Binaural processing and phonological awareness in Australian Indigenous children from the Northern Territory: A community based study

International Journal of Pediatric Otorhinolaryngology 128 (2020) 109702

Contents lists available at ScienceDirect

International Journal of Pediatric Otorhinolaryngology journal homepage: www.elsevier.com/locate/ijporl

Binaural processing and phonological awareness in Australian Indigenous children from the Northern Territory: A community based study

T

Mridula Sharmaa,b,∗, Gillian Wigglesworthc,d, Gemma Savagea, Katherine Demutha,b,e a

Department of Linguistics, Macquarie University, Sydney, Australia HEARing Co-operative Research Centre, Melbourne and Sydney, Australia c School of Languages and Linguistics, University of Melbourne, Parkville, Australia d ARC Centre for Excellence for the Dynamics of Language, University of Melbourne, Parkville, Australia e ARC Centre for Excellence in Cognition and Its Disorders, Macquarie University, Sydney, Australia b

ARTICLE INFO

ABSTRACT

Keywords: Indigenous Australian Aboriginal hearing Otitis media Phonological awareness Dichotic listening

Objective: Research has found that otitis media (OM) is highly prevalent in Australian Indigenous children, and repeated bouts of OM is often associated with minimal-to-moderate hearing loss. However, what is not yet clear is the extent to which OM with hearing loss impacts auditory signal processing specifically, but also binaural listening, listening in noise, and the potential impact on phonological awareness (PA) – an important, emergent literacy skill. The goal of this study was to determine whether auditory abilities, especially binaural processing, were associated with PA in children from populations with a high incidence of OM, living in a remote Australian Indigenous community in the Northern Territory (NT). Methods: Forty-seven 5-12-year-olds from a bilingual school participated in the study. All were tested to determine hearing sensitivity (pure tone audiometry and tympanometry), with PA measured on a test specifically developed in the first language of the children. OM often results in a hearing loss that can affect binaural processing: the Dichotic Digit difference Test (DDdT) was used to evaluate the children's dichotic listening and the Listening in Spatialized Noise-sentences test (LiSN-S) was used to evaluate their abilities to listen to speechin-noise. Results: Seventeen (36%) and 16 (34%) had compromised middle ear compliance (combined Type-B and –C) in the right and left ear respectively. Six children demonstrated a bilateral mild hearing loss, and another five children demonstrated a unilateral mild hearing loss. Thirty-one children were able to complete the DDdT listening task, whereas only 24 completed the speech in noise task (LiSN-S). Forty-four children (94%) were able to complete the letter identification subtask, comprising part of the PA task. The findings revealed that age was significantly correlated with all tasks such that the older children performed better across the board. Once hearing thresholds were controlled for, PA also correlated significantly with both binaural processing tasks of dichotic listening (r = 0.59, p < 0.001) and listening to speech in noise (r = −0.56, p = 0.005); indicating a potential association between early, emergent literacy and listening skills. Conclusions: The significant correlations between phonological awareness and dichotic listening as well as phonological awareness with listening to speech-in-noise skills suggests auditory processing, rather than hearing thresholds per se, are associated to phonological awareness abilities of this cohort of children. This suggests that the ability to process the auditory signal is critical.

Abbreviations: LiSN-S, Listening in spatialized Noise-Sentences; NT, Northern Territory (Australia); OM, otitis media; OME, otitis media with effusion; PTA, Pure Tone Audiogram; (DDT), Dichotic Digits Test; (DDdT), Dichotic Digits difference Test ∗ Corresponding author. Audiology Program, Department of Linguistics, Macquarie University, Room 1.609, S2.6, NSW, 2109, Australia. E-mail addresses: [email protected] (M. Sharma), [email protected] (G. Wigglesworth), [email protected] (G. Savage), [email protected] (K. Demuth). https://doi.org/10.1016/j.ijporl.2019.109702 Received 3 May 2019; Received in revised form 28 September 2019; Accepted 30 September 2019 Available online 05 October 2019 0165-5876/ Crown Copyright © 2019 Published by Elsevier B.V. All rights reserved.

International Journal of Pediatric Otorhinolaryngology 128 (2020) 109702

M. Sharma, et al.

1. Introduction

A high incidence of middle ear disorders places young children at higher risk of having poor auditory processing [1] as well as poor prereading skills [2]. This is partly due to the common incidence of middle ear disorders occurring during early childhood and developmental timing associated with expressive language and speech. Animal research has shown that middle ear disorders leading to temporary hearing loss can contribute to structural and functional changes in the auditory pathway [3]. It is also suggested that auditory deprivation, a consequence of middle ear disorder, results in a compromised or degraded auditory signal [4]. This degraded signal is speculated to formulate an atypical phonological representation that impacts binaural processing – including speech recognition, and phonological awareness [5]. Therefore, the current research aims to determine associations between auditory and phonological skills in children who experience a high incidence of middle ear disorders. Otitis media (OM) leads to middle ear disorders and can present as serous fluid (effusion). The prolonged presence of fluid can ultimately become viscous (glue ear), or suppurative (active infection) [6,7]. OM is pervasive among children (6–30 months) from remote Aboriginal communities in Australia [8]. In Indigenous Australians, 90% of children below the age of 3 have been reported to have some form of OM, and 24% have been shown to have perforation of the tympanic membrane for one or both ears [9]. Most children in remote Aboriginal communities experience at least one episode of OM and have an associated hearing impairment before they enter school [8,10,11]. They also tend to experience OM earlier than their urban peers, and have more recurrent episodes [6,12]. These episodes typically last longer and result in more complications than for children elsewhere [9,13].

Children with earlier onset and longer duration of OM are reported to have poorer outcomes for recognition of speech in noise, regardless whether they demonstrate normal hearing at the time of testing [27]. One study reported that seven-percent of children in Aboriginal communities in Australia had spatial processing disorder where binaural cues were impacted; despite exhibiting normal hearing thresholds [28]. Other studies have reported similar difficulties with spatial processing tasks in children with a known history of OM [1,29]. Effusion related to OM can cause a delay of low-frequency signals to inner ear by up-to 150 μs [30]. Thus, there exists a body of research supporting the theory that the neural consequences of spatial processing difficulties can persist even after OM is resolved (i.e., hearing sensitivity is technically within normal limits) [31]. However, not all studies concur regarding findings suggestive of speech perception difficulties emerging as a consequence of early childhood OM history [32,33], but Hartley and Moore (2005) used monaural tasks, rather than binaural tasks in their small sample of children with a reported history of OM. In contrast, more recent studies have used binaural tasks such as Masking Level Difference, Dichotic Digit Test (DDT) and Listening in Spatialized Noise (LiSN-S) [1,29]. A study applying a variation of the DDT that used monosyllabic words instead of numbers, found no difference between children with and without, OM [34]. The authors acknowledged that rapid intervention (grommets) for OM in the children assessed might have accounted for the lack of impact on auditory task performance. They also noted that the effect of OM may be most apparent on competitive listening tasks [34]. Therefore, the current study used two tasks of binaural processing, the Dichotic Digit difference Test (DDdT) (a variation of the original DDT), and a speech-in-noise test (LiSN-S). The Dichotic Digit difference task has the advantage of using assessment components that were less language-dependent (digits), compared to the sentence-based LiSN-S repetition task.

