Mandarin Consonant Contrast Recognition Among ... - Semantic Scholar
Otology & Neurotology 34:471Y476 Ó 2013, Otology & Neurotology, Inc.
Mandarin Consonant Contrast Recognition Among Children With Cochlear Implants or Hearing Aids and Normal-Hearing Children *Qiaoyun Liu, †‡Ning Zhou, ‡Rebecca Berger, *Daniel Huang, and ‡Li Xu *Key Laboratory of Hearing and Speech Sciences, East China Normal University, Shanghai, People’s Republic of China; ÞKresge Hearing Research Institute, University of Michigan, Ann Arbor, Michigan; and þSchool of Rehabilitation and Communication Sciences, Ohio University, Athens, Ohio, U.S.A.
Hypothesis: The purpose of the present study was to investigate the consonant recognition of Mandarin-speaking children with cochlear implants (CIs) and hearing aids (HAs) and to determine if they reach a level of consonant recognition similar to that of normal-hearing (NH) children. Background: Little information is available in the literature regarding the consonant perception abilities of prelingually deafened young children with either CIs or HAs. No studies have compared Mandarin-Chinese consonant contrast recognition in CI and HA children. Methods: Forty-one prelingually deafened children with CIs, 26 prelingually deafened children with HAs, and 30 NH children participated in this study. The 3 groups were matched for chronologic age (3Y5 yr). The hearing-impaired groups were matched for age at fitting of the devices, duration of device use, and aided hearing threshold. All subjects completed a computerized Mandarin consonant phonetic contrast perception test.
Results: CI and HA children scored, on average, approximately 8 percentage points below the mean NH group performance on the consonant contrast recognition. Approximately 40% of the CI and HA children had not reached a performance level of the NH group. No significant differences in the consonant recognition scores were found between the CI and HA groups. Age of implantation was correlated with consonant contrast recognition in the CI group. Conclusion: When age at fitting of the devices, duration of device use, and aided thresholds are matched at the group level, consonant recognition is similar between the CI and HA children after 2 years of device use. Early implantation tends to yield better consonant contrast recognition in the young children with CIs. However, a large amount of variance in performance was not accounted for by the demographic variables studied. Key Words: Cochlear implantsVConsonant recognitionV Mandarin ChineseVPediatric. Otol Neurotol 34:471Y476, 2013.
The population of prelingually deafened children who have received cochlear implants (CIs) has increased rapidly in China. As of November 2012, more than 15,000 profoundly deafened children have received multichannel CIs in China. The number of children with CIs enrolled in mainstream preschools and elementary schools in China is also rapidly increasing. It is essential for CI children to achieve a high level of listening ability in order to succeed in mainstream schools in China be-
cause assistive services, such as note takers or sign language interpreters, are not readily available. Although there is a wealth of data to indicate whether CI children can reach a level of listening ability that is within the normal range for the English-speaking population, there is very little data for the Chinese-speaking CI users. There is plenty of evidence showing that, for prelingually deafened children, early implantation is important for developing normal or close-to-normal speech and language skills (e.g., 1Y6). Several research laboratories have examined speech perception ability and its relationship with age at implantation in Mandarin-Chinese speaking children with CIs. For example, Zhu et al. (7) tested the ability of Mandarin speaking CI users to understand disyllable words and sentences. On average, prelingually deafened children with CIs achieved approximately 80% correct for disyllabic word and sentence recognition. Age at implantation predicted the recognition performance for both disyllabic words and sentences. Wang et al. (8)
Address correspondence and reprint requests to Li Xu, M.D., Ph.D., School of Rehabilitation and Communication Sciences, Ohio University, Athens, OH 45701; E-mail: [email protected] This study was supported by the Open Fund of Key Laboratory of Speech and Hearing Sciences, East China Normal University, Ministry of Education, Shanghai, China. Declaration of Interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this paper. Supplemental digital content is available in the text.
471
Copyright © 2013 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
472
Q. LIU ET AL.
examined how speech perception in Mandarin-speaking CI children compare with their normal-hearing (NH) peers. Thirty normal-hearing children and 36 CI children (4Y11.6 yr old) were given a newly developed version of the Monosyllabic Lexical Neighborhood Test and the Disyllabic Lexical Neighborhood Test in Mandarin Chinese to determine the effects of linguistic and cognitive demands on speech perception performance. Monosyllabic and disyllabic words with frequencies above the median and lexical densities below the median were labeled as ‘‘easy’’ words, and the opposite were labeled as ‘‘hard’’ words. Lexical effects were demonstrated for disyllabic word recognition. More than half of the CI children fell into T1 standard deviation (SD) of the mean word recognition of the NH group. More recently, Liu et al. (9,10) reported results obtained from 230 children with CIs using their own version of the Standard-Chinese Lexical Neighborhood Test. Lexical effects similar to those found in the NH children were shown for both monosyllabic and disyllabic words in the CI group. The performance of the CI group was poorer than that reported in Wang et al. (8), with only a small portion of the CI children falling within the normal range. An association between age at implantation and open-set word recognition performance was also demonstrated (10). Few studies have examined consonant recognition abilities in Chinese-speaking children with CIs. Wu and Yang (11) evaluated how speech perception of CI children continues to improve after implantation. In that study, 16 Mandarin-speaking CI children between the ages of 1.6 and 6 years were given the Mandarin Auditory Perception Test Battery every 6 months for 2 years after implantation. Between 12 and 24 months after implantation, spondee, vowel, and consonant recognition improved significantly. Interestingly, although spondee, vowel, phrase, and sentence recognition showed a negative correlation with age at implantation, consonant recognition was not correlated with age at implantation. Lin and Peng (12) found that the average scores for the initial consonant discrimination in a group of 30 Mandarin-speaking children (6Y12.5 yr old) with CIs was 76.7% correct. More recently, Zheng et al. (13) showed that 4 of 25 CI children between 3 and 16.9 years old were able to achieve consonant recognition performance close to that of the NH children of similar hearing ages when tested with a closed-set Mandarin Early Speech Perception Test. A number of studies have attempted to compare the English speech and language skills of children with CIs to children fitted with conventional hearing aids (HAs) (4,14Y19). Comparison between these groups is often difficult because of the differences in factors such as age, length of device use, amount of hearing loss, and educational setting. However, the consensus in recent literature is that contemporary multichannel CIs provide prelingually deafened children with enhanced speech and language development compared with conventional HAs. In particular, Connor et al. (4) demonstrated that children who received CIs before the age of 7 exhibited a steeper slope in the development of consonant production accu-
racy when compared with the predicted trajectory for the HA children. However, no study has compared the speech and language development between the CI and HA users in the Mandarin Chinese population. In a recent study, Law and So (20) examined both speech production and speech perception in Cantonese-speaking CI and HA users. Seven HA users and 7 CI users, matched for mean chronologic age (range, 5.1Y6.4 yr) and years of speech therapy (range, 1.25Y3.25 yr), were given the Cantonese Segmental Phonology Test and the Cantonese Lexical Comprehension Test. The CI users demonstrated significantly better consonant production skills than the HA users but not better vowel production. There was no significant difference in the speech perception tasks between the 2 groups. Results indicated that CIs seemed to promote the development of consonant production compared with HAs. There is an urgent need to study the many facets of speech and language development in the rapidly increasing population of Chinese-speaking children with CIs and HAs. The present study focuses on consonant contrast perception in the 2 groups. Consonant landmarks have been found to contribute greatly to speech recognition in noise for NH listeners and CI users. We examined the differences in consonant contrast recognition abilities between Mandarin-speaking CI and HA children with group-matched chronologic age, duration of device use, age at implantation or HA fitting, and aided soundfield pure-tone thresholds. We determined the proportion of CI and HA users who could reach a level of consonant contrast recognition that is equivalent to the NH controls. A systematic comparison of the consonant confusion patterns in the HA and CI users is not available. In acoustic hearing, consonant confusion by hearing-impaired adult listeners might be influenced by high-frequency hearing loss even when audibility is compensated (21), whereas the voicing features of consonants may be well received in listeners with HAs. In electrical hearing, it seems that the frequency range per se is not the issue; instead, performance may be affected by frequency resolution as well as distortion because of mismatch in the electrode position and the frequency allocation (22). We hypothesize that because of the inherent differences between electrical stimulation with a CI and acoustic hearing with amplification, the 2 groups would be better at perceiving different consonant features. Specifically, the voicing feature may be better transmitted acoustically than electrically because of the loss of pitch information in the envelope-based speech coding in CI devices. On the other hand, highfrequency fricatives may be better recognized via a CI because amplification of the high frequencies could distort the speech cues. Finally, we examine the effects of various demographic variables as potential predictors for consonant contrast recognition in both HA and CI children. MATERIALS AND METHODS Subjects Forty-one prelingually deafened children with CIs (24 boys and 17 girls), 26 prelingually deafened children with HAs
Otology & Neurotology, Vol. 34, No. 3, 2013
Copyright © 2013 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
CONSONANT CONTRAST RECOGNITION TABLE 1.
