Chinese language is based on 4 (Mandarin) to 6 (Cantonese) lexical tones to differentiate the meaning of a word with the same phonetic segment. Pitch perception transcribing into tone differentiation is therefore important in speech perception of adults and children undergone auditory brainstem implant. Assessment materials are designed for gross discrimination in fundamental frequency and lexical tone identification as well as the ability to tonal speech production in children. In a cohort of 13 adult NF2 implantees, 78% may differentiate environmental sound and 60% were able to have closed set word identification with lipreading. Only one patient was able to use the device alone without lipreading. In a second cohort of children who were pre-lingually deaf undergone ABI with a follow-up of up to 5 years, all children who had normal intelligence and functioning implants were able to use the device and be integrated into an integrated normal school environment using auditory-verbal means of communication. The mean vowel and consonant identification scores of this group were 59% and 63% respectively. When comparing to cochlear implant controls who often had their performance plateaued within 3 years, the ABI children showed slow but progressive improvement of up to even 5 years of usage. The mean tone imitation score was 53% and the mean tone production score was 63% with some children achieving over 90% performance. An observation that young implantees may achieve tone perception and production earlier and with better performance was made and case illustrations of children with different outcomes were made in this chapter. The observations of the slow progress, etiology of deafness and co-morbidities were discussed with reference to the outcomes. Successful methods of habilitation of these children include focusing on pitch and tone-pairs discrimination, nonsense syllables and meaningful word identification, and bilingualism with the use of sign language.
15 Auditory Brainstem Implantation in Tone Language Speakers
Globally, the Chinese language or its dialects are the most commonly spoken first languages, with over 900 million native mandarin Chinese speakers (rank 1 in Wikipedia) and over 70 million Cantonese speakers (rank 17). The specific nature of the Chinese languages and other tone-based or tone languages allow speech and hearing scientists to explore the ability of different methods to rehabilitate the hearing impaired, from amplification to electrical stimulation of the cochlea. Difference in pitch patterns affects tone and these changes cannot be detected by lip reading alone. One of the challenges regarding the use of electrical stimulation to restore hearing is reduced ability with pitch perception, affecting notably the perception of tones in speech and music. Over the past 20 years, tests have been developed to assess the perception and production of tones in tone language speakers. We have shown that tones are generally recognized through pitch differentiation problem in younger cochlear implantees as well as better performing adults with difficulties, 31 and it is possible that the same problem may exist in auditory brainstem implant (ABI) users. In this chapter, we outline the characteristics, the assessment, and the outcome of ABI implantation in our series of Cantonese-speaking Chinese adults and children to compare with that available in the literature.
15.2 What is a Tone Language?
A tone language is a language in which the variation of the fundamental frequency contour of a syllable can change the meaning of the word. Mandarin and Cantonese are both lexical tone languages widely spoken by the Chinese population. While Mandarin is the official spoken language in China, Cantonese is the second-largest spoken language in terms of scope of use. Cantonese is widely spoken in the southern part of China including the Guangdong Province, Hong Kong, Macau, and other Asian countries such as Malaysia and Singapore, and the communities of residents in various parts of the world including Australia, United Kingdom, Canada, and the United States of America. It is estimated that 62 to 70 million people worldwide speak Cantonese. 20
The use of pitch configurations to distinguish one lexical item from another is one important feature distinguishing tone languages from Western languages such as English. Lexical tones are identified from one another by the fundamental frequency (F0) according to the F0 height and contour. 12 There are six and four tones in Cantonese and Mandarin, respectively. A change in tone within a phonemic segment contributes to a change in lexical meaning. 21 Fig. 15.1 and Fig. 15.2 illustrate the different fundamental frequency patterns of the Cantonese and Mandarin tones. Table 15.1 shows sets of examples.
