Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 May 30.
Published in final edited form as: Cochlear Implants Int. 2025 May 30;26(2):120–129. doi: 10.1080/14670100.2025.2507477

Reading abilities of adolescents with cochlear implants: The role of early speech perception in a longitudinal study

Lisa S Davidson 1, Ann E Geers 2, Rosalie M Uchanski 1
PMCID: PMC12219913  NIHMSID: NIHMS2085742  PMID: 40443258

Abstract

Objective:

To examine reading skills of adolescents who use cochlear implants (CIs) and who were implanted early. Three questions are asked: (1) Do these pediatric CI recipients read at age-appropriate levels at adolescence, as they had done at elementary-ages (Grantham et al., 2022)? (2) Are the previously-documented effects of early segmental and suprasegmental speech perception on elementary-age reading skills obtained for adolescent reading skills? (3) Do the effects on adolescent reading differ for early speech perception versus elementary-age speech perception?

Design:

Eighty-six orally-educated children with a mean age of CI of 2 years completed a standardized reading comprehension test at elementary ages (7-11 years) and at adolescent ages (11-16 years). They completed speech perception tests when they were 5-9 years old (early) and again at 7-11 years of age (elementary-age). Reading comprehension scores from elementary and adolescent ages were compared, and the effects of early and elementary-age speech perception on adolescent reading were examined.

Results:

The mean reading passage comprehension standard score was within the normative range for typically hearing age-mates at both elementary-age and adolescent testing. Early suprasegmental speech perception was the only significant speech perception predictor of adolescent reading skills, after controlling for demographic and audiological variables.

Conclusions:

These data bolster the recent literature reporting good reading skills, on average, for children who received early CIs and highlight the long-term benefits of good early suprasegmental speech perception skills. Both segmental and suprasegmental perception should be included in audiological candidacy criteria and educational intervention for children with CIs.

Keywords: cochlear implant, reading, speech perception

Introduction

The ability to read and comprehend written text provides the foundation for long-term academic and occupational success (Kutner, 2007). Mayer and colleagues recently challenged the frequently-cited historical data showing that few children who are deaf achieve age-appropriate reading skills (Mayer, Trezak, & Hancock, 2021). A recent review and a meta-analysis sought to determine the current state of reading for children with early diagnosis of hearing loss and early receipt of cochlear implants (CIs) (Mayer & Trezek, 2018; Wang, Sibaii, Lee, Gill, & Hatch, 2021). Mayer and Trezek (2018) reviewed 21 studies from 1997-2016 that examined reading comprehension of children with CIs (age range ~5 years to18 years). When reported, the mean age at CI ranged from 2.5 to 8.7 years, and children used spoken language or a combination of spoken language and sign (see Mayer and Trezek, 2018, for details). Fourteen of these studies used standardized scores referenced to normative samples of peers with typical hearing (TH) (a standard score of 100 is the average age-appropriate performance while 15 is the standard deviation). For the majority of the studies reviewed, children with CIs achieved reading comprehension scores within the low average to average range of peers with TH (Mayer & Trezek, 2018). Wang et al. (2021) conducted a meta-analysis of 41 articles, published from 2002 to 2019, in which reading comprehension skills among other literacy and language outcomes of children with CIs are compared to those of children with TH. For reading comprehension, Wang et al. showed a positive trend similar to that reported by Mayer and Trezek (2018). Specifically, they found that children with CIs scored 1.39 SD below their hearing peers, within the normal range (defined in their review as within +/− 1.5 SD of the mean) (see Wang, et al. 2021 for details). Both reviews discuss factors that have been associated with better reading outcomes across studies including higher nonverbal intelligence (NVIQ), higher socioeconomic status (SES), earlier receipt of CIs and rehabilitation services, use of two CIs, and a period of early bimodal experience (CI with a hearing aid [HA] at the non-implanted ear). Wang and colleagues found that of all the audiological characteristics evaluated as moderators of performance differences between the CI and TH groups (age at onset of hearing loss, age at CI, duration of CI use, percentage of unilateral CI users), only unilateral CI use of the CI group impacted (negatively) the group differences between children with CIs and with TH.

Reading outcomes for adolescents with CIs

Compared to children in elementary grades, fewer studies have examined reading outcomes when such children reach adolescent ages. In 2004, Spencer, Gantz, & Knutson reported an average reading comprehension standard score of 92 for 27 high-school-age CI recipients (age at 1st CI between 2 and 12 years old). In 2011, Geers and Hayes reported an average reading comprehension standard score of 90, for 112 adolescents (mean age 16.7 years) who received a CI at an average age of 3.3 years. Later, Edwards et al reported an average reading comprehension standard score of 83, for 18 adolescents (mean age 14.3 years) who received a CI at an average age of 3.5 years (Edwards, Aitkenhead, & Langdon, 2016). Mayer and colleagues (Mayer, Watson, Archbold, Ng, & Mulla, 2016) studied reading comprehension skills of 32 pediatric CI recipients (9-16 years of age), who received their 1st CI at an average age of 3.6 years. Instead of standard scores, they reported the percentage of participants who scored below the normal range (13%), within the normal range (75%), and above the normal range (13%) (below and above the normal range are defined as < − 1 SD and > + 1 SD from the mean).

