Effect of Age and Speech Perception at Different Signal-to-Noise Ratios Among Children With Normal Hearing

Shanjana Kumar; Dawson Gladys Prathiba

doi:10.7874/jao.2025.00080

Abstract

Background and Objectives

Children constantly gain auditory experiences that promote speech and language development. Therefore, evaluating the auditory capacities of children in noisy environments is imperative. In this study, we sought to evaluate the speech perception abilities of children with normal hearing in noisy environments.

Subjects and Methods

A cross-sectional study design was employed, enrolling 162 Tamil-speaking children aged 3-6 years with normal hearing. These children were further categorized into three groups based on age range: group 1 (3;0-3;11 years), group 2 (4;0-4;11 years), and group 3 (5;0-5;11 years). The two-talker babble was recorded and presented under four conditions: silence, 0 dB signal-to-noise ratio (SNR), +10 dB SNR, and +20 dB SNR. Scoring was carried out and recorded appropriately.

Results

A significant difference (p<0.05) in performance was observed at 0 dB SNR and +10 dB SNR compared to that in the other two presentation conditions. Among all the age groups assessed, performance was poorer in those exposed to 0 dB SNR, whereas near-perfect scores were recorded among those exposed to 20 dB SNR.

Conclusions

Our findings provide a basis for comparing the performance of children utilizing hearing devices in noisy environments. This normative data can also serve as a benchmark for evaluating the performance of children with hearing loss fitted with hearing devices.

Keywords: Speech in noise test · Children with normal hearing · Signal-to-noise ratio · Speech perception.

Introduction

Speech constitutes a complex acoustic signal characterized by both spectral and temporal features [1]. Individuals across various age groups encounter difficulties in communicating within noisy environments. Speech components necessitate detailed neural representation of temporal information, particularly in background noise. The auditory system is required to distinguish between noise and speech for this task, which can be evaluated through the speech-in-noise (SIN) test [2]. Speech perception refers to the capacity of an individual to understand and differentiate among sounds, words, and syllables for effective communication [3].

Research indicates that children with normal hearing exhibit effective localization skills in the presence of background noise, attributed to their use of auditory cues such as interaural time difference, interaural level difference, monaural spectral cues, and signal envelopes [4]. Stuart [5] indicated that school-age children exhibit inherently lower central processing efficiency, rather than reduced temporal resolution. Younger children may experience overall auditory perception impairment due to underdeveloped neural systems that have difficulty interpreting SIN, resulting in increased internal noise and decreased intrinsic attention [6]. Children require a higher signal-to-noise ratio (SNR) than adults for effective speech signal processing in noisy environments [7].

SIN tests are utilized to evaluate listening ability in noisy environments for both adults and children, with results determining the realistic expectations of an individual’s listening capacity. Cognitive factors, including memory, fatigue, and attention, affect speech perception in noisy environments [8]. Developmental changes significantly influence individuals’ perception of speech in noisy environments. Wilson, et al. [9] found that children’s performance was lower than that of adults in noisy environments because of delayed maturity, manifested as neurophysiologic problems, listening strategy problems, or a combination of both.

Crandell [10] and Hall, et al. [11] indicated that children experience greater challenges in comprehending speech in noisy environments compared to adults until the age of 12 years. It was observed that SIN test performance improves with increasing age and consequently, a higher SNR is necessary for children aged 11 to 14 years to achieve performance levels comparable to adults [12,13]. Elliott and Katz [14] investigated speech perception in noise by presenting continuous and interrupted noises at three different SNRs. They found that children performed 7% worse than adults in quiet conditions, with performance improving from 15% to 18% in noisy conditions. Hnath-Chisolm, et al. [15] found that the age effect was predominantly observed in children under 7 years old. These studies indicate that as age increases, there is an improvement in auditory and listening tasks.