Lexical tones in Cantonese
Lexical Tones in Mandarin
15.3 Speech Assessment Unique to Tone Languages
Due to the inherent different linguistic systems between tone and non-tone languages, it is essential to design specific clinical assessment tools gearing to the particular properties of the language system. For individuals with ABI, the progress in general is slower when compared with individuals with cochlear implants (CIs). Besides the stage of sound detection, areas of assessment need to be more detailed at the basic level of suprasegmental perception. In the validated standardized test of the Cantonese Basic Speech Perception Test (CBSPT), 17 test facets include assessing suprasegmental features of pitch and stress perception, which are of paramount importance in tonelanguages.
15.3.1 Assessment of Pitch Perception
In tone language, pitch perception assessment could further be divided into
Gross discrimination in fundamental frequency
Basic lexical tone identification
Lexical tone identification
Gross fundamental frequency discrimination testing is done by assessing the differentiation between a typical male speaker and female speaker producing the same sentence. The average fundamental frequency (F0) of a male and female is around 150 and 250 Hz, respectively. This is to test if one is able to make use of the great difference in pitch information for meaningful perception.
The next stage of basic lexical tone perception is to employ pairs of relatively easy-to-identify tone contrasts as test items. The earlier acquired tones and the tones with the greatest fundamental frequency pattern contrasts should be selected for discrimination in this basic test. In Cantonese, the high-level, high-rising, and low-falling tones are regarded as basic tones. 17 However, in Mandarin, research consistently shows that a high-falling tone is the earliest acquired tone and is the easiest to discriminate when paired with the other tones. 23
The next level of evaluating pitch information is assessing the full set of lexical tone identification in the language. In Cantonese, there are six lexical tones. All possible combinations yield a total of 15 tone pair contrasts. The Cantonese Tone Identification Test (CanTIT) 16 is a validated and standardized assessment tool for measuring tone perception ability of the Cantonese-speaking population. Normed scores from age 3 to adults are provided. Reference scores of hearing-impaired children with various degrees of hearing loss are also included. With similar test construction principles, Mandarin speakers employ the Mandarin Tone Identification Test 33 as the assessment tool to assess the six tone pair contrasts.
15.3.2 Assessment on Stress Perception
The role of stress perception in English is very different from the tone languages of Cantonese and Mandarin. Depending on whether the two syllables receive equal or unequal stress, disyllabic words in English could further be divided into spondee and trochee. For example, “football” is a spondee with equal stress on the two syllables whereas “happy” is a trochee where only the first syllable is stressed. Assessment on whether one could perceive the stress/unstressed pattern is important in English. Tone languages like Cantonese and Mandarin, however, do not have such a difference in stress pattern in disyllabic words. Both syllables receive equal stress. The rule of equal stress applies in all multisyllabic words, meaning every syllable in a word receives the same amount of stress. The more important aspect of stress in Cantonese and Mandarin, thus, lies in whether one could correctly identify the number of syllables so as to derive if the perceived word is a monosyllabic/disyllabic/trisyllabic or multisyllabic word. Assessment should include testing on identification of number of syllables.
15.3.3 Assessment of Segmental Aspects Involving Words
When moving up to the more advanced levels of speech perception involving use of meaningful words, linguistic characteristics including the phonological system, the grammar, and the syntactic structure must be considered. Frequency of occurrence of words used and sentence type should also be considered.
15.3.4 Assessment of Speech Production
Speech sound production assessment in tone languages should also include tone production accuracy in addition to segmental phonology of vowel and consonant production. Tone production skills maybe assessed at word level, phrase level, sentence level, or discourse level. The key is that all lexical tones in the concerned tone language should be included in the assessment items.
15.4 ABI in NF2 Chinese Speakers
The incidence and prevalence of neurofibromatosis type 2 (NF2)among Chinese has not been previously studied although it is estimated to be much lower than the quoted UK rate of 1:33,000. 11 The ABI poses further challenges for the tone language speakers. First, postlingually deafened adults who need an ABI have bilateral vestibular schwannomas without their auditory nerves preserved as well as multiple comorbidities. Here we reviewed a series of Chinese adult patients with NF2 who underwent ABI surgery in the Chinese University of Hong Kong.