Some studies report that the gap between reading scores for children with CIs and children with TH widens as children age (Geers, Tobey, Moog, & Brenner, 2008; Thoutenhoofd, 2006; van der Kant, Vermeulen, De Raeve, & Schreuder, 2010). Geers et al. (2008) compared total reading scores for 85 children with CIs tested at elementary age and again at secondary-school age; these children’s mean standard scores decreased from 88 to 83. Moreover, fewer participants achieved reading scores within the normal range at secondary-school age (44%) than at elementary age (56%).

Suprasegmental and Segmental Speech perception and reading outcomes

For children with TH, Zhang and McBride-Chang (Zhang & McBride-Chang, 2010) suggest that early speech perception serves as a primary developmental marker for reading. Learning to read begins when a child hears and listens to spoken language; speech perception skills underpin early literacy. For a child with HL, however, speech perception ability and acquisition of the sound-to-letter knowledge needed for reading are rooted in the audibility and perceptual clarity of spoken language provided by the child’s devices (CIs or HAs). The ‘Simple View of Reading’ model considers early oral language skills as a precursor for later literacy skills and an integral component to reading comprehension for all children (see discussion in Wang et al., 2021). The present study focuses on the unique effects of two different types of speech perception skills, rather than the effects of oral language, on literacy. Moreover, at young ages, speech perception assessments rather than language tests are used to make device candidacy decisions, to validate hearing devices, and to inform early therapy goals.

Zhang and McBride-Chang (2010) suggest that sensitivity to suprasegmental aspects of speech (intonation, stress, and rhythm) plays a unique role in the development of reading. Segmental perception has, however, been explored more widely than suprasegmental perception. Wade-Woolley and colleagues (Wade-Woolley, Wood, Chan & Weidman, 2022) assert that while prosodic competence (suprasegmental perception) is involved in all aspects of reading, much is unknown about prosodic competence. Longitudinal studies are needed to understand how prosodic competence develops and how it influences children from pre-reading through skilled reading. The influence of phoneme-level segmental perception on reading has been documented for children with TH (National Reading Panel; National Institute of Child Health Human Development, 2000). Spencer and Oleson (2008) found that early perception and production of vowels, consonants, phonemes and words (segmental speech units) were predictive of reading decoding and comprehension for children with CIs who were tested approximately 7.4 years after receiving a CI (age at CI: mean = 3.5 years; SD = 1.6 years).

The increased interest in the role of suprasegmental perception in reading for children with TH combined with the known limitations of CIs in processing suprasegmental properties of speech, (Carroll & Zeng, 2007; Dorman, Gifford, Spahr, & McKarns, 2008; Gantz, Turner, Gfeller, & Lowder, 2005), motivated Grantham and colleagues (Grantham, Davidson, Geers, & Uchanski, 2022) to examine the contribution of early suprasegmental speech perception to elementary-age reading skills of children who received early CIs, after controlling for segmental perception. Children with bilateral CIs or bimodal devices (N=117) completed tests of segmental perception (word recognition in quiet and in noise, recognition of vowels and consonants in quiet) and of suprasegmental perception (emotion identification, talker and stress discrimination, and nonword repetition of stress) at early school ages (age range: 5-9 years). Regression analyses of these CI recipients’ data revealed that suprasegmental perception accounted for unique variance in both receptive vocabulary and language scores [14% & 24%, respectively (Davidson, Geers, Uchanski, & Firszt, 2019)]. Two years later, 104 of these same 117 children completed standardized tests of reading decoding and reading comprehension. Regression analyses determined that the contributions of early suprasegmental perception operates independently from early segmental perception in facilitating good reading outcomes for these children with CIs (Grantham et al., 2022).

Clinical protocols for children with hearing loss focus primarily on phoneme, word, and sentence recognition (segmental perception) with less attention, if any at all, given to suprasegmental perception (Ching et al., 2013; Uhler, Warner-Cyz, & Gifford, 2017). Recently, however, Karimi-Boroujeni, Dajani, and Giguère (Karimi-Boroujeni, Dajani, & Giguère, 2023) advocate for increased attention to suprasegmental speech perception as it may provide useful information when assessing benefits from device technology and intervention strategies.

In this study we present longitudinal reading comprehension data for 86 of the 104 CI participants in Grantham et al. (2022). These children were tested at elementary ages (7-11 years old) and again when they were between 11 and 16 years old. The primary goal of this study was to examine reading skills of adolescents who use CIs. Three questions are asked: (1) Do these pediatric CI recipients read at age-appropriate levels at adolescence, as they had done at elementary-ages (Grantham et al., 2022)? (2) Are the previously-documented effects of early segmental and suprasegmental speech perception on elementary-age reading skills obtained for adolescent reading skills? (3) Do the effects on adolescent reading differ for early speech perception versus elementary-age speech perception? [see Figure 1.]

Figure 1.

Figure 1.

Open in a new tab

Tests examined at three sessions, “early,” “elementary,” and “adolescent,”. Speech perception results from the “early” session were reported in Davidson et al (2019). The effects of early speech perception on elementary reading were reported in Grantham et al (2022) (see gray arrows). A comparison of early speech perception effects on elementary vs adolescent reading (gray vs open-black arrows) is our question (2). A comparison of the effects of early vs elementary speech perception on adolescent reading (open-black vs filled-black arrows) is our question (3).