This study aimed to establish age-based norms for the SIN test in Tamil, utilizing various SNRs of +20 dB, +10 dB, and 0 dB for children with normal hearing. The current research examined the impact of age on children aged 3 to 6 years under varying noise conditions. Children aged 3 to 6 years are in a crucial developmental stage for acquiring language and knowledge [16]. The data may be used as a benchmark for comparison of test performances with children using hearing devices. The results can inform recommendations for the fine-tuning of hearing devices, cochlear implant (CI) mapping, rehabilitation, and follow-up.

Subjects and Methods

A cross-sectional study design was used, and this study was approved by the Institutional Ethics Committee of Sri Ramachandra Institute of Higher Education and Research (Ref no. CSF/23/FEB/123/150). The testing was conducted in three schools with the appropriate authorization from the schools. A total of 162 Tamil-speaking children aged 3 to 6 years participated in the study. Table 1 indicates the characteristics of the study participants. The sample size was calculated based on a study conducted by Wilson, et al. [9], yielding a required sample size of 164, which was adjusted to 162 for this study.

The study included children with normal hearing sensitivity (<25 dB HL) aged 3 to 6 years, whose first language was Tamil, and who had no history of otological, neurological, or cognitive problems. The participants were classified into three age groups: group 1 (3;0–3;11 years), group 2 (4;0–4;11 years), and group 3 (5;0–5;11 years). The study was carried out in two phases. In Phase 1, the SIN test in Tamil was developed. In Phase 2, two conditions were included: condition 1, conducted in quiet; and condition 2, conducted at three different SNRs (0, +10, and +20 dB).

Phase 1: Development of speech test in noisy environments for children

The assessment included Tamil vocabulary in the context of two-talker babble (TTB) words featuring a consonant–vowel–consonant–vowel (CVCV) structure that were sourced from Tamil textbooks, storybooks, and everyday conversations. The stimuli comprised 50 words [17]. Four lists, designated as A, B, A1, and B1, each containing 25 words, were generated. Two half-lists, A1 and B1, were generated using the existing 50 words to ensure phonetic balance. The TTB in Tamil was developed as part of the study. An adult Tamil-speaking male and female recorded the TTB while reading a standardized Tamil passage developed by Subramaniyan [18]. The recording was carried out in a studio setting, using a professional microphone positioned 1 m from the speaker. The sampling frequency was established at 44,000 Hz. Two independent recordings of each speaker were digitally combined to create TTB. Noise was allocated to one track, while word material was assigned to the other track. Using Adobe Audition, the following SNRs were generated: 0 dB, +10 dB, and +20 dB. In the 0 dB SNR condition, list B1 was used in the right ear, while list A1 was used in the left ear. In the +10 dB SNR condition, list A was administered to the right ear, while list B was administered to the left ear. In the +20 dB SNR condition, list B was utilized in the right ear, while list A was used in the left ear. The stimuli underwent a goodness test involving 20 children with normal hearing to assess their efficacy.

Phase 2: Testing procedure

Parents were informed of the test descriptions and procedures prior to testing, and informed consent was obtained from them. Before the testing, an otoscopy examination was conducted to assess for any middle ear infections, and relevant history was gathered from both the teacher and the parent. Hearing screening was carried out, air conduction thresholds were measured at frequencies of 500 Hz, 1 kHz, 2 kHz, and 4 kHz. Participants with hearing thresholds below 25 dB HL were selected for the study. Testing was conducted in a quiet room, with ambient noise levels maintained within the permissible limits as specified by the American National Standards Institute [19]. The noise floor in the room was assessed using a sound level meter (SLM). The stimulus was delivered through a calibrated Xiaomi Notebook Ultra laptop and Sennheiser HD 206 headphones. Speech stimuli were presented at 70 dB SPL. The procedure was carried out under two conditions. In condition 1, the words were presented in a quiet environment. Children were directed to repeat the specified target words, and the scores were recorded in the response sheet. In condition 2, the words were presented in TTB at three distinct SNRs: 0 dB, +10 dB, and +20 dB. The stimuli were presented in the following order: quiet, +20 dB SNR, +10 dB SNR, and 0 dB SNR. The ear tested was randomized for each participant.

