Evaluation of speech perception in noisy environments for normally hearing children was conducted in order to provide normal data for speech perception testing in children with hearing impairments thus improving early intervention alternatives for Mandarin-speaking children with hearing impairments. The speech perception abilities of 174 developmentally normal children ranging aged 2-5 years, in four age groups, were evaluated in environments that were quiet or with high levels of competing noise using the Mandarin pediatric speech intelligibility (MPSI) test. The mean score of MPSI between the four age groups showed notable statistical differences, including a variation in mean score between the four age groups, clearly indicating that the speech perception abilities of young children in noisy environments improved greatly with age, most notably between the ages of 3 and 4 years old. Speech perception ability in noisy environments was shown to be significantly, but weakly, related to age, implying the presence of other, possibly environment factors, in speech perception development. Furthermore, no statistically significant difference between boys and girls was noted in the experimental MPSI scores. The ability of children to increasingly perceive speech in environments containing high competing noise levels was shown to gradually and progressively increase with age. These results indicated that the developing Mandarin speech perception abilities in noisy environments in normal hearing children develops substantially after the age of 3-4 years, suggesting that similar age ranges may be even more critical intervention points for children with hearing impairments. More studies are still needed to confirm that.
Keywords: Childhood speech, competing noise, hearing impairment speech, Mandarin speech perception
|How to cite this article:|
Meng Z, Zheng Y, Wang K, Li D. Evaluation of speech perception in competing noise conditions for normally hearing children. Noise Health 2013;15:178-82
| Introduction|| |
The development of audiology and the universal implementation of newborn hearing screenings have increased early identification of hearing impaired children. Timely intervention, often utilizing hearing aids (HA) and rehabilitation, during critical developmental periods in childhood are a direct benefit of these programs. As these programs become increasingly ubiquitous, the need to evaluate the success of the various intervention alternatives employed in the treatment of these very young children increases dramatically.  Previous procedures used to measure the success of various intervention strategies have included longitudinal comparisons of the language development progression experienced by children using various HA. Alternatively, cross-sectional evaluations by HA type and classification have also been employed. Unfortunately, too few of these studies have been completed to generate a full profile of treatment for childhood hearing impairment, particularly in Mandarin-speaking populations.
Most studies that examine hearing impairment strategies include the comparison of two important measurements, the language abilities of normally hearing and hearing impaired children. The majority of treatments aim for the rehabilitation of hearing impaired children, with the desired outcome of ultimately approaching the level of language development observed in normally hearing children. The language development of normally hearing children provides a benchmark that can not only could guide rehabilitation of hearing impaired children but can also provide insight into the developmental process of speech recognition, a process which does not occur linearly with age.  The primary purpose of the current analysis of the language development in normally hearing children is to better understand the contextual process of speech recognition development in hearing impaired' children as they develop use of the Mandarin language.
Understanding the developmental process of speech recognition can result in a greater understanding of treatment strategies for rehabilitation of hearing impaired children, particularly targeting critical points in development. Previously, researchers analyzed speech development in normally hearing children, resulting in the classification of speech development into three distinct stages: Speech and sound detection, complex and phonetic speech perception and phonemic-syllabic representation. ,,, Each stage is a milestone for language development, wherein children acquire speech detection, speech perception and, ultimately, speech recognition. ,,, This three-stage process suggests that children learn language in a series of steps of increasing complexity with a progression that may move in a step-wise instead of linear fashion. A step-wise progression in speech and sound differentiation would suggest that certain stages, potentially represented by age groups, are likely to be the most receptive to treatment interventions.