Methods

Participants

Participants were recruited from CI centers and schools across the United States; these children were participating in a longitudinal study of the effects of early acoustic hearing on spoken language skills. Please refer to Figure 1 for an overview of the tests conducted at each session, and for the age range of the participants at each session. At the early and elementary sessions, segmental and suprasegmental speech tests were administered as part of a comprehensive test battery (see Davidson et al., 2019). Reading passage comprehension tests were administered at the elementary and adolescent sessions. For this study, focusing on reading, 86 of the 1041 participants who had completed reading passage comprehension at the elementary session and who had also completed speech perception tests at the early session, returned to participate in reading tests at the adolescent session.

Two additional demographic variables were gathered at the early session. A block design task was administered to estimate NVIQ. Depending on the age of the child at that time, the Block Design subtest was either from the Wechsler Intelligence Scale for Children–Fourth Edition (Wechsler, 2003) or from the Wechsler Abbreviated Scale of Intelligence–Second Edition (Wechsler, 2011). Maternal education was used as a socioeconomic variable, categorized by total number of years of reported education: some or all of high school (≤ 12 years), postsecondary (13–16 years), and postgraduate (> 16 years). Table 1 summarizes demographic variables for these 86 children who participated in all three test sessions: early, elementary, and adolescent.

Table 1.

Demographics of adolescents with cochlear implants, n = 86.

mean (sd) min – max
Age at Testing (years) 13.9 (1.4) 11 – 16
Block Design (NVIQ) (standard score) 100.1 (17.2) 60 – 145
Maternal Education (years) 16.3 (2.1) 11 – 20
 
Type: Number of Participants (%)
Gender Female: 42 (49) Male: 44 (51)
Race (9 DNR1) Minority2: 17 (20) White: 60 (70)
Ethnicity (1 Unknown/DNR1) Hispanic: 10 (12) Not Hispanic: 75 (87)
Open in a new tab
1

DNR = Did Not Report

2

Minority: African American – 3; Asian – 8; Multi-Racial – 6

Table 2 summarizes the audiological characteristics for these 86 participants. At the adolescent session, all children, except 3, used either bilateral CIs (received simultaneously or sequentially) or bimodal devices. For these exceptions, 2 children who had been bimodal users and 1 child with bilateral CIs were using one CI at adolescence. During the period between the early and adolescent sessions, 13 children with bimodal devices received a second CI, becoming sequential CI participants at the adolescent session. Also in Table 2 is a summary of Age at 1st CI/s, Age at 2nd CI, unaided PTAs (as described in Davidson et al, 2019) and device configurations at two sessions (early and adolescent), and device manufacturer.

Table 2.

Audiological characteristics of cochlear implant (CI) participants.

N mean (sd) min – max
Age at CI surgery (years)
Age at 1st CI surgery 86 2.0 (1.1) 8 months – 4.5 years
Age at 2nd CI surgery 74 3.2 (2.3) 8 months – 10.1 years
 
Unaided PTA§ (dB HL) at Early Session
Bimodal (CI and hearing aid) 23 77.4 (16.4) 41.7 – 100
Sequentially-implanted bilateral CIs 45 94.2 (18.2) 44.5 – 120
Simultaneously-implanted bilateral CIs 18 104.1 (15.1) 71.7 – 125
Unaided PTA§ (dB HL) at Adolescent Session
Bimodalǂ 12 77.2 (15.5) 41.7 – 96.7
Sequentially-implanted bilateral CIs* 56 91.0 (18.4) 44.5 – 120
Simultaneously-implanted bilateral CIs 18 104.1 (15.1) 71.7 – 125
 
Manufacturer
Cochlear 65 -- --
Advanced Bionics 19 -- --
MED-EL 2 -- --
Open in a new tab
§

PTA: 3-frequency (500, 1000, 2000 Hz) pure-tone average

ǂ

Two participants, who had used Bimodal devices at the Early Session, no longer used their hearing aids at the Adolescent Session

*

One participant, who had sequentially-implanted bilateral CIs at the Early Session, no longer used their 2nd CI at the Adolescent Session

Test Battery

All speech perception tests were presented in person from a laptop computer to a loudspeaker placed at 0° azimuth approximately 1 m from the child at a calibrated level of 60 dBA. Due to the COVID-19 pandemic, reading tests were given to participants virtually via a live-video chat platform. Before the start of the test, an approved tester confirmed that the subject: i) was wearing their device/s, ii) was in a quiet space free of distractions, and iii) had a reliable internet connection. The tester also ensured that the subject could hear the test instructions at a comfortable level. This was accomplished by showing the subject a loudness rating system and having them choose their most comfortable listening level using a 7-point scale ranging from too soft to too loud. Last, the tester spoke to the child’s parent or guardian, if available, to check whether there were any questions regarding the test procedure.

Segmental speech perception.