1.2. OM, hearing and auditory processing

1.3. Phonological processing, OM and auditory skills

Hearing loss due to OM is typically conductive in nature and often fluctuating. In some cases it can be persistent, resulting in mild-tomoderate conductive hearing loss [14]. Fluid in the middle ear due to OM can cause a conductive hearing loss ranging from 20 to 30 dB [15]. A recent review found that the range of hearing loss for the low-to-mid frequencies (500–2000 Hz) was 18–35 dB HL, but losses at 500 Hz can be as high as 50 dB HL [16]. Fluctuating hearing loss due to OM during the development of the central auditory system in children younger than 3 years has also been connected to atypical auditory processing abilities [4,17–20] and poorer language skills [21]. It is suggested that fluctuations in hearing may cause disruptions in neural connections that can outlast any peripheral consequences, even once OM has resolved [22]. The auditory system, and specifically the brainstem relative to higher processing centres such as the cortex, is highly sensitive to temporal cues. The medial olivary complex (MSO) within the brainstem receives auditory signals from both ears, performing very fine timing analysis (in the order of microseconds). This is where the signals from the two ears are integrated/evaluated for binaural processing, resulting in the interaural timing differences necessary for sound localisation and speech recognition in noise [19]. There are reports of at least two negative types of consequences of this auditory deprivation during the first 3 years of life, both of which involve binaural processing; namely poorer dichotic listening and speech-in-noise abilities [23,24]. Any dysfunction of the brainstem centres, or a compromised signal arriving at the brainstem (especially due to OM), would send an imprecisely-timed signal further along the auditory pathway [19,25]. If there is OM affecting one ear, resulting in asymmetric hearing loss, it can influence binaural processing. This can affect interaural temporal and level difference cues, compromising the sound localisation necessary for listening in noise, as in a classroom [26].

Phonological awareness is the ability to segment and manipulate the sound structure of a language. This ability is widely-considered necessary for learning to read [35–37]. Some have argued there is a relationship between children's performance on phonological awareness tasks during the first years of schooling, and later reading ability [38–42]. Others argue that phonological awareness plays a causal role in the development of literacy skills (e.g., Refs. [43,44]. The link between OM and phonological processing is less well-understood. OM occurs most frequently during the first 3 years of life, peaking at about 6–18 months [9]. To develop phonological awareness, children need to be able to perceive the sounds in the language they are acquiring, and to analyse those sounds into meaningful (phonemic) units; ultimately learning the relationship between graphemes and sounds [45]. This means that where hearing is impaired, the development of phonological awareness may be delayed and/or protracted. Phonological awareness includes different levels of awareness which may involve separating out syllables and phonemes, identifying parts of a syllable (e.g., onset and rhyme), and blending syllables and phonemes to form words [46]. In situations where children suffer from chronic OM, phonological awareness may develop inappropriately because the OM may impact on the child's auditory (and language) development [47]. The 6–18 month period is crucial for language development, a time when children are learning to map sound to meaning, and are acquiring new words [48]. As noted above, the incidence of OM and particularly OM amongst Australian Indigenous children, is considerably higher than in the general population; in both rural and urban settings [9]. Todate, there have been variable results in studies exploring the impact of OM on children's developing phonological awareness. For instance, one study [49] focused on 19 urban, Aboriginal children from Western Australia who were in the second year of primary/elementary school

1.1. Background

2

International Journal of Pediatric Otorhinolaryngology 128 (2020) 109702

M. Sharma, et al.

(around 5–6 years of age). Nine children had recent evidence of OM and 10 acted as controls. Results suggested OM associated with hearing loss during the early school years had a deleterious effect on the development of phonological awareness, reading, and spelling performance during their second year of school. The children with OM scored significantly-lower on subtests of phonological awareness including: rhyme, alliteration, blending, segmentation [49]. Other studies, however, report no link between an early history of OM and subsequent learning or reading deficits [50,51]. A longitudinal study where OM history was prospectively determined and the number of days impacted due to OM was recorded showed no significant relationship between the number of days with OM and subsequent academic performance [52]. A study by Roberts et al. [50] reported no evidence of a significant relationship between a history of OM (or related hearing loss) and children's later reading skills. They also found home environment, interaction, and language stimulation were better predictors of subsequent language and academic success [53]. These findings have been supported by other studies where the link between the OM and academic success was found to be lacking [54,55]. A more recent study by Fougner et al. [56], prospectively gathered systematic information about the number of OM episodes from over 100,000 children. They found no significant relationship with self-reported, later school performance after accounting for: gender, prematurity, breastfeeding until 6 months, maternal smoking in the first 7 years, day-care attendance by 18 months, mother's education level, or socioeconomic status. However, a longitudinal investigation of health, including OM, in 1000 children born in Dunedin, New Zealand, determined that after adjusting for socioeconomic status, children with OM were more inattentive and had deficits in reading ability into their teens [57]. Thus, there is some evidence linking children with repeated OM with poorer phonological awareness [47,58,59], consistent with suggestions that persistent OM leads to poorly-established phonological representations, or “fuzzy” phonological categories [4]. However, none of the above mentioned studies measured dichotic listening and speech in noise along with phonological awareness (PA) in the children with history of OM.

and television. SAE is learned in the school context, which is often English only, although in some cases (as in the community discussed here), the school is bilingual. The current project took place in a remote community in Arnhem Land where one of the major Indigenous languages, Yolngu Matha, remains widely-spoken; consisting of a number of dialects, as well as multiple clan languages [63]. Children initially speak a somewhat simplified version of Yolngu Matha, a koine communilect, resulting from contact between the multiple dialects [64]. Children grow up speaking their mother's dialect, and traditionally transition to speaking their Clan dialect (their father's dialect) during their early teens. There is evidence suggesting that this timeline is shifting, with the transition to Clan dialect now taking place later [65]. In the community in which this project took place, the communilect is spoken by the younger generation regardless of Clan affiliation. The school that the children attend is bilingual, teaching literacy in the communilect in the early years while English is taught orally. In grade 3 or 4 the children transition to English literacy. Literacy materials are developed by a schoolbased Literacy Production Centre, in the form of books and other activities at a range of levels. Overall, children have limited access to preliteracy and literacy materials at home, so this is a focus of the early years of schooling. Most Aboriginal languages have only 3 vowels, lack fricatives (e.g., s, z, f, v) and/or affricates (ch, j) and do not have voicing contrasts (e.g., p vs. b) [60]. 2. Materials and methods 2.1. Participants Participants were child speakers of the local communilect who were learning English in the bilingual program at the school, whose parents had consented to their participation, and who agreed on the day to participate in the study. A total of forty-seven children (18 male, 29 female) aged 5–12 years (mean 7.9 years, SD 1.7), who had started learning English orally upon entry to school, participated in the study. The study was carried out with the approval of The Human Research Ethics Committee of the Northern Territory Department of Education, the Menzies School of Health Research, and Macquarie University (Ethics # 5201600743).

1.4. Phonological processing, OM and English as second language Many Indigenous children in Australia speak English as a second language. This raises questions about how they would perform on a PA test in a language they are only beginning to learn. A small study (n = 12; half with a history of OM and half without) conducted in the Tiwi Islands, investigated the effect of the children's first language (Modern Tiwi) on their phoneme discrimination abilities [60]. The authors observed that compared to children whose first language was English (n = 7), children with Modern Tiwi as their first language, had difficulty discriminating across consonants. Although the results did not always achieve statistical significance, the difficulties noted included poorer discrimination between continuant sounds (e.g., l, r, n) versus affricates (e.g., ch, j) and stops (p, t, k etc.). This was more challenging when the children had some level of hearing loss.