All patients underwent thorough audiovestibular, lipreading/oral communicative skills, and full radiological assessments before surgery. The implant used in all patients was the Nucleus ABI24 M and later ABI-5 series auditory brainstem device (Cochlear, Sydney, Australia) and one subject had a Combi-40 ABI (MED-EL, Innsbruck, Austria). The Nucleus SPEAK spectral peak speech coding strategy in monopolar mode was used in all patients except the patient using the MED-EL implant.
The extended trans labyrinthine approach for tumor removal and ABI insertion was used in all adults except for one revision surgery in which the retrosigmoid approach was employed. We use the facial nerve, cochlear nerve and other lower cranial nerves (glossopharyngeal and vagus nerves), and especially the choroid plexus as landmarks. The implant is inserted under direct vision and secured in position with a combination of adipose tissue, muscle grafts, and surgical tissue glue, before wound closure. In case the retrosigmoid approach was employed, direct access to the lateral recess was obtained posteriorly. Evoked auditory brainstem response (EABR) is used to confirm proper positioning of the implant electrode array during surgery. In two patients with functional hearing in the contralateral ear, ABI implanted during first-side surgery (as a “sleeper” device) was not switched on in one subject until hearing was lost in the contralateral ear, either from tumor progression or surgery. At the first switch-on session, threshold and maximum comfort levels were set, together with loudness balancing and pitch ranking of the electrodes. Electrodes were activated sequentially, and the intensity gradually increased to obtain thresholds while avoiding adverse effects under cardiovascular monitoring and with standby emergency resuscitation equipment. Nonauditory stimulated electrodes were disabled, and electrode array tonotopy was deduced by paired comparisons of pitch.
Remapping and re-evaluation were performed regularly over the first year. Speech and sound assessment measures, wherever possible, were performed and recorded at 6 months, 1 year, and 2 years, postoperatively. Open-set speech perception tests were conducted in Cantonese (Hong Kong Speech Perception Test Manual) except for the only Mandarin speaker in quiet listening conditions using live voice at normal conversational levels in three conditions: (1) use of ABI alone without the help of lipreading mode (A-mode), (2) lipreading alone without the use of ABI (visual or V-mode), and (3) lipreading together with use of ABI (audiovisual or AV-mode).
From 1997 to 2016, there were a total of 13 adult patients (12Cantonese-speakingand 1 Mandarin-speaking patients) with NF2 who underwent surgery for removal of vestibular schwannoma and surgical placement of ABIs. Ten of these patients received ABIs after surgical excision of either their first-side vestibular schwannoma or second-side vestibular schwannoma. Two patients declined to have ABI implanted. One patient had consented for ABI but the surgery was abandoned intraoperatively because distorted anatomy prohibited access to the lateral recess. Two patients had their first-side vestibular schwannoma surgery done elsewhere. The bilateral implantee had implant migration and subsequent revision. The mean age of patients at presentation to our department was 28 years, ranging from 14 to 51 years. Ten (77%) patients were female. Most of the patients in the series were reported in Thong et al 28 and are being re-cited here.
The mean size of vestibular schwannoma at presentation for this group of patients was 27 mm (range, 15–41 mm) and the mean size at surgery was 30 mm (range, 15–55 mm).The mean age of patients at the time of ABI implantation was 25 years (range, 16–54 years). In all patients except one, ABI implantation was performed in conjunction with removal of the tumor. Only one patient had a second ABI at the time of removal of second-side tumor because of absence of responses on EABR testing postoperatively after the first surgery.