Children were directed to concentrate and reiterate the specified target words while ignoring the persistent background noise. Breaks were provided during the testing to sustain participant attention. A score of 1 was given for each correct response, while a score of 0 was given for each incorrect response. The maximum possible score for each condition list is 25. The total scores obtained from the children under each condition were tabulated, and the data were analyzed statistically.

Statistical analyses

A test of normality was done, and it was observed that the values did not follow normal distribution; therefore, non-parametric test was used to analyze the data. Kruskal–Wallis test was used to find the overall effect in all four noise conditions. The least significance difference (LSD) post hoc test was used for group comparisons. Cohen’s d was used to calculate the effect size. Statistical analyses were performed using IBM SPSS Statistics, version 23.0 (IBM Corp.).

Results

The study aimed to assess speech perception performance in noise among Tamil-speaking children with normal hearing, aged 3 to 6 years. The scores obtained from 162 children with normal hearing indicated an upward trend between age and SIN performance (group 1). The median scores and interquartile ranges (IQR) across groups were: group 1 (3;0–3;11 years), median 23.6 (IQR 13.50–25.00); group 2 (4;0–4;11 years), median 23.9 (IQR 15.50–25.00); and group 3 (5;0–5;11 years), median 24.3 (IQR 15.50–25.00). The mean raw scores of the SIN test across four conditions revealed that the oldest group (group 3) achieved full scores for +20 dB SNR, in contrast to the other three conditions (Fig. 1). The Kruskal–Wallis test was employed to examine the impact of noise across various listening conditions, revealing a significant overall effect in all four noise conditions (p<0.01). A significant difference (p<0.01) in performance was observed at 0 dB SNR and +10 dB SNR when compared to the other two listening conditions. It can be concluded that children experienced difficulty in identifying the signal at 0 dB SNR, followed by the +10 dB SNR listening condition. All age groups exhibited poor scores in the 0 dB SNR listening condition. The youngest age group (group 1, ages 3 to 3;11 years) achieved the lowest mean score of 15.25 out of 25 in the 0 dB SNR listening condition.

Group comparison was conducted using the LSD post hoc test to analyze age and listening conditions. A within-group comparison was carried out under four different listening conditions. A significant difference was observed between age groups 1 and 3. For age groups 1 and 2 under 0 dB SNR and +10 dB SNR listening conditions, Cohen’s effect size was calculated to assess the magnitude of impact for the p value at 0 dB SNR (d=0.62) and +10 dB SNR (d=0.41). The effect size was found to be moderate for both groups across the two noise conditions. No significant statistical differences were observed in the other two listening conditions.

It was found that there was no statistically significant difference in any of the SNR conditions when comparing the quiet condition across the three age groups. Group comparisons revealed a significant difference in performance at both 0 dB SNR and 10 dB SNR (p<0.01) between groups 1 and 2, as well as between groups 1 and 3. Children in the older age group (group 3) demonstrated superior performance across all listening conditions. No difference was observed in the +20 dB SNR listening condition. The near perfect scores at +20 dB SNR suggest that all children require a higher SNR to enhance speech recognition in noisy environments. The performance in noisy conditions exhibited an age-related trend, indicating that as age increased, listening ability in noise also improved (Fig. 2).

Discussion

All 162 children achieved near-perfect scores in quiet conditions; however, the highest performance was observed in the +20 dB SNR compared to the other two noise conditions. The performance was comparable across all age groups for both the quiet and +20 dB SNR listening conditions. SIN recognition is influenced by age effects, with findings from a study by Zheng, et al. [20] indicating that children aged 3 to 6 years exhibited comparable performance in quiet, +5 dB SNR, and +10 dB SNR conditions. The youngest age group (3;0–3;11 years) exhibited the lowest performance compared to the other two age groups. The authors reported corresponding findings, indicating that scores decreased within each age group as test conditions became increasingly challenging.