2.2. Assessments Due to the remote nature of this community, there were no audiological clinics with sound treated booths. All assessments were conducted within the school in the quietest room possible given the location, via portable audiometric equipment. Noise levels were measured by the Bruel & Kjaer sound level meter Type 2250. Average test room ambient noise ranged between 40 and 50 dB SPL. Noise levels were measured at different times (morning, midday and afternoon) throughout the day in the unoccupied room where testing was conducted. The noise levels were determined at different frequencies using Z-weighting (LZF). Table 1 provides the noise levels for the three timelines, across the 4.5 days of data collection. Biologic calibration was carried out using three of the researchers with confirmed normal hearing sensitivity acting as a reference [66], to

1.5. The context for the present study The linguistic context in Indigenous Australia is hugely diverse. Traditionally, Indigenous people were multilingual, often speaking 4–5 different languages. Today Indigenous people speak a range of languages including: Standard Australian English (SAE), Aboriginal English, varieties of Kriol (novel, mixed/hybrid languages) or a traditional language (see Refs. [61,62] for a more in-depth discussion detailing the linguistic environment in Indigenous Australia). This is the context into which Indigenous children are born, so Standard Australian English is typically not their first language, though children living in remote areas will have some limited access to SAE through the health centre, the shops, organic, SAE communication opportunities

Table 1 Unoccupied Noise levels (LZF) as measured in the school test room during the day, averaged across the 4.5 days.

500 Hz 1 kHz 2 kHz 4 kHz

3

Morning (9:00–9:55AM)

Mid-day (12:00–12:30PM)

Afternoon (2:00–2:35PM)

47.6 44.0 36.6 31.2

54.4 45.4 41.0 37.1

51.2 48.3 42.5 39.0

International Journal of Pediatric Otorhinolaryngology 128 (2020) 109702

M. Sharma, et al.

corroborate noise level findings and their impact on testing. Researcher thresholds measured in an industry-standard sound-treated booth using the audiometer and transducer employed for the study, were compared to thresholds gained within the test environment at the study venue for 500 and 1000 Hz. The full test battery of hearing tests and the phonological awareness test were administered to each child on the same day, taking approximately 1 h to complete.

Condition one consisted of dichotic listening, where two pairs of prerecorded numbers (four different numbers total) were presented to each ear, causing the numbers to temporally overlap. The participant needed to listen with both ears to hear all four numbers correctly. Condition two involved diotic listening, where four different digits were presented to both ears at the same time (no overlap). For both conditions, participants were instructed to “Listen carefully to the numbers arriving at both ears and repeat all the numbers.” A total of 25 trials were completed for each condition. Stimuli were presented at a pre-determined sound level for each individual, set by a word recognition test using spondee words (disyllabic words such as baseball or hotdog) to estimate word threshold level. Digits were presented at 40 dB sensational level above the word threshold level. Dichotic listening involves attention, memory and binaural processing. In contrast, diotic listening only involves attention and memory, as the same numbers are presented to both ears [69]. A comparison of the dichotic and diotic listening scores thus helps identify those at-risk of binaural processing deficits [69].

2.2.1. Otoscopy Otoscopy was performed bilaterally. This was to account for any obvious abnormalities of the ear canal and tympanic membrane, such as: inflammation, redness, discoloration, wax (occluding or non-occluding), perforation of the tympanic membrane and/or foreign objects in the ear canal. 2.2.2. Tympanometry Tympanometry was performed bilaterally to assess middle ear function. An Interacoustics “Titan” Tympanometer was used to generate participant tympanograms. Participants were instructed to sit quietly during assessment. Ear canal volume, peak pressure, and peak compliance were recorded for both ears. Normal middle ear function (a Type-A tympanogram) was classified as peak pressure between −100 and +100 daPa and peak compliance of 0.3–1.4 mmhos [67]. A Type-B tympanogram classification was applied when there was no appreciable change in compliance across the changing pressures from −200 to +200 daPa (i.e., peak compliance of 0.2 mmhos or less, and a lack of a clear, observable peak). A Type-C tympanogram classification was given when the middle ear peak compliance fell within the normal function range, but peak pressure was −100 daPa or lower. Tympanometry results were recorded, and participants were assigned Type-A, -B or –C tympanograms [67]. Acoustic reflexes were attempted however, most children found the sound uncomfortable and the task could not be completed for any participant.

2.2.4.2. Speech perception in noise: Listening in Spatialized Noise – sentences (LiSN-S) test. This test involved listening and repeating English sentences in the presence of noise. Sentences were presented initially at 62 dB SPL, in the presence of two distracter stories presented at a fixed intensity of 55 dB SPL [70]. Distracter stories varied in their position in space (coming from either directly in front of the listener, or at the sides of the listener), and in the vocal quality of the speakers (two same, or different, female speakers telling two different stories). The task required repeating each sentence heard. Target sentence intensity was adjusted to find the level at which the child achieved 50% of words correct for each sentence. While the test has four measures, this study only measured the high cue speech recognition threshold (SRT), assessing listening skills when both vocal and spatial cues are available. 2.2.5. Phonological awareness As mentioned above, the children in this study attend a bilingual school in which literacy is taught in the local communilect in the first few years of schooling, whilst the children acquire English orally. It was therefore decided that the most appropriate test of these children's PA would be in their first language. The phonological awareness assessment used was thus one developed, trialled and reported on by Morales (2018) as an iPad app [81]. The app launched with a welcome screen greeting children in their local communilect. Twenty questions followed, each presented serially on a novel screen. All questions were presented with an answer grid, with four cells containing letters or images (see Fig. 1). Students selected an answer by tapping on one of the four cells, which was highlighted in green if correct, or red if not. Children were free to change their answer. Children could not proceed to the following question (by nominating an arrow) without having selected an answer for the existing question. Once completed, a closing screen appeared with a congratulatory message regardless of performance. The app was structured as follows (Appendix A).

2.2.3. Pure tone audiometry Pure tone audiometry was undertaken to determine participant hearing thresholds. The Hughson-Westlake procedure [68] was employed to determine the minimum sound intensity required for each child to reliably hear the test sounds presented at 0.5, 1 k, 2 k, and 4 k Hz. An Interacoustics® AD226 audiometer was used, with TDH39 headsets in amplivox audiocups. Amplivox audiocups are designed for undertaking audiometry in the nonclinical setting to attenuate background noise. Audiometers and headsets were calibrated per ANSI/ASA S3.6-2010 recommendations. Thresholds were determined via play audiometry. Children were conditioned to hold a token to their ear and listen for a sound. When heard, the child was instructed to place the token into a game puzzle (Connect 4), indicating the stimulus was audible. 2.2.4. Binaural processing Both assessments for dichotic listening and speech perception in noise presented stimuli via Sennheiser HD215 over-the-ear headphones. Computerised programs employed for both assessments featured inbuilt, Australian norms for each age. Participant scores were converted into Z-scores to determine how each child compared against normative data [69, 70].

• 6 letter identification questions (e.g., find the letter that makes the ‘l’ sound) • 4 initial syllable identification questions (e.g., find the picture that starts with /gu/) • 3 final consonant identification questions (e.g., find the picture that ends with /k/) • 3 syllable blending questions (e.g., find what you get when you put

2.2.4.1. Dichotic listening: Dichotic digit difference test. Binaural processing was assessed via the DDdT [69]. Testing was performed via a Toshiba Tecra R950 laptop with an integrated with a Phonak Soundcard adaptor.