Speech development in normally hearing children still represents the goal of intervention treatments. Previously, extensive normative data for speech perception in normally hearing children has been compiled; however, few studies address the issue of speech perception in noisy environments where high levels of competing noise are present. In practical situations of daily living, hearing impaired children are often particularly challenged by noisy environments. Pediatric speech intelligibility (PSI) is widely used to analyze speech perception of hearing impaired children and children with auditory processing disorders. ,, Previously, speech testing materials for Mandarin-speaking children were rare, resulting in the development of the Mandarin pediatric speech intelligibility (MPSI), based on the PSI, for speech testing in young children by Yun Zheng et al. The current study utilizes MPSI for assessment of Mandarin-speaking children. Assessment of the speech perception of children in noisy environments more realistically reflects the situations they are likely to encounter in their daily life. The purpose of this study is to evaluate the speech perception of normally hearing children in such environments, containing competing noise, and provide normal data of their speech perception useful in the interpretation of speech perception abilities in children with hearing impairments.
| Methods|| |
Study participants were recruited from four normal kindergartens in Chengdu, the capital of the Sichuan province in Southwest China. This research was approved by the Ethics Committee of the West China Hospital in Sichuan University. Subjects were enrolled after parents/guardians of each minor participant provided written letters of informed consent.
A total of 184 children aged 2-5 years were screened for inclusion in the current study, resulting in enrollment of 174 children that met the inclusion criteria. No subjects withdrew from the study. The subjects were divided into four age groups: 2 years, 3 years, 4 years and 5 years groups. The participants in the 2 years group ranged from 2 years to 2 years and 11 months in age, the 3 years group ranged from 3 years to 3 years and 11 months in age, the 4 years group ranged from 4 years to 4 years and 11 months in age and the 5 years group ranged from 5 years to 5 years and 11 months in age. The age and gender distribution of subjects enrolled in the current study is described in [Table 1].
Inclusion criteria included normal cognitive development as determined by each subject's individual teacher or instructor, passing scores on a standard hearing screen applied using transient evoked otoacoustic emission, age ranging from 2 to 5 years old or 11 months old. Exclusion criteria included abnormal cognitive development as determined by each subject's individual teacher or instructor, failing scores by standard hearing screening methods previously described or failure of subject cooperation resulting in incomplete MPSI test results.
Speech perception in noisy environments was tested by MPSI. The technique is a close-set test for Mandarin-speaking children consisting of 12 sentences and 12 competition sentences. The 12 sentences were grouped into two picture plates. Another two sentences were grouped into a 3 rd picture plate as a practice test for instructional purposes only. The results of the practice test (3 rd plate) are not calculated into final scores.
In each test, the child was asked to point to the corresponding picture upon hearing an appropriate sentence. Sentences were delivered either in a quiet environment or in the presence of competition sentences presented at five different signal to noise ratios (+10 SNR, +5 SNR, 0 SNR, −5 SNR, −10 SNR). Starting with quiet conditions, if the child scored five sentences among 12 sentences ( 41.7%) or above, the child progressed to the next condition from +10 SNR to −10 SNR, until the score fail to exceed 41.7%. The final condition that the score passed 41.7% is defined as the final SNR reached by the child. The MPSI test sentences, practice sentences and competition sentences were recorded at the House Ear Institute using the procedures described by Nilsson et al. The sentence set and sentences contained within the set were presented randomly via a loudspeaker by computerized software and were similarly automatically scored. The output of the test sentences was fixed to 65 dB (A).
The test was conducted in a quiet room. The background noise was about 33-35 dB (A). Most of the children could finish the test by themselves. A few children needed their teachers to be present in the test room when they were tested. Testing was initiated with an individual practice session in order to ensure that the child understood the test procedure. After the practice session, individual testing started using quiet conditions. Upon passing each of the SNR-conditions, the competing noise levels were increased. The test sentences were played from the loudspeaker in front of the child and the competition sentences were played from the loudspeaker behind the child. Calibration was administered before test, allowing the output levels of the loudspeakers to be calibrated at the position of the center of the child's head. If at any time the child became uncooperative or expressed distress, testing was discontinued and incomplete results were discarded. Each child was tested only once, and completed resulted were recorded for analysis.