Device activations were performed between 5 and 9 weeks postoperatively for ABIs implanted during second-side surgery. In the two patients who had ABI implantation at first-side tumor removal, device activation was much later at 4 months (after removal of second-side tumor) and 23 months (after deterioration of hearing in the contralateral ear that had hearing-preservation tumor removal surgery done). The average number of active electrodes was 14 (range, 9–18). The patient who had bilateral sequential ABI implantation did not have EABR response on the second side postoperatively as well, although responses were present intraoperatively. A revision surgery had shown migration of the electrode, yet repositioning resulted in minimal auditory perception. Other surgical complications included: two (25%) patients had permanent facial palsy (House-Brackmann Grades II and III) and one (13%) patient had temporary facial palsy that resolved by 1 year. One patient had cerebrospinal fluid (CSF) leak postoperatively that was managed conservatively with bed rest and insertion of lumbar drain. One patient developed delayed postauricular wound infection at 2 months and this resolved with intravenous antibiotics and wound debridement and did not require implant removal.
15.4.1 Speech and Hearing Outcomes in NF2 Adults
Sound and speech assessment measures were performed for audiological outcomes at 6 months, 1 year, and 2 years, postoperatively. Including the patient who had two sequential ABIs that did not have EABR postoperatively, environmental sounds could be differentiated by six (75%) patients (67% of implants) after 6 months of ABI use (mean score 46% [range, 28–60%]). One patient (Patient 5) stopped using the ABI after 6 months for reasons as mentioned later and therefore at the 1-year and 2-year postoperative assessments, only seven (78%) patients (56% of implants) continued to be able to differentiate between environmental sounds (1-year mean score 57% [range, 36–76%]; 2-year mean score 48% [range, 24–76%]). Closed-set word identification was possible in six (60%) patients (67% of implants) at 6 months (mean score 39% [range, 12–72%]), 1 year (mean score 68% [range, 48–92%]), and 2years postoperatively (mean score 62% [range, 28–100%]). One patient demonstrated open-set sentence recognition in quiet in A-mode (use of ABI alone). However, sentence recognition was possible in AV-mode (lipreading together with use of ABI) in six (60%) patients (67% of implants) at 6 months postoperatively (mean score 49% [range, 27–67%]). At 6 months postoperatively, two patients had improvement with scores of 40 and 52% in AV-mode compared with V mode, whereas two patients had no improvement in AV-mode. After 1 and 2 years of ABI use, five (63%) patients (56% of implants) could recognize sentence in AV-mode (1-year mean score 31% [range, 12–79%]); 2-year mean score 35% [range, 12–67%]) and all scores were better than in V-mode (average improvement in scores of 25% at 1 year and 28% at 2 years).
At 2 years postoperatively, only five (50%) patients remained ABI users. In two patients, initial encouraging results with the ABI deteriorated over time and they stopped using the ABI. Patient 7 had the best results with the ABI but she unfortunately developed loud and persistent tinnitus after 2 years and this negatively affected her use of the ABI. In another patient, deterioration of vision soon after ABI implantation affected her ability to lip-read and the ABI was thereafter deemed unhelpful.
ABI users reported that there was improved environmental sound awareness and they were able to differentiate between everyday sounds such as the telephone and television. In one patient, the ABI was found to be helpful in improving understanding of speech and that there was less dependence on lipreading. In others, it was thought that the ABI helped lipreading and allowed them to communicate normally without writing.
In patients who were nonusers, the main complaint was that the ABI was too noisy, especially outdoors, and that it was difficult to tolerate for long periods. It was also thought that actual sound sensations were weak and of poor quality.
15.5 ABI in Non-NF2 Children with Tone Languages
It is an important step towards understanding the question of ABI in detecting tone changes in speech through the study of its development in children who are prelingually deafened. The introduction of ABIs in other non tumor inner ear diseases and deformities such as cochlear aplasia, labyrinthine aplasia (Michel deformity), and cochlear nerve aplasia9 allows such opportunity for researchers to look into the issue. Following the pioneering work of Colletti and others, the indication for auditory brainstem implantation was extended to the treatment of pediatric prelingual deafness.5,6
The following are the data and findings from a published series of Cantonese-speaking children who underwent auditory brainstem implantations at our Center. 27 The audiological and tone language developmental outcomes of pediatric patients with ABIs followed-up for 1 to 5 years are compared with age-matched outcomes of a parallel group of CI users. This remains the only series presenting the outcomes of pediatric ABI users in a tone language setting.
