has been used in more individuals than any other cochlear implant system worldwide. The original design of the intracochlear component of the Nucleus 22 device consisted of 22 banded electrodes spaced at equal intervals (approximately 4 mm). As the electrode array was inserted into the cochlea, the anatomy of the cochlea caused it to curl around the basal turn. This resulted in an intracochlear array that lay along the outer wall of the cochlea opposite the modiolus.
The original Nucleus 22 device could be programmed to stimulate in one of several different bipolar modes or in a stimulation mode that Cochlear Corporation referred to as common ground. In a bipolar mode the current is passed between two intracochlear electrodes. These electrodes may be adjacent to each other (BP) or spaced slightly more widely apart depending on the subject’s sensitivity (e.g. BP+1, BP+2 etc). With common ground stimulation, one electrode is designated as the active electrode and the other 21 intracochlear electrodes are shorted together and used as the return path. Monopolar stimulation, where stimulation is applied between one intracochlear electrode and an extracochlear ground electrode, was not possible with the first version of the Nucleus device.
Research from animals had shown that bipolar stimulation, particularly at low stimulation levels, resulted in activation of a small group of auditory nerve fibers located relatively close to the stimulating electrode pair (van den Honert & Stupulkowski, 1984). Good place specificity, achieved via the use of bipolar stimulation, was considered crucial to the success of a multi-channel cochlear implant because there was no longer a traveling wave to provide frequency selectivity. High frequency signals were routed to the most basal electrode pairs and low frequency signals were routed to the more apical electrode pairs.
The Nucleus 22 cochlear implant and all subsequent Nucleus devices provide only non-simultaneous, pulsatile stimulation. That is, the output of the cochlear implant consists of a series of biphasic, current pulses that vary in amplitude depending on the intensity of the incoming signal. No two electrodes can be stimulated simultaneously and analog stimulation is not possible with this cochlear implant system. The advantage of using non-simultaneous stimulation is that no two electrodes are ever stimulated at exactly the same instant. This minimizes the chance of deleterious channel interactions. Additionally, power consumption is significantly lower when pulsatile rather than analog stimulation is used to encode the speech signal. The disadvantage of using pulsatile stimulation is that the amount of information conveyed per unit time is directly dependent on the duration of the individual pulses and the overall stimulation rate.
Early speech processing strategies used with the Nucleus 22-channel cochlear implant employed feature-extraction schemes that conveyed fundamental frequency information as well as information about the first two formants of speech (F0F2 and F0F1F2). In the early 1990’s the MPEAK processing strategy was introduced. This strategy still used feature extraction algorithms but also provided additional high frequency information by stimulating two or three fixed, basal electrodes. The goal was to provide additional information about frication that would yield improved consonant recognition scores. The maximum stimulation rate used for these early speech-processing strategies was 250 Hz. Most recipients used speech processor programs constructed using bipolar stimulation, 205 mðs/phase biphasic current pulses with 19–20 electrodes available for stimulation.
Over the course of the next decade, the speech processing algorithms that were used with the Nucleus cochlear implant system moved away from feature extraction schemes. In 1995 Cochlear Corporation introduced the spectral peak (SPEAK) processing strategy. This strategy samples the incoming acoustic signal, converts that signal to the frequency domain, and identifies 6–10 peaks in the acoustic spectrum. A look-up table is used to determine how the output of the 20 separate frequency bands will be routed to the individual intracochlear electrodes. On each stimulation cycle a subset of 6 to 10 intracochlear electrodes are stimulated non-simultaneously at a rate that varies adaptively between 180–300 pulses per second depending on the number of spectral peaks identified.
Early in 1998 a new internal device, the Nucleus 24 system was introduced. The intracochlear electrode array of the Nucleus CI24M device was no different from the array used with the previous Nucleus cochlear implants, however, two additional extracochlear electrodes were added. With this version of the Nucleus cochlear implant it was possible to stimulate in a monopolar stimulation mode using pulse durations as short as 25 us/phase. Stimulation rates on an individual electrode as high 2400 Hz pulses per second could be achieved and the device was designed such that the implanted magnet could be removed if necessary in order to allow for magnetic response imaging (MRI).