This may be attributed to integration of linguistic information between the left and right hemispheres of the brain is influenced by the myelination of the corpus callosum, which continues until approximately 15–20 years of age [21]. Exposure to language and phonemes leads to brain organization beginning in childhood, indicating that SIN recognition improves with increased language exposure [22]. The prefrontal cortex, responsible for speech perception, continues to develop throughout childhood [23]. Significant developmental changes in cortical structures may lead older children to utilize sensory and perceptual information differently than younger children, thereby enhancing speech and noise recognition. The neurophysiological changes may underlie the poor performance of young children in low SNRs. Younger children require higher sound levels and frequencies compared to older children to achieve equivalent hearing, as their sensory systems are not fully developed due to neuronal maturation [24]. The current study demonstrates that young children require a quiet environment for learning, as this period is critical for overall development.

Goldsworthy and Markle [25] found that speech recognition develops earlier in the presence of a single-talker masker compared to a two-talker masker during childhood. Speech recognition in children with normal hearing was superior in the presence of time-varying background noise compared to speech-spectrum noise. The current study indicates that children in the youngest age group (3;0–3;11 years) exhibited poor performance on TTB at low SNR levels. Low scores in the youngest age group may be attributed to cognitive factors [26]. Cognitive factors, including memory, fatigue, and attention, affect speech perception in noisy environments [8]. The data indicates that children with normal hearing also require a higher SNR to effectively learn and communicate in environments with competing sounds. Considering the influence of background noise on children’s perception abilities is essential when planning and implementing educational infrastructures. Creating a quiet environment and implementing measures to enhance speech intelligibility, such as amplification systems or noise-reducing strategies, are essential for facilitating optimal communication and language development in children of this age group.

In conclusion, this study highlights that ability of children to listen in noisy environments improves with age, requiring higher SRNs for effective listening and spoken language development. The data from children with normal hearing serves as a reference benchmark for comparing the performance of children with hearing impairment using amplification and other hearing devices. The findings can be applied to make informed clinical decision-making regarding choice of devices, optimization of outcomes, selection of suitable aural rehabilitation strategies, and educational settings for children with hearing loss.

Notes

Conflicts of Interest

The authors have no financial conflicts of interest.

Author Contributions

Conceptualization: Dawson Gladys Prathiba. Data curation: Shanjana Kumar. Formal analysis: Shanjana Kumar. Methodology: Dawson Gladys Prathiba. Supervision: Dawson Gladys Prathiba. Validation: Shanjana Kumar. Writing—original draft: Shanjana Kumar. Writing—review & editing: Shanjana Kumar, Dawson Gladys Prathiba. Approval of final manuscript: Shanjana Kumar, Dawson Gladys Prathiba.

Funding Statement

None

Acknowledgments

None

Fig. 1.

Mean speech-in-noise (SIN) scores in Tamil-speaking children with normal hearing across three listening conditions. Group 1 included children aged 3;0–3;11 years, group 2 included those aged 4;0–4;11 years, and group 3 included those aged 5;0–5;11 years. SNR, signal-to-noise ratio.

Fig. 2.

Mean speech-in-noise (SIN) scores in Tamil-speaking children with normal hearing for each age group across all listening conditions. Group 1 included children aged 3;0–3;11 years, group 2 included those aged 4;0–4;11 years, and group 3 included those aged 5;0–5;11 years. SNR, signal-to-noise ratio.

Table 1.

Demographic characteristics of participants by age group and sex

Age group (yr)	Sex		Total
Age group (yr)	Male	Female	Total
3;0-3;11	38	18	56
4;0-4;11	35	18	53
5;0-5;11	35	18	53
Total	108	54	162

REFERENCES

1. Shannon RV, Zeng FG, Kamath V, Wygonski J, Ekelid M. Speech recognition with primarily temporal cues. Science 1995;270:303–4.