The Lexical Neighborhood Test (LNT) (Kirk, Pisoni, & Osberger, 1995), an open-set, 50-word monosyllabic word identification test, was administered. One list was administered in quiet and another list in noise (multi-talker babble, at a signal-to-noise ratio of +8 dB at 0° azimuth). Using typical clinical procedures, a clinician recorded the participant’s verbal responses in real-time (phonemes and words were indicated on the scoresheet). Later, two trained laboratory associates reviewed and verified scoring. The On-Line Imitative Test of Speech-Pattern Contrast Perception (OlimSpac) (Boothroyd, Eisenberg, & Martinez, 2010) assessed phoneme perception using nonsense syllables. The stimuli consisted of vowel–consonant–vowel (VCV) combinations of three vowels {/ɑ, i, u/} and 12 consonants {/b, d, p, t, f, v, s, ʃ, z, ʒ, ʧ, ʤ/}. The child listened to and imitated each VCV presented in quiet in an auditory-alone condition. The examiner, who was “acoustically blinded” to the test presentation (by the presence of a brief masking noise), listened to the child’s imitation and chose the syllable sequence, from a closed-set of options, that best represented the child’s utterance. Two lists of 16 nonsense syllables were presented. The accuracy of six contrasts, three from each VCV list, were scored automatically in the program (a total of 96 contrasts): vowel height, vowel place, consonant voicing, consonant continuance, pre-alveolar consonant place, and post-alveolar consonant place. An overall percent correct contrast score, adjusted for chance performance, was obtained.

Suprasegmental speech perception.

Talker discrimination was assessed using a subset of sentences produced by eight female talkers from the Indiana multi-talker speech database (Bradlow, Torretta, & Pisoni, 1996 ; Karl & Pisoni, 1994). In each trial, two sentences were presented and the child was asked to respond whether the sentences were spoken by the same talker or by different talkers. Regardless of talker(s) (i.e., same or different), the two sentences were always different (e.g., “The bill was paid every third week.” and “The box was thrown beside the parked truck.”). A percent correct score was calculated from a total of 32 trials. Stress discrimination was assessed using pairs of bisyllabic nonwords (CVCVs) (Thiessen & Saffran, 2007; Wenrich, Davidson, & Uchanski, 2017) with the same phonetic sequence (e.g., “diga”–“diga”) and with the same overall duration. The stress pattern of each bisyllabic nonword was either spondaic (“DI-GA”), trochaic (“DI-ga”) or iambic (“di-GA”). Pairs of nonwords were presented with either the same stress or different stress patterns, and the child reported whether the two nonwords had same or different stress patterns. Syllabic stress was represented acoustically by elongated duration, increased intensity, and increased fundamental frequency of the stressed relative to the unstressed syllable. A percent correct score was calculated from a total of 30 trials. Emotion identification was assessed using sentences spoken by a female talker with four emotional intentions: angry, scared, happy, and sad (Davidson et al., 2019; Geers, Davidson, Uchanski, & Nicholas, 2013). Sentences were simple in grammatical structure and vocabulary, and neutral with respect to emotion (“It’s time to go.”; “Give me your hand.”). Thirty-six trials were presented, and a percent correct score was calculated. APEX3 software was used to administer the talker, stress, and emotion tests (Francart, van Wieringen, & Wouters, 2008). An adapted and shortened version of the Children’s Test of Nonword Repetition was used to assess perception of syllable stress (Dillon, Burkholder, Cleary, & Pisoni, 2004). The test list consists of 20 nonwords, five each with two, three, four, and five syllables. These 20 nonwords were played in random order. Children were told that they should repeat the “funny word”. Their responses were recorded. Later, nonword productions were transcribed by a team consisting of two speech-language pathologists and one audiology graduate student with phonetic-transcription experience. The percentage of nonwords produced with the correct stress pattern, regardless of phoneme accuracy, was calculated.

Reading.

The Passage Comprehension test from The Woodcock Reading Mastery Tests (Woodcock, 2011) was administered at both the elementary and adolescent sessions. Scores from this reading test are expressed in relation to a large normative sample of age-matched children with TH. A standard score of 100 reflects average age-appropriate performance, with a standard deviation of 15. The Passage Comprehension test measures a child’s ability to study a sentence or short passage, and to exercise a variety of comprehension and vocabulary skills in providing the missing word in the passage. During virtual testing, at adolescence, booklet materials were screen-shared with the participant in a Power-Point format. The tester used the instructions from the test manual to administer the tests. The subject’s answers were then recorded onto the test protocol.

Analyses.

The reading passage comprehension scores obtained at elementary age were compared to the scores obtained at adolescence using student’s t -test. Regression models were used to examine the effects of speech perception scores on passage comprehension scores after controlling for gender, NVIQ, maternal education, age at 1st CI, and early device configuration. We used device configuration from the earliest ages (5-9 years) as a predictor variable; results from Nittrouer et al. (2012, 2018) indicate that early device-use, not necessarily device-use at the time of test, is a more important predictor of reading. Passage comprehension scores from adolescence were the dependent variable, and segmental and suprasegmental speech perception scores were independent variables. In the first model, the early speech perception scores are used as independent variables, while in the second model elementary speech perception scores are used as independent variables.

This research was approved by the Human Studies Committee at Washington University School of Medicine (IRB ID 201305136).