‘we’ … pause … ‘ṯi’ together)

4

International Journal of Pediatric Otorhinolaryngology 128 (2020) 109702

M. Sharma, et al.

Fig. 1. Screenshots of the assessment app questions.

• 4 phoneme blending questions (e.g., Find what you get when you

classroom noise in different countries [66]. Unoccupied classroom noise levels are recommended to be less than 35 dB (A-weighted) SPL for school-based hearing screening [73]. As expected, noise levels featured a low-frequency dominance, decreasing with higher octave-band frequencies. This is consistent with the research literature reporting noise consists predominately of low frequency energy, thereby masking 500 and 1000 Hz pure tones during hearing assessments, despite the application of noise-reducing headsets [66]. Biologic calibration showed the average difference in thresholds, when comparing the sound-treated booth results with those gained via the test room at the school, to be 15 and 5 dB HL for 500 and 1000 Hz respectively.

put /rr/ … pause … /a/ … pause … /m/ … pause … /a/ together)

The app recorded the child's responses, and results were downloaded into an Excel file. The phonological awareness assessment was administered to each child individually in the local communilect by an Indigenous teacher familiar with the app. The children and teacher were well-known to each other. Children could make as many attempts as they wished for each question. Due to the very small population of children, the test has not been normed in the community. However, the assessment was closely based on “Get Ready to Read” [71] which is argued to have strong relationships with other literacy and language assessments [72].

3.2. Hearing assessments

2.3. Procedure

For otoscopic examination, normal tympanic membranes and clear external auditory meati were observed for the right ear (n = 14) and left ear (n = 22). The remaining (more-than half) had retracted tympanic membranes, scarring (tympanosclerosis), redness or perforations (Table 2) based on tympanometry results. For tympanometry (Table 3), 51% of the right ears (n = 24) and 47% of the left ears (n = 21) showed normal peak pressure and compliance while 28% (n = 13) and 23% (n = 11) of the children had Type-C tympanograms [67]. Two children had active discharge on the day of testing, and therefore did not undergo tympanometry. There

All assessments were carried out during teaching time to ensure quiet surroundings. Due the remote locale and conditions, occasional external noise from the hallway, neighbouring rooms, and outdoor areas nearby was possible. Sources of internal room noise (e.g., air conditioning units), were conceivable contributors to ambient noise within the test environment – potentially influencing the results. However, given the school-based location, the environment was similar to that which children encounter in an educational setting on a daily basis and in which they have to learn. Assessments required the children to rotate between five different work stations. All children attempted most tasks. The entire session took approximately 30–40 min per participant for the audiological assessments, and a further 15–20 min for PA testing. For each task, practice items were given to ensure understanding.

Table 2 Otoscopic results (n = 47) against the descriptors.

2.4. Data analyses All scores for DDdT, LiSN-S and PA were presented as raw scores as currently, there are no norms available for bilingual children since most test norms are based on monolingual English-speaking, urban children. Pearson correlations were undertaken using the software package STATISTICA to determine the association between hearing, binaural processing tasks of DDdT and LiSN-S and phonological processing.

Descriptors

Right

Left

Normal & clear Dull Retracted Wax Active discharge Perforation Scarring Redness Could not test

14 9 3 11 2 1 4 2 1

22 10 2 6 1 3 2 0 1

Table 3 Tympanometric results (n = 46) shows the number of ears under the clinicallyused (Jerger) classification.

3. Results 3.1. Ambient noise

Right Left

Mean ambient noise in the testing classroom was consistent across the 4.5 days and multiple readings. The general noise levels were higher than ANSI recommendations, but consistent with other reports of

a

5

Type-A

Type-B

Type-C

Could not test

Othera

24 21

4 5

13 11

1 2

5 8

Type-Adeep, Type-Ashallow.

International Journal of Pediatric Otorhinolaryngology 128 (2020) 109702

M. Sharma, et al.

the LiSN-S high cue scores (r = −0.56, p = 0.005; r = −0.49, p = 0.016, respectively), indicating that those with good performance on one task were good at the others as well. The negative correlation was due to different expected thresholds. LiSN-S determines the lowest threshold where one can perceive the sentence, while dichotic and diotic scores reflect the percentage score of correctly identified numbers. The Z-scores when plotted for both dichotic listening and LiSN-S show lower performance compared to published norms for most children. Fourteen children (of n = 33) showed scores better than -2SD for dichotic listening while all 24 children (n = 24) had poorer than -2SD on LiSN-S (High cue) task [69,70] (appendix 1).

Table 4 Audiogram mean and standard deviation results for Right and Left ears, including maxima and minimum.

R N = 46 L N = 46

Mean Standard deviation Minima Maxima Mean Standard deviation Minima Maxima

500

1 kHz

2 kHz

4 kHz

26.3 7.1 15 50 25.3 8.9 15 55

18.0 9.0 5 50 18.3 10.9 5 50

15.9 9.3 5 45 17.0 10.3 5 60

16.4 9.3 0 45 18.7 12.0 0 60

3.4. Phonological awareness

were 4 right and 5 left ears with Type B tympanograms. There was no obvious trend in middle ear status with age. Table 4 shows the averaged thresholds for 46 children (one child could not be conditioned to provide consistent results). Six children (aged 5.6–8.1 years) showed bilateral mild hearing loss (with greater than 20 dB thresholds on two consecutive frequencies), with five children (aged 5.6–8 years) demonstrating a unilateral mild hearing loss. The type of hearing loss could not be determined as bone conduction thresholds could not be reliably measured due to background noise levels. Averaged Pure Tone Audiometry (PTA) thresholds across 500 Hz, 1 kHz, 2 kHz and 4 kHz for the 2 ears were significantly correlated (r = 0.59, p < 0.001, n = 46). Pearson correlations showed no significant association between hearing thresholds and age.

Table 6 shows the phonological awareness (PA) task scores. While all 47 children attempted the task, 44 completed the letter identification task, whereas 34 children were able to do the phoneme blending task. The three children who had difficulty with letter identification were 5.4–6.8 years old and lacked any co-occurring hearing loss. Of the 3 children demonstrating this difficulty, the older two had Type-C tympanometry bilaterally. Pearson product-moment correlations showed significant correlations between all the subtasks relative to the total PA score. As expected, the total PA score was also associated with age (r = 0.71, p < 0.001, n = 47) such that older children produced better scores for PA skills. There was one outlier (< 2SD above the mean) with a total PA score of 26, but even after removal, the correlation remained significant (Fig. 3).

3.3. Binaural processing

4. Links between hearing, binaural listening, speech perception in noise, and phonological awareness

3.3.1. DDdT There were 33 children (aged 6.9 years and older) who were able to complete the dichotic listening task, with 31 completing the diotic listening task (Table 5). Dichotic and diotic scores were significantlycorrelated (p = 0.86, p < 0.001, n = 31). The dichotic and diotic scores were also significantly-correlated with age, improving as children got older (r = 0.67, p < 0.001, n = 33) and (r = 0.72, p = 0.001, n = 31), respectively. Two of the scores for the diotic listening task were removed for the final analysis as both were considered outliers (< 2SD from the mean), therefore the correlation was observed for 31 ear results (Fig. 2).