SPSS 11.5 and Microsoft Excel software were used for all statistical analyses. No parametric analysis was used to analyze the statistical differences and correlations. Results are expressed as mean ± standard deviation (mean ± STD).
| Results|| |
Mandarin pediatric speech intelligibility results
The mean SNR score and standard deviation of MPSI results from children in the 2 year, 3 year, 4 year and 5 year age groups are described in [Table 2]. No statistical difference between boys and girls was observed in the MPSI final SNR scores (P = 0.672). The average SNR score distribution of groups 4 years and 5 years are described in [Figure 1], which shows that the average SNR scores were skewed from the normal distribution. The non-parametric Kruskal-Wallis test was used to compare the mean SNR scores of MPSIs between the 2 years, 3 years, 4 years and 5 years age groups. Statistical differences in the mean score between the four age groups was demonstrated (P = 0.00). The MPSI mean score of the 2 years group was similar to the mean score of the 3 years age group (2 years: 1.17 SNR, 3 years: −0.97 SNR). Similarly, the MPSI mean score of the 4 years group was close to the mean score of the 5 years group (4 years: −4.84 SNR, 5 years: −5.39 SNR).
|Figure 1: Number of children distributions at final signal to noise ratios (SNR) scores in the combined 4 year and 5 year age groups. X-axis indicates final SNR score in the mandarin pediatric speech intelligibility test. Y-axis means numbers of children that reached the score. The curve demonstrated the normal distribution of the scores|
Click here to view
|Table 2: Mandarin pediatric speech intelligibility test results for kindergarten aged subjects|
Click here to view
To explore whether the statistical differences observed in the mean score between the four groups was caused by improvement in speech perception ability in noisy conditions most dramatically between the ages of 3 and 4 years, as the results initially suggested, the data from the 2 years and 3 years groups were combined. Similarly, data from the 4 years and 5 years groups were combined. A non-parametric Mann-Whitney test was used to compare the mean scores of MPSI between the combined 2 years and 3 years group and the combined 4 years and 5 years group. The Mann-Whitney U value was calculated to be 1719.00. Statistical differences in the mean score between the two combined groups were demonstrated (P = 0.00). Non-parametric correlation analysis, specifically Spearman correlation, was performed to examine the relationship between speech perception ability in the presence of competing noise and age. The Spearman correlation coefficient was − 0.522 (P = 0.00). Correlation was determined to be significant at the P = 0.01 level using a two-tailed approach. The results of these analyses demonstrated that the speech perception ability of children in the presence of competing noise was significantly, but weakly, related to increasing age. This may be caused by the similar mean scores observed in both the 2 years and 3 years groups or the 4 years and 5 years groups.
Average performance as a function of the different SNR for MPSI in quiet and competing speech message conditions (+10 SNR, +5 SNR, 0 SNR, −5 SNR, −10 SNR) in 2 years, 3 years, 4 years and 5 years groups are shown in [Figure 2]. In the 4 years and 5 years groups, the average speech intelligibility scores were 100% correct in quiet conditions, reaching almost 100% in the presence of + 10 SNR, +5 SNR competing messages. The average speech intelligibility scores declined gradually when the SNR reached 0 SNR in both the 4 years and the 5 years groups. No ceiling effect was observed in the four groups at the most difficult test condition (−10 SNR). The best average speech intelligibility score at −10 SNR was observed, as expected, in the 5 years group, obtaining a rate of 59.51% correct. These results indicate that normally hearing 5-year-old children can recognize more than half of the target sentences when the competition sentences were at −10 SNR. For the 2 years and 3 years groups, the average speech intelligibility scores failed to reach 100% in quiet conditions, and dropped with increasing competing noise. The MPSI performance levels observed in the 2 years and 3 years groups were similar at both −5 SNR and −10 SNR. The average speech intelligibility scores were closest in value between the 2 years and 3 years groups at −10 SNR.
|Figure 2: Performance-intensity function for mandarin pediatric speech intelligibility in quiet conditions and in the presence of competing speech messages (10 SNR, 5 SNR, 0 SNR, −5 SNR, −10 SNR) in the four experimental age groups, showing a decline in speech recognition success rate with increasing competing noise. No ceiling floor was observed in the four groups signal to noise ratios (SNR)|
Click here to view
To explore the change trend of the speech intelligibility score with changing of competing speech message conditions, repeated measure analysis of variance (ANOVA) was used. No statistic difference was found in the change trend of speech intelligibility scores between the 2 years group and the 3 years group (P = 0.12). As a result, the data from the 2 years and 3 years groups were combined. Statistical differences in the change trend of speech intelligibility scores with changing of competing speech message conditions among the three groups were demonstrated (P = 0.00).