CI (n = 41) HA (n = 26)
Range Mean T SD Range Mean T SD
473
Demographic information for the cochlear implant and hearing aid groups Chronologic age (m)
Duration of device use (m)
Age at implantation or at HA fitting (m)
Aided sound-field PTAa (dB HL)
29Y60 40.2 T 9.1 31Y60 44.5 T 9.3
6Y44 16.3 T 8.2 6Y37 20.4 T 9.1
10Y50 24.0 T 8.9 5Y47 24.2 T 10.3
23Y52 38.6 T 5.5 33Y54 42.7 T 4.8
CI indicates cochlear implant; HA, hearing aid; PTA, pure-tone average; SD, standard deviation. a PTA is the average threshold of 500, 1,000, 2,000, and 4,000 Hz tested with warble tones.
(16 boys and 10 girls), and 30 NH children (15 boys and 15 girls) were recruited from a rehabilitation center and a kindergarten in Shanghai, China. All of the participants are native Mandarin Chinese speakers. The ages of the CI children ranged from 29 to 60 months (mean, 40.2 mo) and amount of implant experience ranged from 6 to 44 months (mean, 16.3 mo). The age at implantation ranged from 10 to 50 months (mean, 24.0 mo), and the mean auditory threshold with the CI was 38.6 dB HL (SD, 5.5 dB). These children were implanted with Advanced Bionics, Nucleus, and Med-El devices. The ages of the HA children ranged from 31 to 60 months (mean, 44.5 mo). The amount of hearing aid experience ranged from 6 to 37 months (mean, 20.4 mo), and the age at hearing aid fitting ranged from 5 to 47 months (mean, 24.2 mo). The mean auditory threshold while using the hearing aids was 42.7 dB HL (SD, 4.8). All HA children were fitted with digital hearing aids. Table 1 summarizes the demographic information of the 2 hearing-impaired groups. The NH children were between 36 and 60 months old. The 3 groups were matched for the mean chronologic age (3Y5 yr old), and the 2 hearing-impaired groups were matched for the mean age at fitting of devices, duration of device use, and aided hearing threshold. The use of human subjects was reviewed and approved by the institutional review board of East China Normal University.
the positions of the 2 words were randomized each time. Familiarization of the test procedure was provided for each child before the test. Percent correct scores for each subcategory as well as the overall recognition accuracy were determined for each child. Percent correct scores were arcsine-transformed to stabilize the variance for the analysis of variance (ANOVA) and other statistical analysis (24,25).
RESULTS Figure 1 shows the overall consonant recognition performance for the 3 subject groups. Scores for the 30 NH children ranged from 83.1% to 98.5% correct (chance = 50% correct) with a mean of 93.3% correct (SD, 3.5%). The HA and CI children as a group scored approximately 7 to 8 percentage points lower than the NH group. Perception scores for the 41 CI children ranged from 53.0% to 98.9% correct with a mean of 86.2% (SD, 10.9%). Scores for the 26 HA children ranged from 65.9% to 96.9% correct with a mean of 84.8% correct (SD, 10.1%). Note that the variation in performance in the hearing-impaired groups was larger than the NH group. A 1-way ANOVA
Mandarin Consonant Phonetic Contrast Perception Test The consonant contrast perception test involved word identification from a total of 87 Mandarin consonant contrasts. Based on the characteristics of Chinese phonetics, the 87 consonant contrasts were divided into the following 6 subcategories: 1) 2 pairs of fricative/nonfricative contrasts (e.g., /he/ versus /e/), 2) 27 pairs of voiceless/voiced contrasts (e.g., /ma/ versus /ba/), 3) 8 pairs of aspirate/nonaspirate contrasts (e.g., /bao/ versus /pao/), 4) 29 pairs of same place/different manner contrasts (e.g., /bei/ versus /fei/), 5) 18 pairs of same manner/different place contrasts (e.g., /bao/ versus /dao/), and 6) 3 pairs of retroflex/nonretroflex contrasts (e.g., /zi/ versus /zhi/). All the pairs contained simple monosyllabic Mandarin Chinese words that children in this age group are familiar with. The vowels and their carrying tones of the 2 words in a contrast were the same. The speech materials were recorded from a female speaker (whose average voice F0 was 355 Hz) using a 44,100-Hz sampling rate and 16-bit resolution. A word list for the Mandarin Consonant Phonetic Contrast Perception Test is provided in the Supplemental Digital Content (http://links.lww.com/MAO/A143). The subjects participated in a computerized consonant contrast test. The test used a 2-alternative forced-choice paradigm. A consonant contrast token was played via a loudspeaker at a level of 70 dB SPL, while 2 pictures representing the consonant contrast were shown on a computer screen. The subject was asked to choose the 1 photograph that represented the word that he or she had heard. Each contrast pair was tested 3 times, and
FIG. 1. Overall consonant contrast recognition scores for the NH, CI, and HA groups. Each symbol represents 1 subject. Each box plot has lines at the lower quartile, median, and upper quartile values. The vertical whiskers show the extent of the rest of the data (with the exception of 1 data point in the NH group and 1 in the CI group that were considered to be outliers). The bracket at the top represents statistical significant differences between the 2 groups indicated. Otology & Neurotology, Vol. 34, No. 3, 2013
Copyright © 2013 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
474
Q. LIU ET AL.
FIG. 2. The mean consonant contrast recognition scores of the 6 subcategories for the NH, CI, and HA groups. The error bar represents the standard deviation. The bracket at the top represents statistical significant differences between the 2 groups indicated.