The other significant change in the Nucleus 24 device relative to previous versions of the implant was that it was possible to use radio frequency telemetry to transmit information about electrode impedance and device function from the internal device out to the programming system. Additionally, this device has the capability of using implanted electrodes not only to stimulate the ear but also to record electrically evoked auditory potentials from within the cochlea. Such information has proven helpful in programming the speech processor for very young children (Brown, Hughes, Luk, Abbas, Wolaver, and Gervais, 2000). Finally, with the emergence of the Nucleus 24 device, each recipient had the advantage of choosing between a body-worn speech processor or an ear-level speech processor.
Shortly after the Nucleus CI24M cochlear implant was introduced, Cochlear Corporation introduced a revision of this system. They called the new device the Nucleus 24 Contour (CI24RCS). The primary difference between the Nucleus 24 (CI24M) and the Nucleus 24 Contour (CI24RCS) devices was that the intracochlear array of the Contour device is pre-coiled but is held in a straight position during insertion by a stylette, or flexible metal spine, that runs the length of the array. The stylette is removed during the insertion process to allow the array to coil closer to the modiolus of the cochlea where the surviving auditory nerve fibers are located. Closer proximity between the stimulating electrodes and the surviving neural elements within the modiolus resulted in lower thresholds and reduced current spread. Modiolar placement also effectively decreases power consumption and enhances place specificity. The intracochlear electrode contacts are spaced logarithmically along the array with electrodes at the base being more widely separated than electrodes at the apex. The contacts are half bands rather than the full bands used with the Nucleus CI24M device and all earlier versions of this implant. Additionally, the packaging of the internal receiver/stimulator of the Nucleus 24 Contour is thinner and more flexible that earlier versions resulting in a lower profile on the skull.
With these more recent versions of the Nucleus cochlear implant system, the Nucleus 24 (CI24M) and the Nucleus 24 Contour (CI24RCS) devices, it became possible to stimulate in a monopolar mode. With monopolar stimulation, all 22 intracochlear electrodes can be used as active electrodes and stimulation is applied to an intracochlear electrode relative to one of two extracochlear ground electrodes. Monopolar stimulation results in lower thresholds and therefore requires less power consumption than processing strategies using bipolar or common ground stimulation modes. Additionally, the threshold and maximum stimulation levels that are obtained when monopolar stimulation is used are more consistent across the electrode array than those obtained when bipolar stimulation is used. Initial concerns that monopolar stimulation would not be place specific proved unfounded. Persons who use monopolar stimulation are able to pitch rank and generally perceive a monotonic decrease in pitch as the stimulating electrode is moved from the base to the apex of the cochlea. This finding indicates that the electric fields that result when monopolar stimulation is used are concentrated near the stimulating electrode and as such are still relatively place specific.
The Nucleus 24 and the Nucleus 24 Contour devices also offer two additional speech coding strategies. The first was a strategy Cochlear Corporation describes as the n-of-m strategy. This speech processing strategy, better known as ACE (Advanced Combined Encoder), allows the programming audiologist to specify both the specific number of spectral peaks (n) that should be identified as well as the number of different bandpass filters (m) that should be used to divide up the acoustic spectrum on any stimulation cycle. ACE is typically implemented by selecting 8–12 spectral peaks (n) and speech is subdivided into a total of 22 bandpass filters (m). This strategy is similar to the SPEAK strategy but operates at a faster stimulation rate. The majority of persons being fitted with Nucleus cochlear implants today use the ACE strategy at stimulation rates between 900 and 1200 Hz per channel.
The second new speech processing strategy available in the Nucleus 24 cochlear implant system that is referred to as Continuous Interleaved Sampling (CIS). The CIS strategy filters the speech signal into a fixed number of bands (typically 8–12), obtains the speech envelope for each band, and provides compression. A look-up table is used to determine which electrode will be stimulated for each of the specified frequency bands. With this stimulation strategy, not all 22 intracochlear electrodes are used, but every electrode is stimulated on each cycle of stimulation and stimulation rates are typically higher than those used with other speech processing strategies. When the CIS programming strategy is used, each electrode is stimulated sequentially with a biphasic current pulse that has an amplitude proportional to the amount of energy in the corresponding frequency band. This strategy is designed to preserve fine temporal details in the speech signal by using high rate, pulsatile stimuli.