Hearing thresholds, as measured for the two ears and averaged across 0.5, 1 k, 2 k and 4 kHz, provided left and right PTA averaged thresholds. Hearing thresholds did not correlate significantly to binaural listening, speech in noise or the phonological awareness total score. There was one exception involving left ear hearing thresholds, which correlated with dichotic listening scores (p = 0.01). Considering the multiple correlations, these findings (possibly due to type-I error), may be random. There were 4 pairs of comparisons, and the significance level of the correlation was relatively-low for the left ear thresholds and dichotic listening scores. As anticipated, the PA total score was significantly correlated with dichotic digits (p = 0.56, p = 0.001, n = 33); diotic digits (r = 0.70, p < 0.001, n = 31) and LiSN-S (r = −0.57, p = 0.004, n = 24). In order to account for any hearing loss, PA was partially correlated with the dichotic and diotic digits, and there remained a significant correlation with total PA score (r = 0.59, p < 0.001, n = 33; r = 0.74, p < 0.001, n = 31). LiSN-S also showed a significant correlation with PA total score after hearing thresholds had been included (r = −0.56, p = 0.005, n = 24) (Fig. 4).

3.3.2. LiSN-S Twenty-four children (7 years and older (mean age 9.09, SD 1.62)), were able to complete the LiSN-S task, requiring them to repeat the sentences spoken in English. Most had problems maintaining attention, potentially due to their still-developing competence with English (Mean score −2.60, SD 5.95). LiSN-S also was significantly correlated with age, (r = −0.47, p = 0.020, n = 24) such that older children performed better on this task relative to the younger children. 3.3.3. DDdT and LiSN-S The dichotic and diotic scores were also significantly correlated to

5. Discussion

Table 5 Mean performance for Dichotic (n = 33) and Diotic (n = 31) listening.

Mean (in percent) Standard deviation Range

Dichotic

Diotic

59.4 17.5 25–88.8

56.7 12.7 42.5–83.8

The main findings of this research are as follows: 1) nearly half of the participating Indigenous children had some level of middle ear disorder on the day their hearing was tested (i.e., other than Type A tympanograms); and 2) there was a significant correlation between phonological awareness and dichotic listening, as well as between phonological awareness and speech perception in noise.

6

International Journal of Pediatric Otorhinolaryngology 128 (2020) 109702

M. Sharma, et al.

Fig. 2. A. Scatter plot with 95% confidence intervals (dashed lines) showing significant correlations between age (x-axis in years) and Dichotic listening (y axis in percentage). B. Scatter plot with 95% confidence intervals (dashed lines) showing significant correlations between age (x-axis in years) and Diotic listening (y axis in percentage).

5.1. Association between OM, binaural listening, speech in noise and PA

determined older children (> 5years) in Indigenous communities continue to report middle ear disorders [9]. Results from the present study are consistent with research reporting an association between OM and dichotic listening and speech perception in noise [1,28,29], but not in accordance with Moore et al. [19] or Paradise et al. [59]. Half of the children in the present study had

Indigenous Australians are five-times more likely to have severe OM compared to non-Indigenous Australians [74]. The high incidence of middle ear disorders in the participating children was commensurate with previous studies. Results were also consistent with findings that

7

International Journal of Pediatric Otorhinolaryngology 128 (2020) 109702

M. Sharma, et al.

Table 6 Mean performance (and standard deviation = Stdev) for the 47 children on each of the Phonological Awareness (PA) tasks.

Letter identification Initial syllable identification Final syllable identification Syllable blending Phoneme blending Phonological Awareness total

Mean

Stdev

Minima

Maxima

Correlations to PA total

3.74 1.49 2.00 2.47 2.40 12.1

2.04 0.86 1.18 0.88 3.03 4.71

0 0 0 0 0 4

6 4 4 4 13 26

R = 0.32, R = 0.47, R = 0.67, R = 0.63, R = 0.76,

p = 0.028 p = 0.001 p < 0.001 p < 0.001 p < 0.001

Fig. 3. Scatterplot showing the correlation between the Phonological Awareness (PA) total score (y-axis) and the age (in years) on x-axis with 95% confidence intervals (dashed lines).

some manner of middle ear disorder on the day of assessment, and performed poorly on the binaural listening and speech perception in noise tasks. However, individual performance on both tasks (Fig. 3) showed no difference between children with the middle ear concern on the day of assessment and those with normal tympanometry. Therefore, the poor performance cannot be completely explained on the basis of middle ear status on the day of testing. One factor highlighted by previous literature as important when investigating OM consequences is socioeconomic status (SES). Castagno and Lavinsky [75] suggested that people from lower SES are at-risk for higher incidence of poorer health. Chadha et al. [76] specifically compared the prevalence of OM in high versus lower SES in Delhi, India. They reported nearly 20% of children from lower SES had OM while 2% from higher SES had OM. The comparison included parental education, hygiene, diet, and number of family members – all factors that contribute to higher incidence of OM [76]. The current study's community base is a low SES region and therefore subject to a higher incidence of OM than would be expected in Australia more generally.1 Consequently, in future studies comparing performance across communities, it would be essential to control for SES to further clarify the

extent to which low SES is potentially driving the association between OM, PA, and binaural processing. The current study included children who learn English as the second language, which they begin to learn when commencing school [60]. This is unlike the Graydon et al. [1], Cameron et al. [28] and Tomlin and Rance [29] studies, which included Aboriginal children who were Kriol speakers, or those who had been exposed to English language for longer durations. In general, over half of the Australian Aboriginal population living in remote areas (similar to the current cohort) speak an Aboriginal language, compared to 1% in the urban towns and cities [60]. This could suggest that the children in the current study performed poorly due to the audiology tests requiring a certain level of competence in English language. This is why the LiSN-S was only attempted by the older children (> 7 years) since it requires a level of English language competence differences. In future, to assess speech in noise abilities, a task with minimal language competence needs to be used. One such task, the Triple Digit Test, has been used with adults with hearing loss [77]. The test is similar to the LiSN-S test but differs in requiring listeners to repeat numbers (rather than sentences) presented in noise. All children, including the younger 5 to 6 year-olds, were more comfortable repeating numbers, yet for the current study's dichotic/ diotic digit task, children's performance was lower than published norms [69]. While it seems unlikely that poorer performance on the digit tests was mainly due to language competence, it is difficult to rule this out without testing children with English as second language with minimal (or no) history of OM. Note that the normative data for the LiSN-S test was collected from

1

In Australia, the ICSEA (Index of Community Socio-Economic Advantage) scale ranges from about 500 (low advantage) to 1300 (high advantage) with a mean of 1300 and a standard deviation of 100; the school in this study had a score of 590, indicating considerable socio-economic disadvantage [82] (Wilson 2014). 8

International Journal of Pediatric Otorhinolaryngology 128 (2020) 109702

M. Sharma, et al.

suggested that the presence of history of conductive hearing loss relates to poor spatial listening [1,29]. However, these studies tested children when their hearing was within normal limits. The current study is one of a few prospective studies that has investigated PA and binaural listening skills when a significant portion (at least half) of participants had a concurrent middle ear disorder. Importantly, the performance of children on PA or binaural processing tasks did not correlate with their hearing thresholds. Information regarding the history of duration of OM episodes, the number of episodes and the age when children experienced their first OM would be useful for better understanding the true impact of OM on these children's auditory and PA abilities. Correlations are not evidence of causation, but the association between binaural processing and PA, independent of hearing thresholds, is an important finding. This suggests that there are common factors that influence auditory processing and PA skills (e.g., including age, attention, and/or language competence). 5.2. Future directions Further studies are needed with a larger sample, where multi-regression analyses can be undertaken to define the complex relationships between age, hearing, language competence, binaural listening, and PA in this population. Until these relationships have been well-defined, the focus needs to be on improving binaural processing and PA. Emphasis also needs to be on changing the current practice of only measuring hearing thresholds in this population to evaluate and characterise the impact of OM. Tasks that include binaural processing need to be included in the current test battery to ensure the long-term impact of OM is well-defined for each child.