2. Cory P, Bell B. Effective use of speech-in-noise testing in the clinic. Hear J 2019;72:40, 42–3.

3. Ren C, Yang J, Zha D, Lin Y, Liu H, Kong Y, et al. Spoken word recognition in noise in Mandarin-speaking pediatric cochlear implant users. Int J Pediatr Otorhinolaryngol 2018;113:124–30.

4. Kerber S, Seeber BU. Sound localization in noise by normal-hearing listeners and cochlear implant users. Ear Hear 2012;33:445–57.

5. Stuart A. Reception thresholds for sentences in quiet, continuous noise, and interrupted noise in school-age children. J Am Acad Audiol 2008;19:135–46.

6. Anderson S, White-Schwoch T, Parbery-Clark A, Kraus N. A dynamic auditory-cognitive system supports speech-in-noise perception in older adults. Hear Res 2013;300:18–32.

7. Corbin NE, Bonino AY, Buss E, Leibold LJ. Development of open-set word recognition in children: speech-shaped noise and two-talker speech maskers. Ear Hear 2016;37:55–63.

8. Wightman FL, Callahan MR, Lutfi RA, Kistler DJ, Oh E. Children’s detection of pure-tone signals: Informational masking with contralateral maskers. J Acoust Soc Am 2003;113:3297–305.

9. Wilson RH, Farmer NM, Gandhi A, Shelburne E, Weaver J. Normative data for the words-in-noise test for 6- to 12-year-old children. J Speech Lang Hear Res 2010;53:1111–21.

10. Crandell CC. Speech recognition in noise by children with minimal degrees of sensorineural hearing loss. Ear Hear 1993;14:210–6.

11. Hall JW 3rd, Grose JH, Buss E, Dev MB. Spondee recognition in a two-talker masker and a speech-shaped noise masker in adults and children. Ear Hear 2002;23:159–65.

12. Buss E, Leibold LJ, Hall JW 3rd. Effect of response context and masker type on word recognition in school-age children and adults. J Acoust Soc Am 2016;140:968–77.

13. Hall JW, Buss E, Grose JH, Roush PA. Effects of age and hearing impairment on the ability to benefit from temporal and spectral modulation. Ear Hear 2012;33:340–8.

14. Elliott L, Katz D. Development of a new children’s test of speech discrimination. St. Louis: Auditec;1980.

15. 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.

16. Nittrouer S, Pennington B. New approaches to the study of childhood language disorders. Curr Dir Psychol Sci 2010;19:308–13.

17. Prathiba DG, Arunachalam AR, Boominathan P. Performance of a closed set picture identification test in tamil for children with cochlear implant. Indian J Otolaryngol Head Neck Surg 2022;74:3794–99.

18. Subramaniyan B. Development of a standardized reading passage in Tamil [dissertation]. Manipal: Manipal Univ.;2005.

19. American National Standards Institute. American National Standard: maximum permissible ambient noise levels for audiometric test rooms. ANSI S3.1-1999. New York: Acoustical Society of America; 1999.

20. 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.

21. Boothroyd A. Auditory development of the hearing child. Scan Audiol 1997;26(Suppl 46):9–16.

22. Rauschecker JP. Auditory cortical plasticity: a comparison with other sensory systems. Trends Neurosci 1999;22:74–80.

23. Vuontela V, Steenari MR, Carlson S, Koivisto J, Fjällberg M, Aronen ET. Audiospatial and visuospatial working memory in 6-13 year old school children. Learn Mem 2003;10:74–81.

24. Eisenberg LS, Shannon RV, Martinez AS, Wygonski J, Boothroyd A. Speech recognition with reduced spectral cues as a function of age. J Acoust Soc Am 2000;107:2704–10.

25. Goldsworthy RL, Markle KL. Pediatric hearing loss and speech recognition in quiet and in different types of background noise. J Speech Lang Hear Res 2019;62:758–67.

26. Buss E, Leibold LJ, Porter HL, Grose JH. Speech recognition in one- and two-talker maskers in school-age children and adults: development of perceptual masking and glimpsing. J Acoust Soc Am 2017;141:2650–60.