showed that the mean performance was significantly different among the 3 groups ( p G 0.001). Post hoc pairwise comparisons indicated that the performance of both hearing-impaired groups were significantly different from the NH group ( p G 0.05) but not significantly different from each other ( p 9 0.05). The brackets plotted in Figure 1 illustrate the significant statistical results of these comparisons ( p G 0.05). The overall consonant contrast recognition performance of 18 (43.9%) and 24 (58.5%) CI children fell within T1 SD and T 2 SDs of that of the NH children, respectively. Similarly, the overall consonant contrast recognition performance of 11 (42.3%) and 15 (57.7%) HA children fell within T1 SD and T2 SDs of that of the NH children, respectively. Figure 2 plots the group mean and SD of the consonant contrast recognition scores for the 6 subcategories. A 2-way ANOVA with subject group and consonant contrast as 2 main factors was performed to test whether the mean performance was different among the 3 subject groups and among the 6 subcategories. Results of the 2-way ANOVA revealed statistical significance for both group and subcategory factors ( p G 0.001). Post hoc pairwise comparisons were then performed to identify group pairs that showed significant differences for each consonant contrast. The brackets plotted in Figure 2 illustrate the significant statistical results of these comparisons ( p G 0.05). The CI and HA groups performed equally well for each of the 6 consonant contrasts. The performance of the NH group was significantly higher than that of the HA group in 3 of the 6 subcategories: fricative/nonfricative, same place/different manner, and retroflex/nonretroflex contrasts. The performance of the NH group was also significantly higher than that of the CI group in 3 of the 6 subcategories: same place/ different manner, same manner/different place, and retroflex/nonretroflex contrasts. No statistically significant differences were found in the voiceless/voiced and aspirate/nonaspirate contrasts between the NH and hearing-
impaired groups. The largest performance gap between the hearing-impaired and the NH groups was for the retroflex/ nonretroflex contrast. A correlational analysis was performed to determine the contribution of various demographic variables for the variance in the performance. Demographic variables for the CI subjects included age at implantation, duration of implant use, chronologic age, and acoustic hearing threshold with the implant. Age at implantation was found to be the only significant predictor. Figure 3 shows the correlation between age at implantation and the overall consonant contrast recognition scores. The correlation coefficient (r) of the linear fit was j0.343 ( p = 0.028). After we removed the lowest performer as an outlier
FIG. 3. Scatter plot of the overall consonant contrast recognition scores as a function of age at implantation for CI children. Each symbol represents 1 subject. The solid line represents the linear fit of the data.
Otology & Neurotology, Vol. 34, No. 3, 2013
Copyright © 2013 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
CONSONANT CONTRAST RECOGNITION (defined as a data point that is 93 SDs away from the mean of the group) (Fig. 1), the correlation strength (r) improved to j0.478 ( p = 0.0013). Demographic variables for the HA subjects included age at fitting, duration of hearing aid use, chronologic age, and aided threshold. No significant correlation was found between any of these variables and the performance.
DISCUSSION The consonant contrast recognition test was a fairly easy task for the NH children between 3 and 5 years of age. They scored on average higher than 90% correct on overall consonant contrast recognition and on recognition of all 6 consonant contrast subcategories (Figs. 1 and 2). The hearing-impaired groups (i.e., CI and HA groups) showed variability in performance, with scores ranging from slightly above chance (i.e., 50% correct) to nearly perfect. The overall recognition performance was 86.2% and 84.8% correct, respectively, for the CI and HA groups (Fig. 1). The performance for the CI group was approximately 10 percentage points higher than that reported by Lin and Peng (2003). The age at implantation in their CI children ranged from 27 to 123 months (mean, 68 mo), much older than that of our CI group (mean, 24 mo). This may explain the differences in the 2 studies (see below for discussion on age at implantation and consonant recognition performance). The overall consonant contrast recognition performance of approximately 60% of subjects in the hearing-impaired groups fell within T2 SDs of the performance of the NH children. In other words, approximately 40% of the CI and HA children performed greater than 2 SDs below the mean performance of the NH children who were of similar chronologic age. Note, however, that the hearing-impaired children started using CIs or HAs at an average age of 2 years (Table 1). Thus, their hearing age was, on average, 2 years younger than the hearing ages of the NH group. It is encouraging to see that more than half of the CI and HA children fell within 2 SDs of the NH controls whose hearing age was, on average, 2 years older. Will the other 40% of the hearing-impaired children catch up with more experience with their devices? Such a question can be answered with a longitudinal study of the same subjects. The lack of correlation between the duration of device use and the consonant contrast recognition performance, however, indicates that hearing age may not predict consonant recognition performance. Factors, such as chronologic age and maturity of cognitive function, are likely to interact with hearing age and may contribute to the performance with hearing devices. The results of the present study show that, when chronologic age, duration of device use, age at implantation or HA fitting, and aided sound-field pure-tone thresholds were matched for the group mean, there was no significant difference in the consonant recognition scores between the CI and HA groups (Figs. 1 and 2). These results are somewhat surprising, given that several
475
studies have shown that CI children demonstrate performance superior to HA children on many speech and language tasks (4,19). Tomblin et al. (19) demonstrated that more than 50% of the CI children, by 2 years after implantation, exceeded the 95% prediction interval of the HA children in English expressive language skills. This contrasted with only approximately 28% of the CI children who exceeded such a level at 1 year postimplantation. Connor et al. (4) also showed that the English consonant production accuracy of the CI and HA children diverged further as duration of device use became longer (1Y5 yr). In the present study, the hearing-impaired children had used their devices for an average of 2 years. No clear difference in Mandarin consonant contrast recognition was evident. Further research with a longer follow-up period will be necessary to determine whether these 2 groups will diverge in consonant contrast recognition. The 2 contrasts, fricatives/nonfricatives and retroflex and nonretroflex contrasts, tended to cause more confusion for the HA children than the CI children, although the statistical analysis did not reveal significant difference (Fig. 2). Nonetheless, both of these subcategories seemed to require more high-frequency hearing than other subcategories, thus rendering HA users at a disadvantage. The detection of voicing feature was also similar between the 2 groups. The CI users probably used cues that do not require a high spectral resolution, such as voice onset time or duration of the following vowel for perceiving the voicing feature. The present study examined the potential contributing factors for the variability in performance within the CI and HA groups. No demographic variables were found to predict consonant contrast recognition performance in the HA children. Only age at implantation was a significant predictor for the overall consonant contrast recognition performance in the CI group (Fig. 3). These results are consistent with the literature that age at implantation is one of the most important factors that contributes to many facets of speech and language development in prelingually deafened children (1Y6,23). However, Lin and Peng (12) reported that, although age at implantation was negatively correlated with the consonant production ability of Mandarin-speaking children with CIs, it was not correlated with consonant perception performance (26). It should be noted that all of our 41 children received their CIs before 50 months of age, whereas only 3 of the 30 children in the study by Lin and Peng received their CIs before that age. Most likely, age at implantation exerts it effects on consonant perception only during the early sensitive stage of speech development. In summary, the present study examined one of the important aspects of speech perception, consonant contrast recognition, in prelingually deafened, Mandarinspeaking children and compared the performance of CI and HA children. Results showed that CI and HA children scored lower than NH children on the overall consonant contrast recognition tasks by, on average, approximately 8 percentage points. Approximately 40% of the CI and HA children had not reached a performance level that was Otology & Neurotology, Vol. 34, No. 3, 2013
Copyright © 2013 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
476
Q. LIU ET AL.
within the range of performance of the NH, age-matched controls after an average of 2 years of device use. The retroflex/nonretroflex and same place/different manner consonant contrasts seemed to be the most difficult contrasts for both groups of hearing-impaired children. CI children had more difficulties with same manner/different place contrasts, whereas HA children had more difficulties with fricative/nonfricative contrasts. When chronologic age, duration of device use, age at implantation or HA fitting, and aided sound-field pure-tone thresholds were similar, there was no significant difference in the consonant recognition scores between the CI and HA groups. Early implantation tended to yield better consonant contrast recognition in the young children with CIs. However, a fairly large amount of variance in performance was not accounted for by the demographic variables studied. Future longitudinal studies are necessary to explore the developmental trajectories for both the CI and HA children in consonant contrast recognition as well as many other facets of speech and language skills in the Mandarin-speaking population. REFERENCES 1. Svirsky MA, Tajudeen BA, Waltzman SB, et al. Speech perception in congenitally deaf children receiving cochlear implants in the first year of life. Otol Neurotol 2010;31:1254Y60. 2. Kirk KI, Miyamoto RT, Lento CL, et al. Effects of age at implantation in young children. Ann Otol Rhinol Laryngol 2002; 111:69Y73. 3. Xu L, Chen XW, Lu HY, et al. Tone perception and production in pediatric cochlear implants users. Acta Otolaryngol 2011; 131:395Y8. 4. Connor CM, Craig HK, Raudenbush SW, et al. The age at which young deaf children receive cochlear implants and their vocabulary and speech-production growth: is there an added value for early implantation? Ear Hear 2006;27:628Y44. 5. Sharma A, Dorman MF, Spahr AJ. A sensitive period for the development of the central auditory system in children with cochlear implants: implications for age of implantation. Ear Hear 2002; 23:532Y9. 6. Tyler RS, Fryhauf-Bertschy H, Kelsay DMR, et al. Speech perception by prelingually deaf children using cochlear implants. Otolaryngol Head Neck Surg 1997;117:180Y7. 7. Zhu M, Fu QJ, Galvin J, et al. Mandarin Chinese speech recognition by pediatric cochlear implant users. Int J Pediatr Otorhinolaryngol 2011;75:793Y800. 8. Wang NM, Wu CM, Kirk KI. Lexical effects on spoken word recognition performance among mandarin-speaking children with
9. 10.