Results

Reading comprehension standard scores for individual participants from the elementary and adolescent sessions are shown in Figure 2. The adolescent scores are shown with bars and are rank-ordered from low to high; the elementary scores are shown with circles. Horizontal green lines indicate one standard deviation above (115) and below (85) the standard mean of 100. Reading scores range from 70 to 143. The mean scores from the elementary (M = 101.6, SD= 15.1) and adolescent sessions (M =101.7, SD= 16.9) were not significantly different t(85) = .05), p = .48. Both means are within the normative range. Sixty-nine (80%) of these adolescent participants scored within or above the normative range (49 participants’ adolescent scores within +/− 1 SD of the mean; 20 participants’ adolescent scores > 1 SD above the mean). At the elementary session, 74 participants (86%) had scored within or above the normative range (61 participants’ elementary scores within +/− 1 SD of the mean; 13 participants’ elementary scores > 1 SD above the mean). As these children developed, there are slight increases in the number who scored below the normative range and in those who score above the normative range. Specifically, there was an increase from 12 participants (~14%) with reading scores < 85 at elementary ages to 17 participants (~20%) with scores below average at adolescence. At the other extreme, there was an increase from 13 participants (~15%) with reading scores > 115 at the elementary session to 20 participants (~23%) with above average scores at adolescence.

Figure 2.

Figure 2.

Open in a new tab

Reading comprehension scores for children with CIs at elementary age (blue dots) and adolescent age (orange bars). Horizontal green lines indicate one standard deviation above (115) and below (85) the standard mean of 100.

Table 3 shows speech perception scores, from the early (age at test: mean =7.0 years, SD =1.3) and elementary sessions (age at test: mean=9.0 years, SD=1.3). Scores from both segmental and suprasegmental speech perception tests are used in the regression analyses. Mean scores for the tests in the segmental battery [LNT word recognition in quiet and in noise, and phoneme contrast scores in quiet (OlimSpac)] ranged from 66% to 86% correct at the early session and from 67% to 92% correct at the elementary session. Mean scores for the tests in the suprasegmental battery (Stress Discrimination, Talker Discrimination, Emotion Identification, and Nonword Stress) ranged from 60% to 79% correct at the early session and from 70% to 91% correct at the elementary session. Principal components analysis (PCA) was used to create composite scores, one each for segmental and suprasegmental speech perception (see Grantham, et al. 2022 for details). These composite scores were utilized to reduce the number of potentially redundant statistical tests, reduce the probability of Type I errors, and increase the reliability of measures used in subsequent analyses.

Table 3.

Segmental and Suprasegmental Speech Perception Scores (Percent-Correct) for cochlear implant participants from Early and Elementary test ages (n=86).

Early Elementary
mean (sd) min – max mean (sd) min –max
Segmental
LNT in Quiet 84.5 (11.7) 38 – 100 88.7 (7.8) 56 – 100
LNT in Noise 66.4 (19.5) 0 – 92 67.3 (12.8) 26 – 86
OlimSpac* scores 86.3 (12.0) 46 – 100 92.3 (6.4) 69 – 100
Suprasegmental
Stress Discrimination 79.3 (17.1) 27 – 100 87.8 (12.6) 47 – 100
Talker Discrimination 60.7 (12.2) 34 – 94 69.9 (11.1) 42 – 91
Emotion Identification 60.2 (24.4) 17 – 97 78.1 (19.7) 19 – 100
Non-Word Repetition of Stress 75.0 (23.3) 5 – 100 90.8 (12.8) 45 – 100
Open in a new tab
*

OlimSpac is the On-Line Imitative test of Speech-Pattern Contrast perception

Effects of Speech Perception, Early vs Elementary, on Adolescent Reading Comprehension

A multi-step regression analysis was employed to examine the contribution of speech perception to reading comprehension scores of adolescent CI participants. The first step included three demographic variables (gender, NVIQ, and maternal education), the second step included device variables (age at 1st CI and device configuration) at the early session and the third step included the segmental perception composite scores. The suprasegmental perception composite score was added at the fourth (final) step. Two separate regression analyses were conducted, the first, ‘Model 1’, uses speech perception scores from the early session in steps 3 & 4, while the second, ‘Model 2’, uses speech perception scores from the elementary session in steps 3 & 4.

Table 4 provides the coefficients, standard errors, and significance values for each predictor variable in the regression model, as well as the F-values, significance (p), and multiple R2 values for the regression model at each step. Models 1 and 2 are identical for the first two steps. At step 1, the three demographic variables (gender, NVIQ, maternal education) account for 30% of the variance in reading scores. NVIQ is the only statistically significant predictor. At step 2 of the model, age at 1st CI and device configuration contribute an additional 3% of the variance in reading scores. Although, on average, children who were early bimodal users have a higher mean reading score [107] than children who were sequential or simultaneous bilateral CIs users [99 and 101, respectively], neither of the two device variables is statistically significant in the model. For Model 1, at step 3, segmental perception from the early session contributed a mere 1% additional variance in reading and is not statistically significant. At step 4, however, early suprasegmental speech perception contributed an additional 8% of reading variance (total variance 41%) and is a statistically significant predictor. For Model 2, neither segmental nor suprasegmental perception, from the elementary session, is a statistically significant contributor (total variance 35%). Segmental scores from neither the early nor elementary sessions contribute significantly to adolescent reading scores. By contrast, early suprasegmental perception contributes a significant additional (8%) variance to reading scores, while elementary suprasegmental perception is not a significant predictor.