Fig. 4. Scatterplot showing the significant correlations between the Dichotic (i.e., binaural) listening score and [B] LiSN-S (Listening in spatialized NoiseSentences) and Phonological Awareness (on x-axis). The dotted line indicates the regression line.

6. Conclusions

48 children with Australian English as their first language and no history of hearing problems [70]. Similarly, the dichotic digit test norms are based on data collected in urban regions. Furthermore, the normative data collection was carried out in an acoustically-treated sound studio, unlike the classroom used in the current study. These factors may explain, in part, the overall lower performance of the children in the present study. However, despite the fact that the PA task was carried out in the children's first language, overall performance was poorer than expected though one 12-year-old was able to achieve scores close to perfect (Fig. 4). When looking at the specific subtasks, all children were able to identify letters and syllables, but had difficulty with phoneme identification and blending (e.g./b/+/æ/+/t/ = bat) Consistent with the current study, a recent study reported that 8-11-year-olds in the UK with OM had difficulties on phoneme blending [78]. Although children in current study from grades 4–6 were able to successfully complete the letter identification, this is a skill most monolingual children in urban settings are expected to master by grade 1 [79]. Several factors may contribute to the lower-than-expected performance on the phonological awareness task, including the low SES levels in remote communities, lack of access to literacy and pre-literacy activities in the communilect (e.g., in terms of notices or signposts), and lack of access to early literacy activities at home; this may be partly a function of the low literacy levels of many parents [80]. Apart from the language differences, other considerations that need to be acknowledged include the 1) non-determination of type of hearing loss; 2) history of OM for each child; and 3) cognitive measures including working memory and attention. Previous literature has

The current study highlights the presence of persistent middle ear disorders in Indigenous children, suggesting an association between binaural processing and phonological awareness, over and above hearing levels. This has implications for current practice, where hearing thresholds are primarily tested in children with long-term otitis media. The fact that hearing thresholds were not associated with either binaural processing or phonological awareness in this study, suggests that tasks of binaural processing should be considered as standard practice for assessing the children with a history of OM. Funding The current study was funded by the following bodies: ARC Centre of Excellence for Cognition and its Disorders (CE110001021), ARC Centre of Excellence for the Dynamics of Language (CE140100041), ARC Laureate Fellowship (FL130100014), and the Macquarie University Centre for Language Sciences (CLaS). The funders have had no role in the study design or the data analysis, interpretation and writing. Availability of data and materials The datasets generated and analysed during the current study are not publicly available (since it is not possible to share the research data without compromising the privacy of individual participants in the small community in which the research was undertaken).

9

International Journal of Pediatric Otorhinolaryngology 128 (2020) 109702

M. Sharma, et al.

Authors’ contributions

Publisher's note

MS, GW and KD were primarily involved in the study design, interpretation of the results, and write up of the manuscript. MS, GW and GS were primarily involved in community consultation and data collection. MS and GS were primarily involved in data analysis. All authors read and approved the final manuscript.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Declaration of competing interest

Ethics approval and consent to participate

The authors declare that they have no competing interests.

The research was approved by the Research Ethics Committee at Menzies School of Health Research (HREC 2016–2702) and the Research Ethics Committee at Macquarie Univeristy (approval number 5201600743). Informed (written) consent to participate was given by parents’/caretakers of all participants.

Acknowledgements The authors gratefully acknowledge the children, parents, teachers and school principal for their participation in the study, as well as the students who collected the data.

Consent for publication Not applicable. Appendix A. (from Morales 2018)

The phonological awareness assessment app is delivered in the communilect although English translations are given here for the instructions for ease of readability; Task items (target words and distractor words) remain in the communilect: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

Find the letter that makes the/a/sound. Find the letter that makes the/l/sound. Find the letter that makes the/ŋ/sound. Find the letter that makes the/i/sound. Find the letter that makes the/r/sound. Find the letter that makes the/k/sound. These pictures are guya, bäpi, waṯu, and mäṉa. Find the one that starts with/gu/. These pictures are bathi, ḏowu, mapu, and lorri. Find the one that starts with/ma/. These pictures are gara, bolu, raŋi, and ṉaku. Find the one that ends with/ku/. These pictures are yalu, mapu, buku, and ŋäṯi. Find the one that starts with/m/. These pictures are wäŋa, lorri, djäri, and ṉaku. Find the one that starts with/dj/. These pictures are bärr, mel, djeṯ, and goŋ. Find the one that ends with/l/. These pictures are borum, dawurr, ḻikan, and gärak. Find the one that ends with /k/. These pictures are bäru, weṯi, guya, and rupu. Find what you get when you put bä … pause … ru together. Find bä … pause … ru. The pictures are djuku, mäṉa, ḻaḻu, and weṯi. Find what you get when you put we … pause … ṯi together. Find we … pause … ṯi. These pictures are daruma, muthali, djitama, and yiḏaki. Find what you get when you put mu … pause … tha … pause … li together. Find mu … pause … tha … pause … li. These pictures are djeṯ, mel, bärr, and goŋ. Find what you get when you put /m/ … pause … /e/ … pause … /l/together. Find/m/ … pause … /e/ … pause … /l/. These pictures are goŋ, ḏetj, djät, and ŋäl. Find what you get when you put /g/ … pause … /o/ … pause … /ŋ/together. Find/g/ … pause … /o/ … pause … /ŋ/. These pictures are raŋi, retja, rräma, and wäŋa. Find what you get when you put /r/ … pause … /a/ … pause … /m/ … pause … /a/together. Find /r/ … pause … /a/ … pause … /m/ … pause … /a/. These pictures are yothu, ḏowu, yiki, and ŋatha. Find what you get when you put /ŋ/ … pause … /a/ … pause … /th/ … pause … /a/together. Find /ŋ/ … pause … /a/ … pause … /th/ … pause … /a/.

10

International Journal of Pediatric Otorhinolaryngology 128 (2020) 109702

M. Sharma, et al.

Appendix 1. Individual performance on [A] dichotic listening (as measured on Dichotic Diotic digit difference Test, n = 33). ‘–’ indicates Dichotic scores and ‘x’ indicates Diotic scores. The dotted arrow in A indicates the child with unilateral hearing loss. [B] shows speech perception in noise scores (as measured by LiSN-S) of participating children against the published norms. The dotted lines in both graphs indicate 2 SD from the normed mean (0). All symbols that fall below the −2 SD dotted line would be regarded as lower thresholds than their age-matched urban peers on both tasks.