11. 12. 13.
14. 15. 16. 17. 18.
19. 20. 21. 22. 23. 24. 25. 26.
normal hearing and cochlear implants. Int J Pediatr Otorhinolaryngol 2010;74:883Y90. Liu C, Liu S, Zhang N, et al. Standard-Chinese lexical neighborhood test in normal-hearing young children. Int J Pediatr Otorhinolaryngol 2011;75:774Y81. Liu H, Liu S, Wang S, et al. Effects of lexical characteristics and demographic factors on Mandarin Chinese open-set word recognition in children with cochlear implants. Ear Hear 2013;doi: 10.1097/AUD.0b013e31826d0bc6. Wu JL, Yang HM. Speech perception of Mandarin Chinese speaking young children after cochlear implant use: effect of age at implantation. Int J Pediatr Otorhinolaryngol 2003;67:247Y53. Lin YS, Peng SC. Acquisition profiles of syllable-initial consonants in Mandarin-speaking children with cochlear implants. Acta Otolaryngol 2003;123:1046Y53. Zheng Y, Soli S, Meng Z, et al. Assessment of Mandarin-speaking pediatric cochlear implant recipients with the Mandarin Early Speech Perception (MESP) test. Int J Pediatr Otorhinolaryngol 2010;74:920Y5. Somers MN. Speech perception abilities in children with cochlear implants or hearing aids. Am J Otol 1991;12:174Y8. Osberger MJ, Robbins AM, Miyamoto RT, et al. Speech perception abilities of children with cochlear implants, tactile aids or hearing aids. Am J Otol 1991;12:80Y8. Dawson PW, Blamey PJ, Rowland LC, et al. Cochlear implants in children, adolescents, and prelinguistically deafened adults: speech perception. J Speech Hear Res 1992;35:401Y17. Geers AE, Moog JS. Assessing the benefits of cochlear implants in an oral education program. In: Uziel AS, Mondain M, eds. Cochlear Implants in Children. Basel, Switzerland: Karger, 1995:119Y24. Snik AFM, Vermeulen AM, Brokx AM, et al. Speech perception performance of children with a cochlear implant compared to that of children with conventional hearing aids: 1. The ‘‘equivalent hearing loss’’ concept. Acta Otolaryngol 1997;117:750Y4. Tomblin BJ, Spencer L, Flock S, et al. A comparison of language achievement in children with cochlear implants and children using hearing aids. J Speech Lang Hear 1999;42:497Y511. Law ZWY, So LKH. Phonological abilities of hearing-impaired Cantonese-speaking children with cochlear implants or hearing aids. J Speech Lang Hear Res 2006;49:1342Y53. Phatak SA, Yoon YS, Gooler DM, et al. Consonant recognition loss in hearing impaired listeners. J Acoust Soc Am 2009;126:2683Y94. Zhou N, Xu L, Lee C-Y. The effects of frequency-place shift on consonant confusion in cochlear implant simulations. J Acoust Soc Am 2010;128:401Y9. Zhou N, Huang J, Chen X, Xu L. Relationship between tone perception and production in prelingually-deafened children with cochlear implants. Otol Neurotol 2013;34:499Y506. Thornton AR, Raffin MJ. Speech-discrimination scores modeled as a binomial variable. J Speech Hear Res 1978;21:507Y18. Studebaker GA. A rationalized arcsine transform. J Speech Hear Res 1985;28:455Y62. Peng SC, Weiss AL, Cheung H, et al. Consonant production and language skills in Mandarin-speaking children with cochlear implants. Arch Otolaryngol Head Neck Surg 2004;130:592Y7.
Otology & Neurotology, Vol. 34, No. 3, 2013
Copyright © 2013 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
Appendix 1. Word list for the Mandarin Consonant Phonetic Contrast Perception Test Item
pinyin
number
Chinese
Item
character
number
Fricative vs nonfricative
pinyin
Chinese
Item
character
number
pinyin
Chinese character
Same place and different
Same manner and different
manners
place
1
hé/é
河/鹅
30
dù/chù
肚/触
59
shí/xí
石/席
2
sè/è
色/饿
31
dù/cù
肚/醋
60
sī/xī
撕/吸
32
tǔ/zhǔ
土/煮
61
zī/jī
姿/鸡
Voiced vs voiceless 3
tā/lā
塌/拉
33
tú/zú
涂/足
62
zhī/jī
织/鸡
4
dā/lā
搭/拉
34
dú/zú
读/足
63
cì/qì
刺/汽
5
tù/rù
兔/褥
35
tù/chù
兔/触
64
chì/qì
翅/汽
6
dù/rù
肚/褥
36
tù/cù
兔/醋
65
bāo/gāo
包/高
7
chòu/ròu
臭/肉
37
dǔ/zhǔ
堵/煮
66
fǔ/hǔ
斧/虎
8
còu/ròu
凑/肉
38
zī/shī
姿/狮
67
pào/kào
炮/靠
9
zòu/ròu
揍/肉
39
chì/sì
翅/四
68
shǔ/hǔ
鼠/虎
10
zhù/rù
柱/褥
40
chù/shù
触/树
69
sū/hū
酥/呼
11
nù/tù
怒/兔
41
zhū/shū
猪/书
70
dāo/gāo
刀/高
12
māo/bāo
猫/包
42
jī/xī
鸡/吸
71
tào/kào
套/靠
13
mào/pào
帽/炮
43
zhī/sī
织/撕
72
fù/shù
父/树
14
nào/dào
闹/稻
44
zì/sì
字/四
73
bāo/dāo
包/刀
15
chòu/lòu
臭/漏
45
cì/shì
刺/室
74
má/ná
麻/拿
16
cā/lā
擦/拉
46
qī/xī
七/吸
75
fù/sù
父/塑
17
zòu/lòu
揍/漏
47
cì/sì
刺/四
76
pào/tào
炮/套
18
zhòu/lòu
皱/漏
48
dù/shù
肚/树
Aspirate vs nonaspirate
19
shù/lù
树/鹿
49
tù/shù
兔/树
77
zhū/cū
猪/粗
20
sù/lù
塑/鹿
50
tù/sù
兔/塑
78
zì/chì
字/翅
21
nù/shù
怒/树
51
gǔ/hǔ
骨/虎
79
jī/qī
鸡/七
22
mǔ/fǔ
母/斧
52
bēi/fēi
杯/飞
80
zhū/chū
猪/出
23
nù/sù
怒/塑
53
dù/sù
肚/塑
81
zì/cì
字/刺
24
ná/zá
拿/砸
54
pǔ/fǔ
谱/斧
82
bāo/pāo
包/抛
25
nù/cù
怒/醋
55
ké/hé
壳/河
83
gū/kū
菇/哭
26
nù/chù
怒/触
56
nù/lù
怒/鹿
84
dào/tào
稻/套
27
nù/zhù
怒/柱
57
nù/rù
怒/褥
28
shòu/ròu
瘦/肉
58
ròu/lòu
肉/漏
29
sù/rù
塑/褥
Pinyin, the phonemic spelling system for Chinese.