Table 4.

Regression Model statistics for predicting Passage Comprehension

Models 1 & 2 Regression Coefficient Regression Model
Step Predictor Variable: Estimate Std. Error t value Pr ( >∣t∣ ) F value p value Multiple R2
1 (Intercept) 40.97 14.19 2.89 .005 F(3,82) = 11.6 < .001 .298
1 Gender −2.56 3.22 −0.80 0.426
1 Block Design, NVIQ 0.47 0.098 4.85 < .001
1 Maternal Education 0.89 0.78 1.14 .258
2 Age at 1st CI −0.82 1.75 −0.47 .6404 F(6,79) = 6.43 < .001 .328
2 Early Device Configuration Con_2* −7.77 4.30 −1.81 .0746
2 Early Device Configuration Con_3* −6.60 5.42 −1.22 .2270
Model 1: Early Speech Perception
3 Early Segmental Perception 1.89 2.053 0.92 .3601 F(7, 78) = 5.63 < .001 .335
4 Early Suprasegmental Perception 6.42 2.047 3.14 .0024 F(8,77) = 6.71 < .001 .411
Model 2: Elementary Speech Perception
3 Elementary Segmental Perception 1.029 1.732 0.59 .554 F(7, 78) = 5.52 < .001 .331
4 Elementary Suprasegmental Perception 3.323 2.223 1.49 .1390 F(8,77) = 5.19 < .001 .350
Open in a new tab
*

Con_2 and Con_3 are dummy-coded categorical variables, ‘Device Configuration’ (bimodal, sequential CIs, simultaneous CIs).

Discussion & Conclusion

These data reinforce recent positive trends in reading achievement by children who received early CIs (Mayer et al., 2018, Wang et al., 2021). On average these adolescent CI users scored within the normative range for reading comprehension, as they had when tested at elementary ages (Grantham et al., 2022). Thus, the answer to our Question (1) is ‘yes.’ Additionally, this group of CI users’ average standard reading score is slightly higher (102) than those reported in earlier studies (Spencer et al. 2004, Geers & Hayes, 2011, Edwards et al., 2016) in which average reading scores range from 82 to 92. Compared to these aforementioned studies the mean age at first CI for these adolescent CI users was somewhat younger (2.2 years vs. ~3.5 years). Moreover, all but three participants in this study used bilateral devices (2CIs or bimodal) compared to the predominantly unilateral device use (one CI only) in the earlier studies. Individually, 80% of these adolescent participants scored within or above the normative range on reading comprehension (i.e., > 85); this percentage is similar to the percentage (76%) reported by Mayer and colleagues (2016) and greater than the 66% reported by Geers and Hayes (2011). Notably, ~57% of the participants in the Mayer study used 2CIs. When examined over time, these results are similar to those of Geers et al. (2008), in that a slightly higher proportion of participants fail to achieve age-appropriate reading scores (<85) at adolescence, or high school (20%), compared to elementary ages (14%). Interestingly, the percentage of participants scoring above the normative range (>115) also increased at adolescence (23%) compared to elementary ages (15%), indicating a ‘hollowing out’ of the middle.

Similar to results from Wang and colleagues, age at 1st CI was not a significant predictor for reading. While Wang et al. did find that unilateral CI use was related to poorer outcomes, the majority of children in this study used two CIs, or a CI and HA combined. Similar to results from Nittrouer and colleagues (2012) who tested emergent literacy skills at kindergarten age, there was a trend for children who had an early period of bimodal use to have higher reading scores. Device configuration, however, was not statistically significant with this adolescent sample. For this current study, comparisons across device-groups should be interpreted with caution due to the small and uneven numbers of participants in each group. Finally, NVIQ was a significant predictor of reading comprehension as has been found in other studies (e.g., Geers and Hayes, 2011).

Examination of suprasegmental speech perception skills may be especially important for children with CIs due to the limits of current CI systems in transmitting acoustic cues related to prosody. Grantham and colleagues (2022) found that early segmental perception was a significant predictor of reading decoding but not of reading comprehension, while early suprasegmental perception was a significant predictor of both decoding and comprehension. Based on these past results, we were interested in determining whether the statistically significant influence of early segmental and suprasegmental perception on reading still holds for these same participants’ reading comprehension skills at adolescence. For these 86 adolescents with CIs, suprasegmental scores obtained at early ages contributed significantly to their reading comprehension after controlling for demographic variables, audiological variables, and early segmental perception. By contrast, early segmental speech perception was not a significant contributor to adolescent reading comprehension. Thus, early suprasegmental speech perception is predictive of both elementary and adolescent reading comprehension, in a positive manner; good early suprasegmental perception is associated with good near- and long-term reading comprehension. Thus, the answer to our Question (2) is ‘yes’ - for reading comprehension (we did not assess reading decoding skills at adolescence and, hence, cannot compare effects of early speech perception on decoding at elementary and adolescent ages). Segmental speech perception is likely most involved in early reading and word-attack skills, including phonological awareness, phoneme recognition and vocabulary. Unlike word-attack and decoding skills, reading comprehension likely depends more on global language comprehension, which may rely on good prosodic perception for detecting phrase and lexical boundaries.