Cochlear introduced the first ear-level speech processor, ESPrit 24, for Nucleus 24 recipients in 1998. Ear-level speech processors compatible with the older Nucleus 22 device, known as the ESPrit 22, became available in 2000. Both of these behind the ear processors are powered by two hearing aid batteries and have an average battery life of 50 hours for Nucleus 24 recipients and 35 hours for Nucleus 22 recipients (ESPrit User Manual). These original ear level processors were less flexible than the body-worn Sprint processor. They were designed to implement the SPEAK processing strategy for Nucleus 22 recipients and either the SPEAK or ACE strategies for Nucleus 24 recipients. The newest Nucleus behind the ear processor, ESPrit 3G, can implement all three Nucleus speech processing strategies, SPEAK, ACE, and CIS. It also has an integrated telecoil connection and is powered by three 675 hearing aid batteries.
The Clarion Cochlear Implant Systems
The second cochlear implant system that is available in the United States today is the Clarion multichannel cochlear implant system manufactured by Advanced Bionics Corporation. This device was approved by the FDA for use in adults in 1996 and in children in 1997. Like the Nucleus device, the Clarion cochlear implant system has undergone a series of changes over the past several years. The original Clarion (Versions 1.0 and 1.2) consisted of an array of 16 intracochlear electrodes arranged in 8 closely spaced electrode pairs that were oriented radially, rather than longitudinally, within the cochlea. This “radial bipolar” configuration was selected based on early physiological, electrophysiological, and computer modeling studies that demonstrated this configuration resulted in optimal place specificity (van den Honert & Stupulkowski, 1984). This device could be programmed in either a monopolar or a bipolar mode and resulted in a maximum of 8 stimulation sites (channels). Because each channel or site of stimulation had an independent output circuit, each channel could be programmed independently allowing for either non-simultaneous or simultaneous patterns of electrode (channel) activation. Later iterations of the Clarion electrode array used a stimulation pattern referred to as Enhanced Bipolar. In this stimulation mode, the medial electrode in one pair was stimulated in a bipolar fashion relative to the lateral electrode in the next most apical electrode pair. The wider electrode spacing resulted in lower thresholds and maximum comfort levels and a maximum of 7 distinct stimulation sites (channels) within the cochlea.
The original intracochlear electrode array of the Clarion was pre-curled and inserted through the cochleostomy using a special insertion tool. This array was designed to conform to the contour of the cochlea. Until recently, newer versions of the Clarion used a silastic positioner that was inserted into the cochlea behind the intracochlear electrode array. The effect of the positioner was to move the electrode contacts closer to the cochlear modiolus (medial wall of the scala tympani) in order to reduce power consumption and to improve frequency selectivity. In October 2002, the Clarion electrode positioner was removed from the market due to concerns that its use may be associated with an increased risk of bacterial meningitis in cochlear implant recipients. Subsequently the FDA approved use of the Hi Focus Clarion electrode array without the positioner.
Although the Clarion electrode array always has had 16 contacts, the earliest device was limited to eight channels of stimulation because it used 8 independent output circuits. Each channel would be routed to a bipolar electrode pair or to the eight medial electrode contacts via monopolar electrode coupling. The newest version of the Clarion, the CII system, has 16 independent output circuits that can stimulate each of the 16 electrode contacts either non-simultaneously, simultaneously, or in various combinations.
The HiFocus electrode contacts are arranged longitudinally and can be activated in either monopolar, bipolar, or multipolar mode. Theoretically, because the Clarion allows simultaneous stimulation of multiple channels, it should also be possible to control the pattern of stimulation within the cochlea to provide up to 31 “virtual” channels. The CII system has 31 filter bands to enable experimentation with this form of stimulation but software is not yet available to allow implementation of this stimulation mode.
The Clarion device is packaged in a ceramic case that is set into a bed drilled into the temporal bone. The magnet is contained within the ceramic case and is neither removable nor MRI compatible. The Clarion cochlear implant system functions with both a body worn speech processor and an ear level device (BTE). The ear level processor is capable of implementing all of the processing strategies available with the body worn processor.
The Clarion is the only cochlear implant system capable of simultaneous stimulation of multiple electrodes within the cochlea. It also is the only device that can stimulate with analog waveforms. Like other commercially available cochlear implant systems, the Clarion offers a wide range of speech processing strategies. Clarion was the first commercially available implant system to implement CIS processing in 1991. The Clarion version of the CIS programming strategy was available with the very early versions of the device and was typically implemented using 8 channels of monopolar stimulation. With the original version of the Clarion system, pulse durations of 75 ms/phase were used with a stimulation rate of 833 Hz per channel. The newly introduced Clarion CII cochlear implant system allows for stimulation rates as high as 2,840 Hz per channel with the CIS programming strategy when all 16 channels are active and 5,980 Hz per channel when 8 channels are programmed.