[13] P. Morris, A systematic review of clinical research addressing the prevalence, aetiology, diagnosis, prognosis and therapy of otitis media in Australian Aboriginal children, J. Paediatr. Child Health 34 (6) (1998) 487–497. [14] J. Jervis-Bardy, L. Sanchez, A. Carney, Otitis media in Indigenous Australian children: review of epidemiology and risk factors, J. Laryngol. Otol. 128 (S1) (2014) S16–S27. [15] M.E. Ravicz, J.J. Rosowski, S.N. Merchant, Mechanisms of hearing loss resulting from middle-ear fluid, Hear. Res. 195 (1–2) (2004) 103–130. [16] T. Cai, B. McPherson, Hearing loss in children with otitis media with effusion: a systematic review, Int. J. Audiol. 56 (2) (2017) 65–76. [17] S. Haapala, et al., Effects of recurrent acute otitis media on cortical speech-sound processing in 2-year old children, Ear Hear. 35 (3) (2014) e75–e83. [18] S. Maruthy, J. Mannarukrishnaiah, Effect of early onset otitis media on brainstem and cortical auditory processing, Behav. Brain Funct. 4 (1) (2008) 17. [19] D.R. Moore, D.E. Hartley, S.C. Hogan, Effects of otitis media with effusion on central auditory function, International Congress Series, Elsevier, 2003. [20] A. Zumach, et al., Speech perception after early-life otitis media with fluctuating hearing loss, Audiol. Neurotol. 16 (5) (2011) 304–314. [21] J.P. Whitton, D.B. Polley, Evaluating the perceptual and pathophysiological consequences of auditory deprivation in early postnatal life: a comparison of basic and clinical studies, JARO J. Assoc. Res. Otolaryngol. 12 (5) (2011) 535–547. [22] E.I. Knudsen, S.D. Esterly, P.F. Knudsen, Monaural occlusion alters sound localization during a sensitive period in the barn owl, J. Neurosci. 4 (4) (1984) 1001–1011. [23] J.H. Grose, The masking-level difference in children, J. Am. Acad. Audiol. 1 (2) (1990) 81–88. [24] J.W. Hall, J.H. Grose, H.C. Pillsbury, Long-term effects of chronic otitis media on binaural hearing in children, Arch. Otolaryngol. Head Neck Surg. 121 (8) (1995) 847–852.

References [1] K. Graydon, et al., Consequences of early conductive hearing loss on long-term binaural processing, Ear Hear. 38 (5) (2017) 621–627. [2] C.J. Williams, A.M. Jacobs, The impact of otitis media on cognitive and educational outcomes, Med. J. Aust. 191 (S9) (2009) S69–S72. [3] M.C. Liberman, L.D. Liberman, S.F. Maison, Chronic conductive hearing loss leads to cochlear degeneration, PLoS One 10 (11) (2015) e0142341. [4] M. Mody, et al., Speech perception and verbal memory in children with and without histories of otitis media, J. Speech Lang. Hear. Res. 42 (5) (1999) 1069–1079. [5] J. Roberts, et al., Otitis media, hearing loss, and language learning: controversies and current research, J. Dev. Behav. Pediatr. 25 (2) (2004) 110–122. [6] A.J. Leach, et al., Bacterial colonization of the nasopharynx predicts very early onset and persistence of otitis media in Australian aboriginal infants, Pediatr. Infect. Dis. J. 13 (11) (1994) 983–989. [7] A.J. Leach, P.S. Morris, Antibiotics for the prevention of acute and chronic suppurative otitis media in children, Cochrane Database Syst. Rev. (4) (2006). [8] P. Morris, et al., Otitis media in young Aboriginal children from remote communities in Northern and Central Australia: a cross-sectional survey, BMC Pediatr. 5 (1) (2005) 27. [9] M. Mahadevan, et al., A review of the burden of disease due to otitis media in the Asia-Pacific, Int. J. Pediatr. Otorhinolaryngol. 76 (5) (2012) 623–635. [10] A.-C. Nyquist, et al., Antibiotic prescribing for children with colds, upper respiratory tract infections, and bronchitis, JAMA 279 (11) (1998) 875–877. [11] H. Gunasekera, et al., Primary care management of otitis media among Australian children, Med. J. Aust. 191 (9) (2009) S55. [12] P. Morris, et al., An overview of acute otitis media in Australian Aboriginal children living in remote communities, Vaccine 25 (13) (2007) 2389–2393.

11

International Journal of Pediatric Otorhinolaryngology 128 (2020) 109702

M. Sharma, et al.

[54] M.R. Burchinal, et al., Otitis media in early childhood and its relationship to later phonological development, JSHD (J. Speech Hear. Disord.) 53 (4) (1988) 424–432. [55] I. Augustsson, I. Engstand, Otitis media and academic achievements, Int. J. Pediatr. Otorhinolaryngol. 57 (1) (2001) 31–40. [56] V. Fougner, et al., Early childhood otitis media and later school performance–A prospective cohort study of associations, Int. J. Pediatr. Otorhinolaryngol. 94 (2017) 87–94. [57] K. Bennett, et al., Behaviour and developmental effects of otitis media with effusion into the teens, Arch. Dis. Child. 85 (2) (2001) 91–95. [58] S. Nittrouer, L.T. Burton, The role of early language experience in the development of speech perception and phonological processing abilities: evidence from 5-yearolds with histories of otitis media with effusion and low socioeconomic status, J. Commun. Disord. 38 (1) (2005) 29–63. [59] J.L. Paradise, et al., Language, speech sound production, and cognition in threeyear-old children in relation to otitis media in their first three years of life, Pediatrics 105 (5) (2000) 1119–1130. [60] S. Aithal, A. Yonovitz, V. Aithal, Perceptual consequences of conductive hearing loss: speech perception in Indigenous students learning English as a'school'language, Aust. N. Z. J. Audiol. 30 (1) (2008) 1. [61] J. Simpson, G. Wigglesworth, Language diversity in Indigenous Australia in the 21st century, Curr. Issues Lang. Plann. (2018) 1–14. [62] G. Wigglesworth, J. Simpson, Going to school in a different world, Language Practices of Indigenous Children and Youth, Springer, 2018, pp. 1–20. [63] R. Amery, A New Diglossia: Contemporary Speech Varieties at Yirrkala in North East Arnhem Land, (1985). [64] R. Amery, An Australian koine: dhuwaya, a variety of Yolŋu Matha spoken at Yirrkala in worth East Arnhemland, Int. J. Sociol. Lang. 99 (1) (1993) 45–64. [65] G. Morales, J. Vaughan, M. Ganambarr-Stubbs, From home to school in multilingual Arnhem Land: the development of Yirrkala school's bilingual curriculum, Language Practices of Indigenous Children and Youth, Springer, 2018, pp. 69–98. [66] A.H.C. Lo, B. McPherson, Hearing screening for school children: utility of noisecancelling headphones, BMC Ear Nose Throat Disord. 13 (1) (2013) 6. [67] J. Jerger, Clinical experience with impedance audiometry, Arch. Otolaryngol. 92 (4) (1970) 311–324. [68] R. Carhart, J.F. Jerger, Preferred method for clinical determination of pure-tone thresholds, J. Speech Hear. Disord. 24 (4) (1959) 330–345. [69] S. Cameron, et al., The Dichotic Digits Difference Test (DDdT): Development, Normative Data and Retest Reliability Studies, (2015). [70] S. Cameron, H. Dillon, Development of the listening in spatialized noise-sentences test (LISN-S), Ear Hear. 28 (2) (2007) 196–211. [71] G.J. Whitehurst, The NCLD Get Ready to Read. Screening Tool Technical Report, (2001). [72] B.M. Phillips, C.J. Lonigan, M.A. Wyatt, Predictive validity of the Get Ready to Read! screener: concurrent and long-term relations with reading-related skills, J. Learn. Disabil. 42 (2) (2009) 133–147. [73] Institute, A.N.S., Acoustical Performance Criteria, Design Requirements, and Guidelines for Schools, Part 1: Permanent Schools, in ANSI S12. 2010, Author New York: New York. [74] D. Lehmann, et al., The Kalgoorlie Otitis Media Research Project: rationale, methods, population characteristics and ethical considerations, Paediatr. Perinat. Epidemiol. 22 (1) (2008) 60–71. [75] L.A. Castagno, L. Lavinsky, Otitis media in children: seasonal changes and socioeconomic level, Int. J. Pediatr. Otorhinolaryngol. 62 (2) (2002) 129–134 62(2): p. 5. [76] S.K. Chadha, et al., A comparative evaluation of ear diseases in children of higher versus lower socioeconomic status, J. Laryngol. Otol. 120 (1) (2005) 16–19. [77] H.D. Semeraro, et al., Development and evaluation of the British English coordinate response measure speech-in-noise test as an occupational hearing assessment tool, Int. J. Audiol. 56 (10) (2017) 749–758. [78] J.M. Carroll, H.L. Breadmore, Not all phonological awareness deficits are created equal: evidence from a comparison between children with otitis media and poor readers, Dev. Sci. 21 (3) (2018). [79] N.K. Lesaux, L.S. Siegel, The development of reading in children who speak English as a second language, Dev. Psychol. 39 (6) (2003) 1005. [80] G.L. Webb, C.J. Williams, Factors affecting language and literacy development in Australian Aboriginal children: considering dialect, culture and health, J. Early Child. Res. 16 (1) (2018) 104–116. [81] G. Morales, Technology and Bilingual Education: Helping Yolŋu Students Crack the Alphabetic Code, Unpublished PhD thesis, University of Melbourne, 2018. [82] B. Wilson, A Share in the Future: Review of Indigenous Education in the Northern Territory. The Education Business, (2014) Retrieved from http://www.education. nt.gov.au/data/assets/pdf_file/0007/37294/A-Share-in-the-Future-The-Review-ofIndigenous-Education-in-the-Northern-Territory.pdf.