We were also interested in examining whether elementary speech perception skills measured two years later than early skills would influence adolescent reading skills. Here, we found that neither elementary segmental nor suprasegmental speech perception skills contributed significant variance to adolescent reading comprehension. Thus, the answer to our Question (3) is ‘yes’ – effects of early vs elementary speech perception on adolescent reading differ. It appears that children receive a long-term benefit for reading from having had good suprasegmental speech perception at early ages. Suprasegmental speech perception ability at later ages (i.e., elementary), however, is not predictive of adolescent reading. Either this relationship weakens or is diminished compared to other factors such as development of world knowledge. As stated by Wade-Woolley and colleagues (2022), there is a need for longitudinal studies that concurrently assess both prosody (perception) and reading skills in children with TH and with hearing loss.

Speech perception skills provide the underpinnings for early literacy. Furthermore, for children who use CIs or HAs, speech perception assessments are fundamental to guiding all aspects of habilitation including device candidacy, device validation, and therapy goals. Current clinical device validation protocols and CI candidacy guidelines focus primarily on segmental speech perception while suprasegmental perception is largely ignored. Given that both CI and HA technology degrade, though in different ways, acoustic properties associated with suprasegmental information (Karimi-Boroujeni, et al., 2023), a greater understanding of the suprasegmental speech perception abilities of children, with a wide range of HL using HAs and CIs, is needed. Results from this study and from our previous work (Grantham et al., 2022) highlight the importance of assessing suprasegmental skills at early ages. We suggest that an expanded view of speech perception, that includes both segmental and suprasegmental perception, has broad implications for device technology, targeted therapies and device decisions, which may ultimately improve language and literacy skills of children with HL.

Strengths and Limitations

Two strengths of this study are: 1) the fairly large number of participants and 2) the reporting of longitudinal reading data, which includes skills assessed at adolescent ages when readers are well beyond rudimentary decoding, etc. Limitations to the generalizability of these results include participating families with higher SES than found in the overall US population and the predominance of oral educational intervention during infancy and early childhood.

Acknowledgements:

This work is supported by NIDCD grant RO1 DC012778. Appreciation is expressed to the students, families and staff from the following sites: Arkansas Children’s Hospital, Central Institute for the Deaf, Child’s Voice, Children’s Hospital of Philadelphia-Philadelphia, Children’s Choice for Hearing and Talking, Hearts for Hearing, Listen and Talk, Memphis Oral School for the Deaf, Moog Center for Deaf Education, Ohio Valley Voices, Presbyterian Ear Institute, University of Miami, University of Minnesota Children’s Hospital, University of Texas at Dallas, Vanderbilt University.

Footnotes

All authors have no conflicts related to this work.

1

Grantham et al. (2022) present elementary reading scores for 104 children with CIs; the 86 children with CIs, here, are a proper subset of those 104.