Retroflex vs nonretroflex 85
shì/sì
室/四
86
zhǐ/zǐ
纸/籽
87
chū/cū
出/粗
Mandarin Consonant Contrast Recognition Among Children With Cochlear Implants or Hearing Aids and Normal-Hearing Children *Qiaoyun Liu, †‡Ning Zhou, ‡Rebecca Berger, *Daniel Huang, and ‡Li Xu *Key Laboratory of Hearing and Speech Sciences, East China Normal University, Shanghai, People’s Republic of China; ÞKresge Hearing Research Institute, University of Michigan, Ann Arbor, Michigan; and þSchool of Rehabilitation and Communication Sciences, Ohio University, Athens, Ohio, U.S.A.
Hypothesis: The purpose of the present study was to investigate the consonant recognition of Mandarin-speaking children with cochlear implants (CIs) and hearing aids (HAs) and to determine if they reach a level of consonant recognition similar to that of normal-hearing (NH) children. Background: Little information is available in the literature regarding the consonant perception abilities of prelingually deafened young children with either CIs or HAs. No studies have compared Mandarin-Chinese consonant contrast recognition in CI and HA children. Methods: Forty-one prelingually deafened children with CIs, 26 prelingually deafened children with HAs, and 30 NH children participated in this study. The 3 groups were matched for chronologic age (3Y5 yr). The hearing-impaired groups were matched for age at fitting of the devices, duration of device use, and aided hearing threshold. All subjects completed a computerized Mandarin consonant phonetic contrast perception test.
Results: CI and HA children scored, on average, approximately 8 percentage points below the mean NH group performance on the consonant contrast recognition. Approximately 40% of the CI and HA children had not reached a performance level of the NH group. No significant differences in the consonant recognition scores were found between the CI and HA groups. Age of implantation was correlated with consonant contrast recognition in the CI group. Conclusion: When age at fitting of the devices, duration of device use, and aided thresholds are matched at the group level, consonant recognition is similar between the CI and HA children after 2 years of device use. Early implantation tends to yield better consonant contrast recognition in the young children with CIs. However, a large amount of variance in performance was not accounted for by the demographic variables studied. Key Words: Cochlear implantsVConsonant recognitionV Mandarin ChineseVPediatric. Otol Neurotol 34:471Y476, 2013.
The population of prelingually deafened children who have received cochlear implants (CIs) has increased rapidly in China. As of November 2012, more than 15,000 profoundly deafened children have received multichannel CIs in China. The number of children with CIs enrolled in mainstream preschools and elementary schools in China is also rapidly increasing. It is essential for CI children to achieve a high level of listening ability in order to succeed in mainstream schools in China be-
cause assistive services, such as note takers or sign language interpreters, are not readily available. Although there is a wealth of data to indicate whether CI children can reach a level of listening ability that is within the normal range for the English-speaking population, there is very little data for the Chinese-speaking CI users. There is plenty of evidence showing that, for prelingually deafened children, early implantation is important for developing normal or close-to-normal speech and language skills (e.g., 1Y6). Several research laboratories have examined speech perception ability and its relationship with age at implantation in Mandarin-Chinese speaking children with CIs. For example, Zhu et al. (7) tested the ability of Mandarin speaking CI users to understand disyllable words and sentences. On average, prelingually deafened children with CIs achieved approximately 80% correct for disyllabic word and sentence recognition. Age at implantation predicted the recognition performance for both disyllabic words and sentences. Wang et al. (8)
Address correspondence and reprint requests to Li Xu, M.D., Ph.D., School of Rehabilitation and Communication Sciences, Ohio University, Athens, OH 45701; E-mail: [email protected] This study was supported by the Open Fund of Key Laboratory of Speech and Hearing Sciences, East China Normal University, Ministry of Education, Shanghai, China. Declaration of Interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this paper. Supplemental digital content is available in the text.
471
Copyright © 2013 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
472
Q. LIU ET AL.
examined how speech perception in Mandarin-speaking CI children compare with their normal-hearing (NH) peers. Thirty normal-hearing children and 36 CI children (4Y11.6 yr old) were given a newly developed version of the Monosyllabic Lexical Neighborhood Test and the Disyllabic Lexical Neighborhood Test in Mandarin Chinese to determine the effects of linguistic and cognitive demands on speech perception performance. Monosyllabic and disyllabic words with frequencies above the median and lexical densities below the median were labeled as ‘‘easy’’ words, and the opposite were labeled as ‘‘hard’’ words. Lexical effects were demonstrated for disyllabic word recognition. More than half of the CI children fell into T1 standard deviation (SD) of the mean word recognition of the NH group. More recently, Liu et al. (9,10) reported results obtained from 230 children with CIs using their own version of the Standard-Chinese Lexical Neighborhood Test. Lexical effects similar to those found in the NH children were shown for both monosyllabic and disyllabic words in the CI group. The performance of the CI group was poorer than that reported in Wang et al. (8), with only a small portion of the CI children falling within the normal range. An association between age at implantation and open-set word recognition performance was also demonstrated (10). Few studies have examined consonant recognition abilities in Chinese-speaking children with CIs. Wu and Yang (11) evaluated how speech perception of CI children continues to improve after implantation. In that study, 16 Mandarin-speaking CI children between the ages of 1.6 and 6 years were given the Mandarin Auditory Perception Test Battery every 6 months for 2 years after implantation. Between 12 and 24 months after implantation, spondee, vowel, and consonant recognition improved significantly. Interestingly, although spondee, vowel, phrase, and sentence recognition showed a negative correlation with age at implantation, consonant recognition was not correlated with age at implantation. Lin and Peng (12) found that the average scores for the initial consonant discrimination in a group of 30 Mandarin-speaking children (6Y12.5 yr old) with CIs was 76.7% correct. More recently, Zheng et al. (13) showed that 4 of 25 CI children between 3 and 16.9 years old were able to achieve consonant recognition performance close to that of the NH children of similar hearing ages when tested with a closed-set Mandarin Early Speech Perception Test. A number of studies have attempted to compare the English speech and language skills of children with CIs to children fitted with conventional hearing aids (HAs) (4,14Y19). Comparison between these groups is often difficult because of the differences in factors such as age, length of device use, amount of hearing loss, and educational setting. However, the consensus in recent literature is that contemporary multichannel CIs provide prelingually deafened children with enhanced speech and language development compared with conventional HAs. In particular, Connor et al. (4) demonstrated that children who received CIs before the age of 7 exhibited a steeper slope in the development of consonant production accu-
racy when compared with the predicted trajectory for the HA children. However, no study has compared the speech and language development between the CI and HA users in the Mandarin Chinese population. In a recent study, Law and So (20) examined both speech production and speech perception in Cantonese-speaking CI and HA users. Seven HA users and 7 CI users, matched for mean chronologic age (range, 5.1Y6.4 yr) and years of speech therapy (range, 1.25Y3.25 yr), were given the Cantonese Segmental Phonology Test and the Cantonese Lexical Comprehension Test. The CI users demonstrated significantly better consonant production skills than the HA users but not better vowel production. There was no significant difference in the speech perception tasks between the 2 groups. Results indicated that CIs seemed to promote the development of consonant production compared with HAs. There is an urgent need to study the many facets of speech and language development in the rapidly increasing population of Chinese-speaking children with CIs and HAs. The present study focuses on consonant contrast perception in the 2 groups. Consonant landmarks have been found to contribute greatly to speech recognition in noise for NH listeners and CI users. We examined the differences in consonant contrast recognition abilities between Mandarin-speaking CI and HA children with group-matched chronologic age, duration of device use, age at implantation or HA fitting, and aided soundfield pure-tone thresholds. We determined the proportion of CI and HA users who could reach a level of consonant contrast recognition that is equivalent to the NH controls. A systematic comparison of the consonant confusion patterns in the HA and CI users is not available. In acoustic hearing, consonant confusion by hearing-impaired adult listeners might be influenced by high-frequency hearing loss even when audibility is compensated (21), whereas the voicing features of consonants may be well received in listeners with HAs. In electrical hearing, it seems that the frequency range per se is not the issue; instead, performance may be affected by frequency resolution as well as distortion because of mismatch in the electrode position and the frequency allocation (22). We hypothesize that because of the inherent differences between electrical stimulation with a CI and acoustic hearing with amplification, the 2 groups would be better at perceiving different consonant features. Specifically, the voicing feature may be better transmitted acoustically than electrically because of the loss of pitch information in the envelope-based speech coding in CI devices. On the other hand, highfrequency fricatives may be better recognized via a CI because amplification of the high frequencies could distort the speech cues. Finally, we examine the effects of various demographic variables as potential predictors for consonant contrast recognition in both HA and CI children. MATERIALS AND METHODS Subjects Forty-one prelingually deafened children with CIs (24 boys and 17 girls), 26 prelingually deafened children with HAs
Otology & Neurotology, Vol. 34, No. 3, 2013
Copyright © 2013 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
CONSONANT CONTRAST RECOGNITION TABLE 1.