| Discussion|| |
The speech perception ability of normally hearing children was shown to development gradually from 2 years old to 5 years old in a non-linear fashion, exhibiting the largest stepwise increase between ages 3 and 4 years, according to MPSI testing. While 2-year-old children were able to recognize target sentences from competitive sentences at low competing noise conditions (SNR > 0), these younger children demonstrated greater difficulty in processing speech at elevated competing noise levels. The 3-year-old children were able to recognize target sentences from competitive sentences even at increased levels of competing noise (SNR < 0), although with a lower correct rate than children aged 4 and 5 years. Statistical difference between mean SNR scores in MPSI was demonstrated between the ages of 3 and 4 years. This may be due to the cognitive development. But, this was not observed among children aged 4 and 5 years. More research could be devoted to exploring whether 3-4 years of age is the critical period for normally hearing children to develop speech perception ability in the presence of competing noise, an important skill in everyday conditions where competing noise is common.
The first few years of life, particularly from ages 3 to 4 years, may lead to accelerated growth of speech perception in daily conditions; however, many hearing impaired children may experience the most difficulty during this period. These findings confirm previous results, indicating that ages 3-4 years are crucial for development in normally hearing children and intervention in hearing impaired children by Nishi et al.  Speech perception in the presence of noise in children of ages 4-9 years was reported previously by Nishi et al., wherein the signal-to-noise ratios (SNR: 0, +5, +10 dB) were tested. An improved performance as a function of age was observed, and SNR improved at a similar rate for each age group. Nishi et al. used 15 American English consonants as testing materials presented in a background of speech-shaped noises. Unlike the present study, Nishi et al. failed to observe accelerated growth in the younger groups, likely due to the use of older subjects. At the hardest MPSI test (−10 SNR), more than half of the 5-year-old children (57%) achieved correct rates up to −10 SNR, indicating that 5-year-old children demonstrate good speech perception abilities in noisy environments. Not all the 5-year-old group subjects reached the ceiling at −10 SNR however indicating that future studies will be necessary to examine children 6 years and older using MPSI testing in the future to determine the exact progression of speech recognition development. While such studies are commonplace in English-speaking children, few studies have examined these findings in Mandarin-speaking children, partially due to the difficulty in obtaining and applying testing materials prior to the development of the MPSI testing system.
Gender was not shown to impact the development of speech perception in noise. The steepness of the PI Performance-Intensity (PI) function in our research appears, instead, to be more gradual than the steepness of previous PSI results.  This may be caused by variance in test materials and methods, or by language distinctions. The test sentences used in MPSI are different than those of PSI. Also, the SNR is decreased from +10 SNR to −10 SNR in MPSI testing. According to a report by Jerger, the message-to-competition ratio was constant at 0 dB for correct identification of target sentences. The target sentence speaker was placed in front of the child and the competition sentence speaker was placed behind the child during MPSI testing, a condition differing from Jerger's report, in which a loud speaker was placed opposite the test ear or the children were tested using earphones. This resulted in a PI function obtained on only one ear of each subject.  This difference in procedure may have made the PSI test, when performed by this method, more difficult than the currently applied MPSI, and thus more sensitive to small fluctuations. This is a likely cause for the observed steepness of PSI-PI functions compared with those of MPSI-PI functions. The steepness of the PI function demonstrated a statistical difference among the three groups in our research. The difference existed at difficult competing noise conditions (0,−5,−10 SNR) according to [Figure 2]. Fluctuation of MPSI-PI function was observed in the 4 years and 5 years groups. The result demonstrated that children vary in speech perception ability in difficult competing noise conditions in the early stages. The result, in this study, was agreement with the Brown's work that children vary in the rate at which their language develops, especially in the early stages. 