The second strategy available with the original Clarion cochlear implant system was Compressed Analog stimulation (CA). In more recent versions, this strategy has been refined and is referred to as Simultaneous Analog Stimulation (SAS). This speech processing strategy is typically used with bipolar or enhanced bipolar electrode coupling. With SAS, the incoming speech signal is sampled and filtered into seven different frequency bands. The output of each frequency band is routed to an individual electrode or electrode pair. Compression is used to insure that the signal stays within the user’s dynamic range. With this strategy, biphasic current pulses are not used. Rather, the amount of current applied to a given electrode varies almost instantaneously according to the energy within that frequency band. When 7 channels are stimulated simultaneously in analog mode the overall stimulation rate is 91,000 samples per second. This processing strategy is designed to preserve the relative amplitude information in each channel and the temporal details of the waveforms.
One potential limitation of speech processing strategies that use simultaneous analog stimulation is that the simultaneous activation of multiple electrodes can result in deleterious channel interactions. Wilson, Lawson, Finley, and Wolford (1993) compared simultaneous analog stimulation to sequential pulsatile stimulation in persons who used monopolar coupling. They demonstrated that speech recognition scores were higher when the non-simultaneous processing strategy was used. Both the original CA strategy and the current SAS strategy are most successfully implemented using a bipolar or enhanced bipolar mode rather than monopolar stimulation. The probability of deleterious channel interactions is minimized with the Clarion device through the use of only 7–8 channels of stimulation and closely spaced bipolar coupling.
In 1999, a variation on these two basic speech-processing strategies was introduced. This variation is the Paired Pulsatile Sampler (PPS). PPS is similar to CIS except that instead of each electrode in the array being stimulated sequentially—without simultaneous stimulation—with PPS pairs of electrodes that are widely spaced across the array are stimulated simultaneously. The advantage of PPS over CIS is that it is possible to achieve stimulation rates that are twice as fast as those used with a fully sequential or non-simultaneous CIS strategy. Increasing the stimulation rate has the effect of increasing the amount of information about the acoustic signal that is transmitted per unit of time. By simultaneously stimulating electrode pairs that are spaced far apart, the effects of channel interaction can be minimized.
Recently, Advanced Bionics has developed and is testing a High Resolution processing strategy for the CII cochlear implant. Although this new programming software has been FDA approved for both adults and children, general release of this software is still pending. When the High Resolution mode is implemented with the CII device, the Clarion system should be capable of reaching the fastest stimulation rates of any of the commercially available cochlear implant systems.
Like all commercially available cochlear implant systems, the Clarion speech processor is flexible, allowing the user to listen to a range of different processing strategies. The current Clarion body worn speech processor, the Platinum Sound Processor, is smaller than the Nucleus body worn processors. Additionally there are two versions of the ear level system, one for use with the earlier Clarion implant system, called the Platinum BTE, and one for use with the new implant system, called the Clarion CII BTE. Because the Clarion Platinum BTE has high power demands, the retrofitted Platinum BTE with a custom designed rechargeable battery has a limited battery life averaging 5 to 6 hours. The Clarion CII BTE is somewhat more efficient, using custom designed rechargeable batteries that average anywhere from 8 to 11 hours of use, depending on the individual’s processing strategy and required stimulation levels. Rechargeable batteries typically must be replaced periodically.
Advanced Bionics was the first cochlear implant system equipped with telemetry capabilities for monitoring electrode integrity and compliance voltages. With their most recent implant, the Clarion CI-II, it is also possible to use the intracochlear electrodes to record electrically evoked auditory potentials. This system offers the same monitoring capabilities as the Nucleus Neural Response Telemetry (NRT) system. Clarion’s version is known as Neural Response Imaging (NRI). The software for measuring neural responses from within the cochlea is FDA approved and should be released in the near future.