CI (n = 41) HA (n = 26)
Range Mean T SD Range Mean T SD
473
Demographic information for the cochlear implant and hearing aid groups Chronologic age (m)
Duration of device use (m)
Age at implantation or at HA fitting (m)
Aided sound-field PTAa (dB HL)
29Y60 40.2 T 9.1 31Y60 44.5 T 9.3
6Y44 16.3 T 8.2 6Y37 20.4 T 9.1
10Y50 24.0 T 8.9 5Y47 24.2 T 10.3
23Y52 38.6 T 5.5 33Y54 42.7 T 4.8
CI indicates cochlear implant; HA, hearing aid; PTA, pure-tone average; SD, standard deviation. a PTA is the average threshold of 500, 1,000, 2,000, and 4,000 Hz tested with warble tones.
(16 boys and 10 girls), and 30 NH children (15 boys and 15 girls) were recruited from a rehabilitation center and a kindergarten in Shanghai, China. All of the participants are native Mandarin Chinese speakers. The ages of the CI children ranged from 29 to 60 months (mean, 40.2 mo) and amount of implant experience ranged from 6 to 44 months (mean, 16.3 mo). The age at implantation ranged from 10 to 50 months (mean, 24.0 mo), and the mean auditory threshold with the CI was 38.6 dB HL (SD, 5.5 dB). These children were implanted with Advanced Bionics, Nucleus, and Med-El devices. The ages of the HA children ranged from 31 to 60 months (mean, 44.5 mo). The amount of hearing aid experience ranged from 6 to 37 months (mean, 20.4 mo), and the age at hearing aid fitting ranged from 5 to 47 months (mean, 24.2 mo). The mean auditory threshold while using the hearing aids was 42.7 dB HL (SD, 4.8). All HA children were fitted with digital hearing aids. Table 1 summarizes the demographic information of the 2 hearing-impaired groups. The NH children were between 36 and 60 months old. The 3 groups were matched for the mean chronologic age (3Y5 yr old), and the 2 hearing-impaired groups were matched for the mean age at fitting of devices, duration of device use, and aided hearing threshold. The use of human subjects was reviewed and approved by the institutional review board of East China Normal University.
the positions of the 2 words were randomized each time. Familiarization of the test procedure was provided for each child before the test. Percent correct scores for each subcategory as well as the overall recognition accuracy were determined for each child. Percent correct scores were arcsine-transformed to stabilize the variance for the analysis of variance (ANOVA) and other statistical analysis (24,25).
RESULTS Figure 1 shows the overall consonant recognition performance for the 3 subject groups. Scores for the 30 NH children ranged from 83.1% to 98.5% correct (chance = 50% correct) with a mean of 93.3% correct (SD, 3.5%). The HA and CI children as a group scored approximately 7 to 8 percentage points lower than the NH group. Perception scores for the 41 CI children ranged from 53.0% to 98.9% correct with a mean of 86.2% (SD, 10.9%). Scores for the 26 HA children ranged from 65.9% to 96.9% correct with a mean of 84.8% correct (SD, 10.1%). Note that the variation in performance in the hearing-impaired groups was larger than the NH group. A 1-way ANOVA
Mandarin Consonant Phonetic Contrast Perception Test The consonant contrast perception test involved word identification from a total of 87 Mandarin consonant contrasts. Based on the characteristics of Chinese phonetics, the 87 consonant contrasts were divided into the following 6 subcategories: 1) 2 pairs of fricative/nonfricative contrasts (e.g., /he/ versus /e/), 2) 27 pairs of voiceless/voiced contrasts (e.g., /ma/ versus /ba/), 3) 8 pairs of aspirate/nonaspirate contrasts (e.g., /bao/ versus /pao/), 4) 29 pairs of same place/different manner contrasts (e.g., /bei/ versus /fei/), 5) 18 pairs of same manner/different place contrasts (e.g., /bao/ versus /dao/), and 6) 3 pairs of retroflex/nonretroflex contrasts (e.g., /zi/ versus /zhi/). All the pairs contained simple monosyllabic Mandarin Chinese words that children in this age group are familiar with. The vowels and their carrying tones of the 2 words in a contrast were the same. The speech materials were recorded from a female speaker (whose average voice F0 was 355 Hz) using a 44,100-Hz sampling rate and 16-bit resolution. A word list for the Mandarin Consonant Phonetic Contrast Perception Test is provided in the Supplemental Digital Content (http://links.lww.com/MAO/A143). The subjects participated in a computerized consonant contrast test. The test used a 2-alternative forced-choice paradigm. A consonant contrast token was played via a loudspeaker at a level of 70 dB SPL, while 2 pictures representing the consonant contrast were shown on a computer screen. The subject was asked to choose the 1 photograph that represented the word that he or she had heard. Each contrast pair was tested 3 times, and
FIG. 1. Overall consonant contrast recognition scores for the NH, CI, and HA groups. Each symbol represents 1 subject. Each box plot has lines at the lower quartile, median, and upper quartile values. The vertical whiskers show the extent of the rest of the data (with the exception of 1 data point in the NH group and 1 in the CI group that were considered to be outliers). The bracket at the top represents statistical significant differences between the 2 groups indicated. Otology & Neurotology, Vol. 34, No. 3, 2013
Copyright © 2013 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
474
Q. LIU ET AL.
FIG. 2. The mean consonant contrast recognition scores of the 6 subcategories for the NH, CI, and HA groups. The error bar represents the standard deviation. The bracket at the top represents statistical significant differences between the 2 groups indicated.
showed that the mean performance was significantly different among the 3 groups ( p G 0.001). Post hoc pairwise comparisons indicated that the performance of both hearing-impaired groups were significantly different from the NH group ( p G 0.05) but not significantly different from each other ( p 9 0.05). The brackets plotted in Figure 1 illustrate the significant statistical results of these comparisons ( p G 0.05). The overall consonant contrast recognition performance of 18 (43.9%) and 24 (58.5%) CI children fell within T1 SD and T 2 SDs of that of the NH children, respectively. Similarly, the overall consonant contrast recognition performance of 11 (42.3%) and 15 (57.7%) HA children fell within T1 SD and T2 SDs of that of the NH children, respectively. Figure 2 plots the group mean and SD of the consonant contrast recognition scores for the 6 subcategories. A 2-way ANOVA with subject group and consonant contrast as 2 main factors was performed to test whether the mean performance was different among the 3 subject groups and among the 6 subcategories. Results of the 2-way ANOVA revealed statistical significance for both group and subcategory factors ( p G 0.001). Post hoc pairwise comparisons were then performed to identify group pairs that showed significant differences for each consonant contrast. The brackets plotted in Figure 2 illustrate the significant statistical results of these comparisons ( p G 0.05). The CI and HA groups performed equally well for each of the 6 consonant contrasts. The performance of the NH group was significantly higher than that of the HA group in 3 of the 6 subcategories: fricative/nonfricative, same place/different manner, and retroflex/nonretroflex contrasts. The performance of the NH group was also significantly higher than that of the CI group in 3 of the 6 subcategories: same place/ different manner, same manner/different place, and retroflex/nonretroflex contrasts. No statistically significant differences were found in the voiceless/voiced and aspirate/nonaspirate contrasts between the NH and hearing-
impaired groups. The largest performance gap between the hearing-impaired and the NH groups was for the retroflex/ nonretroflex contrast. A correlational analysis was performed to determine the contribution of various demographic variables for the variance in the performance. Demographic variables for the CI subjects included age at implantation, duration of implant use, chronologic age, and acoustic hearing threshold with the implant. Age at implantation was found to be the only significant predictor. Figure 3 shows the correlation between age at implantation and the overall consonant contrast recognition scores. The correlation coefficient (r) of the linear fit was j0.343 ( p = 0.028). After we removed the lowest performer as an outlier
FIG. 3. Scatter plot of the overall consonant contrast recognition scores as a function of age at implantation for CI children. Each symbol represents 1 subject. The solid line represents the linear fit of the data.