[25] H.L. Read, J.A. Winer, C.E. Schreiner, Functional architecture of auditory cortex, Curr. Opin. Neurobiol. 12 (4) (2002) 433–440. [26] J.L. Thornton, et al., The conductive hearing loss due to an experimentally induced middle ear effusion alters the interaural level and time difference cues to sound location, JARO J. Assoc. Res. Otolaryngol. 13 (5) (2012) 641–654. [27] A. Zumach, et al., Otitis media and speech-in-noise recognition in school-aged children, Audiol. Neurotol. 14 (2) (2009) 121–129. [28] S. Cameron, et al., Prevalence and remediation of spatial processing disorder (SPD) in Indigenous children in regional Australia, Int. J. Audiol. 53 (5) (2014) 326–335. [29] D. Tomlin, G. Rance, Long-term hearing deficits after childhood middle ear disease, Ear Hear. 35 (6) (2014) e233–e242. [30] D.E. Hartley, D.R. Moore, Effects of conductive hearing loss on temporal aspects of sound transmission through the ear, Hear. Res. 177 (1–2) (2003) 53–60. [31] T.M. Mowery, V.C. Kotak, D.H. Sanes, Transient hearing loss within a critical period causes persistent changes to cellular properties in adult auditory cortex, Cerebr. Cortex 25 (8) (2014) 2083–2094. [32] D.E. Hartley, D.R. Moore, Effects of otitis media with effusion on auditory temporal resolution, Int. J. Pediatr. Otorhinolaryngol. 69 (6) (2005) 757–769. [33] D.R. Moore, D.E. Hartley, S.C. Hogan, Effects of otitis media with effusion (OME) on central auditory function, Int. J. Pediatr. Otorhinolaryngol. 67 (2003) S63–S67. [34] A. Schilder, A. Snik, H. Straatman, The effect of otitis media with effusion at preschool age on some aspects of auditory perception at school age, Ear Hear. 15 (3) (1994) 224–231. [35] L. Bradley, P.E. Bryant, Categorizing sounds and learning to read—a causal connection, Nature 301 (5899) (1983) 419. [36] A.G. Bus, M.H. Van IJzendoorn, Phonological awareness and early reading: a metaanalysis of experimental training studies, J. Educ. Psychol. 91 (3) (1999) 403. [37] A. Castles, M. Coltheart, Is there a causal link from phonological awareness to success in learning to read? Cognition 91 (1) (2004) 77–111. [38] G.W. MacDonald, A. Cornwall, The relationship between phonological awareness and reading and spelling achievement eleven years later, J. Learn. Disabil. 28 (8) (1995) 523–527. [39] C. Lonigan, C. Schatschneider, L. Westberg, The National Early Literacy Panel, Identification of Children's Skills and Abilities Linked to Later Outcomes in Reading, Writing, and Spelling, (2008). [40] N.R. Panel, N.I.o.C. Health, H. Development, Report of the National Reading Panel: Teaching Children to Read: an Evidence-Based Assessment of the Scientific Research Literature on Reading and its Implications for Reading Instruction: Reports of the Subgroups, National Institute of Child Health and Human Development, 2000 (National …). [41] C.M. Schuele, D. Boudreau, Phonological awareness intervention: beyond the basics, Lang. Speech Hear. Serv. Sch. 39 (1) (2008) 3–20. [42] J. Snowling Margaret, et al., Educational attainments of school leavers with a preschool history of speech‐language impairments, Int. J. Lang. Commun. Disord. 36 (2) (2001) 173–183. [43] G.T. Gillon, Facilitating phoneme awareness development in 3-and 4-year-old children with speech impairment, Lang. Speech Hear. Serv. Sch. 36 (4) (2005) 308–324. [44] W. Schneider, et al., Short-and long-term effects of training phonological awareness in kindergarten: evidence from two German studies, J. Exp. Child Psychol. 66 (3) (1997) 311–340. [45] M.J. Snowling, The development of graphene-phoneme correspondence in normal and dyslexic readers, J. Exp. Child Psychol. 29 (2) (1980) 294–305. [46] T. Nicholson, At the Cutting Edge: the Importance of Phonemic Awareness in Learning to Read and Spell, New Zealand Council for Educational Research Wellington, New Zealand, 2005. [47] H. Winskel, The effects of an early history of otitis media on children's language and literacy skill development, Br. J. Educ. Psychol. 76 (4) (2006) 727–744. [48] J.F. Werker, et al., Developmental Aspects of Cross-Language Speech Perception, Child development, 1981, pp. 349–355. [49] N. Walker, G. Wigglesworth, The effect of conductive hearing loss on phonological awareness, reading and spelling of urban Aboriginal students, Aust. N. Z. J. Audiol. 23 (1) (2001) 37. [50] L. Timms, et al., Literacy skills of Australian Indigenous school children with and without otitis media and hearing loss, Int. J. Speech Lang. Pathol. 16 (3) (2014) 327–334. [51] C.J. Williams, J.J. Masterson, Phonemic awareness and early spelling skills in urban Australian Aboriginal and non-Aboriginal children, Int. J. Speech Lang. Pathol. 12 (6) (2010) 497–507. [52] J. Roberts, et al., Otitis media in early childhood and its relationship to later verbal and academic performance, Pediatrics 78 (3) (1986) 423–430. [53] J. Roberts, M.R. Burchinal, S.A. Zeisel, Otitis media in early childhood in relation to children's school-age language and academic skills, Pediatrics 110 (4) (2002) 696–706.

12