Several studies have suggested that fitting HA or implanting cochlear implants (CI) early is important for language rehabilitation in hearing impaired children. ,, The current results indicate that intervention using these devices is most critical during or before the ages of 3-4 years, ages at which normally hearing children exhibit the most dramatic acceleration of growth in speech perception, particularly in everyday environments containing competing noises. According to our research, hearing impaired children may experience improved outcomes in speech perception ability in noisy conditions if fitted HA or implanted CI at ages younger than 4 years, providing the maximum opportunity to achieve normal levels of speech recognition.
| References|| |
|1.||Zheng Y, Soli SD, Wang K, Meng J, Meng Z, Xu K, et al. Development of the Mandarin pediatric speech intelligibility (MPSI) test. Int J Audiol 2009;48:718-28. |
|2.||Owens RE. Language development: An introduction. Columbus: Merrill; 1984. p. 51. |
|3.||Aslin RN, Smith LB. Perceptual development. Annu Rev Psychol 1988;39:435-73. |
|4.||Carney AE. Audition and the development of oral communication competency. In: Bess FH, Gravel JS, Tharpe AM, editors. Nashville: Bill Wilkerson Center Press; 1996. p. 121. |
|5.||Bess FH, Gravel J, Tharpe AM. Amplification for children with auditory deficits. Nashville, TN: Bill Wilkerson Center Press; 1996. p. 29-54. |
|6.||Eisenberg LS. Current state of knowledge: Speech recognition and production in children with hearing impairment. Ear Hear 2007;28:766-72. |
|7.||Boothroyd A. Developmental factors in speech recognition. Int J Audiol 1970;9:30-8. |
|8.||Elliott LL. Performance of children aged 9 to 17 years on a test of speech intelligibility in noise using sentence material with controlled word predictability. J Acoust Soc Am 1979;66:651-3. |
|9.||Fior R. Physiological maturation of auditory function between 3 and 13 years of age. Audiology 1972;11:317-21. |
|10.||Hnath-Chisolm TE, Laipply E, Boothroyd A. Age-related changes on a children's test of sensory-level speech perception capacity. J Speech Lang Hear Res 1998;41:94-106. |
|11.||Jerger S, Jerger J. Pediatric speech intelligibility test. St. Louis: Auditec of St. Louis; 1984. |
|12.||Wang NY, Eisenberg LS, Johnson KC, Fink NE, Tobey EA, Quittner AL, et al. Tracking development of speech recognition: Longitudinal data from hierarchical assessments in the Childhood Development after Cochlear Implantation Study. Otol Neurotol 2008;29:240-5. |
|13.||Quintas VG, Attoni TM, Keske-Soares M, Mezzomo CL. Auditory processing in children with normal and disordered speech. Braz J Otorhinolaryngol 2010;76:718-22. |
|14.||Nilsson M, Soli SD, Sullivan JA. Development of the Hearing in Noise Test for the measurement of speech reception thresholds in quiet and in noise. J Acoust Soc Am 1994;95:1085-99. |
|15.||Nishi K, Lewis DE, Hoover BM, Choi S, Stelmachowicz PG. Children's recognition of American English consonants in noise. J Acoust Soc Am 2010;127:3177-88. |
|16.||Jerger S, Jerger J. Pediatric speech intelligibility test: Performance-intensity characteristics. Ear Hear 1982;3:325-34. |
|17.||Brown R. Five aspects of sentence construction. A first language: The early stages. Cambridge, MA: The MIT press; 1973. p. 6. |
|18.||Connor CM, Craig HK, Raudenbush SW, Heavner K, Zwolan TA. 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;27:628-44. |
|19.||Manrique M, Cervera-Paz FJ, Huarte A, Molina M. Advantages of cochlear implantation in prelingual deaf children before 2 years of age when compared with later implantation. Laryngoscope 2004;114:1462-9. |
|20.||Robbins AM, Koch DB, Osberger MJ, Zimmerman-Phillips S, Kishon-Rabin L. Effect of age at cochlear implantation on auditory skill development in infants and toddlers. Arch Otolaryngol Head Neck Surg 2004;130:570-4. |
No. 37, GuoXueXiang, Chengdu, 610041, Sichuan
People's Republic of China
Source of Support: None, Conflict of Interest: None
[Figure 1], [Figure 2]
[Table 1], [Table 2]