Otology & Neurotology, Vol. 34, No. 3, 2013
Copyright © 2013 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
CONSONANT CONTRAST RECOGNITION (defined as a data point that is 93 SDs away from the mean of the group) (Fig. 1), the correlation strength (r) improved to j0.478 ( p = 0.0013). Demographic variables for the HA subjects included age at fitting, duration of hearing aid use, chronologic age, and aided threshold. No significant correlation was found between any of these variables and the performance.
DISCUSSION The consonant contrast recognition test was a fairly easy task for the NH children between 3 and 5 years of age. They scored on average higher than 90% correct on overall consonant contrast recognition and on recognition of all 6 consonant contrast subcategories (Figs. 1 and 2). The hearing-impaired groups (i.e., CI and HA groups) showed variability in performance, with scores ranging from slightly above chance (i.e., 50% correct) to nearly perfect. The overall recognition performance was 86.2% and 84.8% correct, respectively, for the CI and HA groups (Fig. 1). The performance for the CI group was approximately 10 percentage points higher than that reported by Lin and Peng (2003). The age at implantation in their CI children ranged from 27 to 123 months (mean, 68 mo), much older than that of our CI group (mean, 24 mo). This may explain the differences in the 2 studies (see below for discussion on age at implantation and consonant recognition performance). The overall consonant contrast recognition performance of approximately 60% of subjects in the hearing-impaired groups fell within T2 SDs of the performance of the NH children. In other words, approximately 40% of the CI and HA children performed greater than 2 SDs below the mean performance of the NH children who were of similar chronologic age. Note, however, that the hearing-impaired children started using CIs or HAs at an average age of 2 years (Table 1). Thus, their hearing age was, on average, 2 years younger than the hearing ages of the NH group. It is encouraging to see that more than half of the CI and HA children fell within 2 SDs of the NH controls whose hearing age was, on average, 2 years older. Will the other 40% of the hearing-impaired children catch up with more experience with their devices? Such a question can be answered with a longitudinal study of the same subjects. The lack of correlation between the duration of device use and the consonant contrast recognition performance, however, indicates that hearing age may not predict consonant recognition performance. Factors, such as chronologic age and maturity of cognitive function, are likely to interact with hearing age and may contribute to the performance with hearing devices. The results of the present study show that, when chronologic age, duration of device use, age at implantation or HA fitting, and aided sound-field pure-tone thresholds were matched for the group mean, there was no significant difference in the consonant recognition scores between the CI and HA groups (Figs. 1 and 2). These results are somewhat surprising, given that several
475
studies have shown that CI children demonstrate performance superior to HA children on many speech and language tasks (4,19). Tomblin et al. (19) demonstrated that more than 50% of the CI children, by 2 years after implantation, exceeded the 95% prediction interval of the HA children in English expressive language skills. This contrasted with only approximately 28% of the CI children who exceeded such a level at 1 year postimplantation. Connor et al. (4) also showed that the English consonant production accuracy of the CI and HA children diverged further as duration of device use became longer (1Y5 yr). In the present study, the hearing-impaired children had used their devices for an average of 2 years. No clear difference in Mandarin consonant contrast recognition was evident. Further research with a longer follow-up period will be necessary to determine whether these 2 groups will diverge in consonant contrast recognition. The 2 contrasts, fricatives/nonfricatives and retroflex and nonretroflex contrasts, tended to cause more confusion for the HA children than the CI children, although the statistical analysis did not reveal significant difference (Fig. 2). Nonetheless, both of these subcategories seemed to require more high-frequency hearing than other subcategories, thus rendering HA users at a disadvantage. The detection of voicing feature was also similar between the 2 groups. The CI users probably used cues that do not require a high spectral resolution, such as voice onset time or duration of the following vowel for perceiving the voicing feature. The present study examined the potential contributing factors for the variability in performance within the CI and HA groups. No demographic variables were found to predict consonant contrast recognition performance in the HA children. Only age at implantation was a significant predictor for the overall consonant contrast recognition performance in the CI group (Fig. 3). These results are consistent with the literature that age at implantation is one of the most important factors that contributes to many facets of speech and language development in prelingually deafened children (1Y6,23). However, Lin and Peng (12) reported that, although age at implantation was negatively correlated with the consonant production ability of Mandarin-speaking children with CIs, it was not correlated with consonant perception performance (26). It should be noted that all of our 41 children received their CIs before 50 months of age, whereas only 3 of the 30 children in the study by Lin and Peng received their CIs before that age. Most likely, age at implantation exerts it effects on consonant perception only during the early sensitive stage of speech development. In summary, the present study examined one of the important aspects of speech perception, consonant contrast recognition, in prelingually deafened, Mandarinspeaking children and compared the performance of CI and HA children. Results showed that CI and HA children scored lower than NH children on the overall consonant contrast recognition tasks by, on average, approximately 8 percentage points. Approximately 40% of the CI and HA children had not reached a performance level that was Otology & Neurotology, Vol. 34, No. 3, 2013
Copyright © 2013 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
476
Q. LIU ET AL.
within the range of performance of the NH, age-matched controls after an average of 2 years of device use. The retroflex/nonretroflex and same place/different manner consonant contrasts seemed to be the most difficult contrasts for both groups of hearing-impaired children. CI children had more difficulties with same manner/different place contrasts, whereas HA children had more difficulties with fricative/nonfricative contrasts. When chronologic age, duration of device use, age at implantation or HA fitting, and aided sound-field pure-tone thresholds were similar, there was no significant difference in the consonant recognition scores between the CI and HA groups. Early implantation tended to yield better consonant contrast recognition in the young children with CIs. However, a fairly large amount of variance in performance was not accounted for by the demographic variables studied. Future longitudinal studies are necessary to explore the developmental trajectories for both the CI and HA children in consonant contrast recognition as well as many other facets of speech and language skills in the Mandarin-speaking population. REFERENCES 1. Svirsky MA, Tajudeen BA, Waltzman SB, et al. Speech perception in congenitally deaf children receiving cochlear implants in the first year of life. Otol Neurotol 2010;31:1254Y60. 2. Kirk KI, Miyamoto RT, Lento CL, et al. Effects of age at implantation in young children. Ann Otol Rhinol Laryngol 2002; 111:69Y73. 3. Xu L, Chen XW, Lu HY, et al. Tone perception and production in pediatric cochlear implants users. Acta Otolaryngol 2011; 131:395Y8. 4. Connor CM, Craig HK, Raudenbush SW, et al. The age at which young deaf children receive cochlear implants and their vocabulary and speech-production growth: is there an added value for early implantation? Ear Hear 2006;27:628Y44. 5. Sharma A, Dorman MF, Spahr AJ. A sensitive period for the development of the central auditory system in children with cochlear implants: implications for age of implantation. Ear Hear 2002; 23:532Y9. 6. Tyler RS, Fryhauf-Bertschy H, Kelsay DMR, et al. Speech perception by prelingually deaf children using cochlear implants. Otolaryngol Head Neck Surg 1997;117:180Y7. 7. Zhu M, Fu QJ, Galvin J, et al. Mandarin Chinese speech recognition by pediatric cochlear implant users. Int J Pediatr Otorhinolaryngol 2011;75:793Y800. 8. Wang NM, Wu CM, Kirk KI. Lexical effects on spoken word recognition performance among mandarin-speaking children with
9. 10.