The MED-EL Cochlear Implant
The third FDA-approved cochlear implant system available in the United States today is the MED-EL Combi 40+ cochlear implant manufactured by the Medical Electronics Corporation. This device has 12 electrode pairs that are inserted deep into the apical regions of the cochlea. The standard array is the longest of all three cochlear implant systems and extends 26.4 mm into the cochlea (2.4 mm contact separation) or two complete turns. The Combi 40+ electrode is a thin, soft, flexible straight array that is threaded into the scala tympani of the cochlea through a cochleostomy and relies on the contour of the cochlear ducts to achieve the spiral form. Like the Clarion device, the internal electronics and the internal magnet of the MED-EL implant are housed in a ceramic case. The MED-EL device has FDA approval for use in MRI machines up to 0.2 Tesla. In Europe, it is used with MRI machines of 1.0 and 1.5 Tesla. A special form available from MED-EL must be submitted to the radiologist before scanning. All safety measures and limitations for scanning are provided on the form. In addition, MED-EL will provide direct information for radiologists if they are contacted. The MED-EL speech processor has up to 9 memories available to hold a range of programs. Originally, a body-worn processor, the CIS-PRO+ was provided with the MedEl device. In 1998, a behind the ear processor, the Tempo+ was introduced. Current recipients standardly are provided with the Tempo+ behind the ear speech processor, even very young children. The Tempo+ offers a variety of wearing options including the option to use a battery pack that is attached to the processor via a cord allowing it to be clipped to a collar, etc. The fact that the processor is tethered to the battery pack, which in turn can be securely mounted on clothing, can help with retention of the behind the ear processor when fitted to young children. The MED-EL Tempo+ offers the longest battery life of all the available behind-the-ear cochlear implant speech processors, with an average battery life of 50 hours.
The MED-EL device has the capacity to provide some of the most rapid stimulation rates of any of the cochlear implant systems currently available (1515 Hz/channel, 18180 Hz overall) using sequential pulsatile stimulation. Older versions of the MED-EL system offered the CIS speech processing strategy, implemented in a similar fashion to the Clarion speech processor. The current MED-EL speech processors offer two sequential stimulation processing strategies, CIS+ which uses a Hilbert transform for envelope detection, thereby eliminating problems with aliasing that may affect other speech processing systems, and n-of-m processing that is similar to ACE processing with the Nucleus device
Special Electrode Arrays
Several special electrode arrays have been designed for individuals who are not candidates for standard electrode arrays. This includes persons with obstructed cochleae (i.e., ossified cochleae) or other cochlear malformations and persons who no longer have an intact auditory nerve. For individuals with ossified or malformed cochleae, MED-EL offer a shorter version of the C40+ electrode array with more closely spaced electrode contacts known as the compressed array. In addition, both MED-EL and Cochlear Corporation offer a special split array for individuals with complete cochlear ossification. With split arrays, the surgeon makes two cochleostomies, one at the basal and one at the apical end of the cochlea. One branch of the split array is inserted into each. Both the compressed and the split electrode arrays are FDA-approved
Both Cochlear Corporation and MED-EL have developed a special electrode for combined electric and acoustic stimulation for use with individuals who have moderate amounts of low-frequency hearing. These devices, which are in the preliminary stages of investigation, are designed to preserve as much residual hearing as possible during implant electrode insertion; this requires special electrode insertion techniques.
Cochlear Corporation offers a special electrode array for people with Neurofibromatosis II. These people typically have surgery to remove an acoustic neuroma leaving them without an intact auditory nerve. Therefore, the electrode array is positioned on or near the cochlear nucleus rather than within the cochlea itself.
Device Selection
The device selected for an individual patient depends on several factors including the center at which the patient is followed, whether or not the device is in FDA clinical trials, and the preference of the surgeon and recipient. Some centers offer cochlear implant candidates a choice of devices from all three major manufacturers whereas other centers may offer only one or two different cochlear implant systems. When a particular device is in FDA clinical trials, availability is limited to individuals who meet the candidacy criteria for that clinical trials’ study. For example, some clinical trials protocols restrict implantation to people with no additional handicapping conditions. Typically, device selection is made by the patient in consultation with the surgeon. With current cochlear implant technology, cochlear implant outcomes are similar across devices from all three manufacturers. There is a wide range of patient outcomes within each group of individuals using a given device. For each device, some people obtain substantial auditory-only speech understanding whereas others use the input from their cochlear implant as an aid to speechreading.
COCHLEAR IMPLANT 