15.5.1 Patient Demographics
ABI was performed in 11 Cantonese-speaking and 2 Mandarin-speaking prelingually deaf children who either failed or had contraindications for cochlear implantation between January 2009 and February 2015 in our unit. The age at implantation ranged from 1.7 to 3.8 years (mean 2.7 years). There were eight males and three females. Etiologies included cochlear nerve deficiency (n=7) and severe cochlear malformations (n=2) as shown radiologically. In the cochlear nerve deficiency group, two of the seven had coexisting auditory neuropathy spectrum disorder (ANSD) features in their auditory brainstem response (ABR) study.
15.5.2 Preoperative Assessment
All subjects were initially identified by the Hong Kong universal newborn hearing screening program and were referred to our unit for further assessment. Children who met the criteria for cochlear implantation would proceed to CI surgery. Children who showed limited or no benefit from hearing aids, and who had abnormalities on imaging such as cochlear aplasia or severe malformations would be further assessed for ABI candidacy. The decision to proceed with cochlear implantation was jointly made with the parents in view of the lack of benefit from hearing aids in speech and language development.
15.5.3 Audiological Perception Outcomes
Each subject’s auditory perception ability was tested using the CBSPT. The following domains of auditory perception were tested:
Sound detection (seven-sound detection):The ability to detect the Ling’s seven sounds in a quiet environment
Suprasegmental (syllable identification):The ability to identify the number of syllables in a sound string
Vowel identification: The ability to identify a word with an appropriate vowel in an array which is different from the vowel only
Consonant identification: The ability to identify a word with an appropriate consonant in an array which is different from the consonant only
Raw scores of these domains were used to determine the subject’s speech perception category (SPC) from 0 to 7 (see Table 15.2). The CBSPT only covers the test scope of up to consonant perception. For patients scoring over 75% in the consonant identification, open-set word recognition and sentence recognition tests were used to assess higher levels of speech perception ability. In addition, tone imitation and production tests were used on this specific group of Cantonese-speaking children. In tone imitation tests we test their ability to imitate words with different Cantonese tones while in tone production tests we assess their ability to correctly produce words with different Cantonese tones. Both were performed by experienced speech and language pathologists.
Speech perception category
Minimal Sound Detection
10–20% Open-set Word Recognition
20–50% Open-set Word Recognition
> 50% Open-set Word Recognition
For comparison of CI outcomes with this group of ABI children, an age-matched group consisting of 17 children implanted between the ages of 1.1 and 3.1 years who had no significant developmental delay was selected. They were identified with severe to profound hearing loss with no indication of ANSD. Imaging including computed tomography (CT) and magnetic resonance imaging (MRI) suggested no significant abnormalities in the cochlear and internal auditory meatus.
Cochlear implantation was performed on 7 of the 11 Cantonese-speaking ABI subjects. One child had a hypoplastic middle ear with an unsuccessful electrode placement attempt. The remaining six subjects, who underwent uneventful surgery with satisfactory electrode position, did not show consistent sound detection nor benefit in speech and language development after 1 year of CI use with regular auditory programming attempts and speech training. Cochlear implantation was not considered for the other 4 of the 11 subjects as they all had severe cochlear malformations, or cochlear nerve deficiency. For the two subjects with cochlear nerve deficiency, the decision of not to implant them with a CI was made in conjunction with parental preference. Of the two Mandarin speakers, one had a prior intraoperative EABR showing no response in another center and one had complete cochlear aplasia. Both of them had no CI before ABI surgeries.