11. 12. 13.
14. 15. 16. 17. 18.
19. 20. 21. 22. 23. 24. 25. 26.
normal hearing and cochlear implants. Int J Pediatr Otorhinolaryngol 2010;74:883Y90. Liu C, Liu S, Zhang N, et al. Standard-Chinese lexical neighborhood test in normal-hearing young children. Int J Pediatr Otorhinolaryngol 2011;75:774Y81. Liu H, Liu S, Wang S, et al. Effects of lexical characteristics and demographic factors on Mandarin Chinese open-set word recognition in children with cochlear implants. Ear Hear 2013;doi: 10.1097/AUD.0b013e31826d0bc6. Wu JL, Yang HM. Speech perception of Mandarin Chinese speaking young children after cochlear implant use: effect of age at implantation. Int J Pediatr Otorhinolaryngol 2003;67:247Y53. Lin YS, Peng SC. Acquisition profiles of syllable-initial consonants in Mandarin-speaking children with cochlear implants. Acta Otolaryngol 2003;123:1046Y53. Zheng Y, Soli S, Meng Z, et al. Assessment of Mandarin-speaking pediatric cochlear implant recipients with the Mandarin Early Speech Perception (MESP) test. Int J Pediatr Otorhinolaryngol 2010;74:920Y5. Somers MN. Speech perception abilities in children with cochlear implants or hearing aids. Am J Otol 1991;12:174Y8. Osberger MJ, Robbins AM, Miyamoto RT, et al. Speech perception abilities of children with cochlear implants, tactile aids or hearing aids. Am J Otol 1991;12:80Y8. Dawson PW, Blamey PJ, Rowland LC, et al. Cochlear implants in children, adolescents, and prelinguistically deafened adults: speech perception. J Speech Hear Res 1992;35:401Y17. Geers AE, Moog JS. Assessing the benefits of cochlear implants in an oral education program. In: Uziel AS, Mondain M, eds. Cochlear Implants in Children. Basel, Switzerland: Karger, 1995:119Y24. Snik AFM, Vermeulen AM, Brokx AM, et al. Speech perception performance of children with a cochlear implant compared to that of children with conventional hearing aids: 1. The ‘‘equivalent hearing loss’’ concept. Acta Otolaryngol 1997;117:750Y4. Tomblin BJ, Spencer L, Flock S, et al. A comparison of language achievement in children with cochlear implants and children using hearing aids. J Speech Lang Hear 1999;42:497Y511. Law ZWY, So LKH. Phonological abilities of hearing-impaired Cantonese-speaking children with cochlear implants or hearing aids. J Speech Lang Hear Res 2006;49:1342Y53. Phatak SA, Yoon YS, Gooler DM, et al. Consonant recognition loss in hearing impaired listeners. J Acoust Soc Am 2009;126:2683Y94. Zhou N, Xu L, Lee C-Y. The effects of frequency-place shift on consonant confusion in cochlear implant simulations. J Acoust Soc Am 2010;128:401Y9. Zhou N, Huang J, Chen X, Xu L. Relationship between tone perception and production in prelingually-deafened children with cochlear implants. Otol Neurotol 2013;34:499Y506. Thornton AR, Raffin MJ. Speech-discrimination scores modeled as a binomial variable. J Speech Hear Res 1978;21:507Y18. Studebaker GA. A rationalized arcsine transform. J Speech Hear Res 1985;28:455Y62. Peng SC, Weiss AL, Cheung H, et al. Consonant production and language skills in Mandarin-speaking children with cochlear implants. Arch Otolaryngol Head Neck Surg 2004;130:592Y7.
Otology & Neurotology, Vol. 34, No. 3, 2013
Copyright © 2013 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
Appendix 1. Word list for the Mandarin Consonant Phonetic Contrast Perception Test Item
pinyin
number
Chinese
Item
character
number
Fricative vs nonfricative
pinyin
Chinese
Item
character
number
pinyin
Chinese character
Same place and different
Same manner and different
manners
place
1
hé/é
河/鹅
30
dù/chù
肚/触
59
shí/xí
石/席
2
sè/è
色/饿
31
dù/cù
肚/醋
60
sī/xī
撕/吸
32
tǔ/zhǔ
土/煮
61
zī/jī
姿/鸡
Voiced vs voiceless 3
tā/lā
塌/拉
33
tú/zú
涂/足
62
zhī/jī
织/鸡
4
dā/lā
搭/拉
34
dú/zú
读/足
63
cì/qì
刺/汽
5
tù/rù
兔/褥
35
tù/chù
兔/触
64
chì/qì
翅/汽
6
dù/rù
肚/褥
36
tù/cù
兔/醋
65
bāo/gāo
包/高
7
chòu/ròu
臭/肉
37
dǔ/zhǔ
堵/煮
66
fǔ/hǔ
斧/虎
8
còu/ròu
凑/肉
38
zī/shī
姿/狮
67
pào/kào
炮/靠
9
zòu/ròu
揍/肉
39
chì/sì
翅/四
68
shǔ/hǔ
鼠/虎
10
zhù/rù
柱/褥
40
chù/shù
触/树
69
sū/hū
酥/呼
11
nù/tù
怒/兔
41
zhū/shū
猪/书
70
dāo/gāo
刀/高
12
māo/bāo
猫/包
42
jī/xī
鸡/吸
71
tào/kào
套/靠
13
mào/pào
帽/炮
43
zhī/sī
织/撕
72
fù/shù
父/树
14
nào/dào
闹/稻
44
zì/sì
字/四
73
bāo/dāo
包/刀
15
chòu/lòu
臭/漏
45
cì/shì
刺/室
74
má/ná
麻/拿
16
cā/lā
擦/拉
46
qī/xī
七/吸
75
fù/sù
父/塑
17
zòu/lòu
揍/漏
47
cì/sì
刺/四
76
pào/tào
炮/套
18
zhòu/lòu
皱/漏
48
dù/shù
肚/树
Aspirate vs nonaspirate
19
shù/lù
树/鹿
49
tù/shù
兔/树
77
zhū/cū
猪/粗
20
sù/lù
塑/鹿
50
tù/sù
兔/塑
78
zì/chì
字/翅
21
nù/shù
怒/树
51
gǔ/hǔ
骨/虎
79
jī/qī
鸡/七
22
mǔ/fǔ
母/斧
52
bēi/fēi
杯/飞
80
zhū/chū
猪/出
23
nù/sù
怒/塑
53
dù/sù
肚/塑
81
zì/cì
字/刺
24
ná/zá
拿/砸
54
pǔ/fǔ
谱/斧
82
bāo/pāo
包/抛
25
nù/cù
怒/醋
55
ké/hé
壳/河
83
gū/kū
菇/哭
26
nù/chù
怒/触
56
nù/lù
怒/鹿
84
dào/tào
稻/套
27
nù/zhù
怒/柱
57
nù/rù
怒/褥
28
shòu/ròu
瘦/肉
58
ròu/lòu
肉/漏
29
sù/rù
塑/褥
Pinyin, the phonemic spelling system for Chinese.
Retroflex vs nonretroflex 85
shì/sì
室/四
86
zhǐ/zǐ
纸/籽
87
chū/cū
出/粗
