People who are deafened by bilateral acoustic tumors are not candidates for a cochlear implant because tumor removal often severs the auditory nerve. The auditory brainstem implant (ABI) is designed for those patients. It is intended to bypass the auditory nerve and electrically stimulate the human cochlear nucleus.

Neurofibromatosis type 2 (NF2) is a genetic disorder that causes multiple tumors of the cranial nerves and spinal cord, among other symptoms. The gene causing NF2 has been located on chromosome 22 (Rouleau et al., 1993; Trofatter et al., 1993). The defining symptom of the disease is bilateral tumors originating on the Schwann cells of the vestibular branch of nerve VIII (vestibular schwannomas). These tumors are life-threatening, and their removal usually produces bilateral deafness. Patients with NF2 cannot benefit from a cochlear implant because they have no auditory nerve that can be stimulated from an intracochlear electrode. In 1979 William House and William Hitselberger attempted to provide auditory sensations for an NF2 patient by placing a single pair of electrodes in the cochlear nucleus following tumor removal (Edgerton, House, and Hitzelberger, 1984; Eisenberg, et al., 1987). The success of that early attempt led to the development of a more sophisticated multichannel ABI device (Brackmann et al., 1993; Shannon et al., 1993; Otto et al., 1998, 2002). The first commercial multichannel ABI was developed in a collaborative effort between the House Ear Institute, Cochlear Corporation, and the Huntington Medical Research Institutes. The multichannel ABI was approved by the U.S. Food and Drug Administration in October 2000.

Several commercial ABI devices are available, all basically similar to the original device. ABIs are virtually identical in design to cochlear implants except for the electrode assembly (Fig. 1). The electrode assembly is a flat, paddle-like structure with platinum electrical contacts along one side. The overall size of the assembly is generally 2–3 mm × 8 mm and is designed to fit within the lateral recess of the IV ventricle. The electrical contacts are 0.5–1.0 mm in diameter, which is sufficient for keeping the electrical charge density at the stimulated neurons within safe limits (Shannon, 1992). All ABI devices have an external speech processor unit that contains a microphone to pick up the acoustic sound, a signal processor to convert the acoustic sound to electrical signals, and a transmitter/ receiver to send the signals across the skin to the implanted portion of the device. The implanted unit decodes the received signal and produces controlled electrical stimulation of the electrodes. The most widely used ABI electrode array (Fig. 2; manufactured by Cochlear Corp.) consists of 21 platinum disk contacts, each 700 µm in diameter. The contacts are placed in three rows along a Silastic rubber carrier that is 8 mm × 2.5 mm (Fig. 3).

Figure 1..  

Overview of the ABI device and placement.

Figure 2..  

Implantable portion of the ABI, showing the receiver/ stimulator and electrode array.

Figure 3..  

Close-up view of the 21-electrode array, which consists of 21 platinum disks mounted on a Silastic substrate. The fabric mesh backing is intended to encourage fibrous ingrowth to fix the electrode array in position.

Present ABI electrodes are designed to be placed within the lateral recess of the IV ventricle. Anatomical studies (Moore and Osen, 1979) of the human cochlear nucleus complex and imaging studies of early ABI patients demonstrated that this location produced the most effective auditory results and the fewest nonauditory side effects (Shannon et al., 1997; Otto et al., 1998, 2002). Electrical stimulation in the human brainstem can potentially produce activation of many nonauditory structures (cranial nerves VII, IX, X, and cerebellum, for example). Fortunately, the human cochlear nucleus complex almost completely surrounds the opening of the lateral recess of the IV ventricle, and the levels of current delivered to the ABI are not large enough to activate other brainstem nuclei more than about 2 mm away (Shannon, 1989, 1992; Shannon et al., 1997).

Vestibular schwannomas can be visualized and removed via several surgical approaches, of which the retrosigmoid and translabyrinthine approaches are the most common. The translabyrinthine approach allows better visualization of the mouth of the lateral recess following tumor removal and thus better access for placement of the ABI electrode array (Brackmann et al., 1993).

Of the first 80 patients implanted with the multichannel ABI at the House Ear Institute, 86% received sufficient auditory sensations that they could use the ABI in daily life. For most ABI users the primary benefit is as an aid to lipreading, since only a few ABI patients can understand words with the ABI without lipreading. For most patients the present ABI device functions at a level similar to that of single-channel cochlear implants, even when many electrodes are used in the ABI speech processor. ABI patients are able to detect sound and are able to discriminate sounds based on coarse temporal properties (Shannon and Otto, 1990; Otto et al., 2002). On average, ABI patients receive a 30% improvement in speech understanding when the sound from the ABI is added to lipreading alone (Fig. 4). A few ABI patients (10%) achieve significant word and sentence recognition with the device, and a few (4/80) can actually converse in a limited fashion on the telephone.

Figure 4..  

Speech recognition results from the first 55 multichannel ABI patients. Lower part of each bar shows the percent correct recognition of simple sentence materials using lipreading alone. The upper part of each bar shows the improvement in recognition obtained when the ABI was added to lipreading.

Most ABI patients perceive variations in amplitude and temporal cues but receive little, if any, spectral cues. ABI patients are able to perceive changes in pitch with changes in pulse rate only up to about 150 Hz, which is about an octave lower than observed for cochlear implant listeners and for temporal pitch discrimination for normal-hearing listeners. Typically, the dynamic range of amplitude for the ABI is only 6 dB or less in terms of electrical current range. We estimate that ABI patients may be able to discriminate only 10 amplitude steps within their dynamic range, in contrast to 20–40 steps for cochlear implants and 200–250 steps for acoustic hearing. Some ABI patients have relatively small differences in pitch across their electrode array, while others show a large change in pitch. In general, patients who do better at speech recognition tend to have a larger pitch range across their electrode array, but not all patients who have a large pitch range have good speech recognition.

Most ABI patients have some electrodes that cause nonauditory side effects. Almost all of these nonauditory effects are benign and produce tingling sensations along the body on the side ipsilateral to the ABI. Nonauditory sensations are produced from stimulation of the cerebellar flocculus (which causes a sensation of eye jitter) and from antidromic activation of the cerebellar peduncle. In patients with an intact facial nerve on the implanted side there is a chance of activation of the facial nerve, causing facial tingling and even motor activation. In most cases, electrodes that produce nonauditory side effects are simply turned off and not stimulated.

One of the possible reasons for the limited success of the ABI is that the present electrode is placed on the surface of the cochlear nucleus. Unfortunately, the tonotopic axis of the human cochlear nucleus is orthogonal to the surface of the nucleus (Moore and Osen, 1979), and thus orthogonal to the axis of the electrode array as well. To obtain better access to the tonotopic dimension of the human cochlear nucleus requires penetrating electrodes (McCreery et al., 1998). A penetrating electrode ABI system is presently under development. Its initial trial is anticipated for 2002.