References

  1. Boothroyd A, Eisenberg LS, & Martinez AS (2010). An on-line imitative test of speech-pattern contrast perception (OlimSpac): Developmental effects in normally hearing children. Journal of Speech, Language and Hearing Research, 53(3), 531–542. doi: 10.1044/1092-4388(2009/08-0260) [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bradlow AR, Torretta GM, & Pisoni DB (1996). Intelligibility of normal speech I: Global and fine-grained acoustic-phonetic talker characteristics. Speech Communication, 20(3), 255–272. doi: 10.1016/S0167-6393(96)00063-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Carroll J, & Zeng F-G (2007). Fundamental frequency discrimination and speech perception in noise in cochlear implant simulations. Hearing Research, 231(1), 42–53. doi: 10.1016/j.heares.2007.05.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Ching T, Galaster J, Grimes A, Johnson C, Lewis D, McCreery R, & Yoshinga-Itano C (2013). American Academy of Audiology clinical practice guidelines: Pediatric amplification. Reston, VA: American Academy of Audiology. [Google Scholar]
  5. Davidson LS, Geers AE, Uchanski RM, & Firszt JB (2019). The Effects of Early Acoustic Hearing on Speech Perception and Language for Pediatric Cochlear Implant Recipients. Journal of Speech, Language and Hearing Research, 62(9), 3620–3637. doi: 10.1044/2019_JSLHR-H-18-0255 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Dillon CM, Burkholder RA, Cleary M, & Pisoni DB (2004). Nonword repetition by children with cochlear implants: Accuracy ratings from normal-hearing listeners. Journal of Speech, Language and Hearing Research, 47(5), 1103–1116. doi: 10.1044/1092-4388(2004/082) [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Dorman MF, Gifford RH, Spahr AJ, & McKarns SA (2008). The benefits of combining acoustic and electric stimulation for the recognition of speech, voice and melodies. Audiology and Neurotology, 13(2), 105–112. doi: 10.1159/000111782 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Edwards L, Aitkenhead L, & Langdon D (2016). The contribution of short-term memory capacity to reading ability in adolescents with cochlear implants. International Journal of Pediatric Otorhinolaryngology, 90, 37–42. doi: 10.1016/j.ijporl.2016.08.017 [DOI] [PubMed] [Google Scholar]
  9. Francart T, van Wieringen A, & Wouters J (2008). APEX 3: A multi-purpose test platform for auditory psychophysical experiments. Journal of Neuroscience Methods, 172(2), 283–293. doi: 10.1016/j.jneumeth.2008.04.020 [DOI] [PubMed] [Google Scholar]
  10. Gantz BJ, Turner C, Gfeller KE, & Lowder MW (2005). Preservation of hearing in cochlear implant surgery: Advantages of combined electrical and acoustical speech processing. The Laryngoscope, 115(5), 796–802. doi: 10.1097/01.MLG.0000157695.07536.D2 [DOI] [PubMed] [Google Scholar]
  11. Geers A, Davidson L, Uchanski R, & Nicholas J (2013). Interdependence of linguistic and indexical speech perception skills in school-aged children with early cochlear implantation. Ear and Hearing, 34(5), 562–574. doi: 10.1097/AUD.0b013e31828d2bd6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Geers A, Tobey E, Moog J, & Brenner C (2008). Long-term outcomes of cochlear implantation in the preschool years: From elementary grades to high school. International Journal of Audiology, 47(Suppl 2), S21–S30. doi: 10.1080/14992020802339167 [DOI] [PubMed] [Google Scholar]
  13. Geers AE, & Hayes H (2011). Reading, writing, and phonological processing skills of adolescents with 10 or more years of cochlear implant experience. Ear and Hearing, 32(1 suppl), 49S–59S. doi: 10.1097/AUD.0b013e3181fa41fa [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Grantham H, Davidson LS, Geers AE, & Uchanski RM (2022). Effects of Segmental and Suprasegmental Speech Perception on Reading in Pediatric Cochlear Implant Recipients. Journal of Speech, Language and Hearing Research, 65(9), 3583–3594. doi: 10.1044/2022_JSLHR-22-00035 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Karimi-Boroujeni M, Dajani HR, & Giguère C (2023). Perception of Prosody in Hearing-Impaired Individuals and Users of Hearing Assistive Devices: An Overview of Recent Advances. Journal of Speech, Language and Hearing Research, 66(2), 775–789. doi: 10.1044/2022_jslhr-22-00125 [DOI] [PubMed] [Google Scholar]
  16. Karl JR, & Pisoni DB (1994). Effects of stimulus variability on recall of spoken sentences: A first report. Retrieved from http://www.iu.edu/~srlweb/publications/progress-report-19/
  17. Kirk KI, Pisoni DB, & Osberger MJ (1995). Lexical effects on spoken word recognition by pediatric cochlear implant users. Ear and Hearing, 16(5), 470–481. doi: 10.1097/00003446-199510000-00004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Kutner MG,E; Jin Y; Boyle B; Hsu Y; Dunleavy E (2007). Literacy in Everday Life: Results from the 2003 National Assessment of Adult Literacy. NCES. (480). US Department of Education. [Google Scholar]
  19. Mayer C, & Trezek BJ (2018). Literacy outcomes in deaf students with cochlear implants: Current state of the knowledge. Journal of Deaf Studies and Deaf Education, 23(1), 1–16. doi: 10.1093/deafed/enx043 [DOI] [PubMed] [Google Scholar]
  20. Mayer C, Trezek BJ, & Hancock GR (2021). Reading Achievement of Deaf Students: Challenging the Fourth Grade Ceiling. Journal of Deaf Studies and Deaf Education, 26(3), 427–437. doi: 10.1093/deafed/enab013 [DOI] [PubMed] [Google Scholar]
  21. Mayer C, Watson L, Archbold S, Ng ZY, & Mulla I (2016). Reading and Writing Skills of Deaf Pupils with Cochlear Implants. Deafness & Education International, 18(2), 71–86. doi: 10.1080/14643154.2016.1155346 [DOI] [Google Scholar]
  22. National Reading Panel; National Institute of Child Health Human Development. (2000). 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, National Institutes of Health. [Google Scholar]
  23. Nittrouer S., et al. , Emergent Literacy in Kindergartners With Cochlear Implants. Ear Hear, 2012. 33(6): p. 683–697. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Nittrouer S, Muir M, Tietgens K, Moberly AC, & Lowenstein JH (2018). Development of phonological, lexical, and syntactic abilities in children with cochlear implants across the elementary 608 grades. Journal of Speech, Language and Hearing Research, 61(10), 2561–2577. 10.1044/2018_JSLHR-H-18-0047 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Spencer LJ, Gantz BJ, Knutson JF. Outcomes and achievement of students who grew up with access to cochlear implants. Laryngoscope. 2004. Sep;114(9):1576–81. doi: 10.1097/00005537-200409000-00014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Spencer L, & Oleson J (2008). Early Listening and Speaking Skills Predict Later Reading Proficiency in Pediatric Cochlear Implant Users. Ear and Hearing, 29(2), 270–280. doi: 10.1097/01.aud.0000305158.84403.f7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Thiessen ED, & Saffran JR (2007). Learning to learn: Infants’ acquisition of stress-based strategies for word segmentation. Language Learning and Development, 3(1), 73–100. doi: 10.1080/15475440709337001 [DOI] [Google Scholar]
  28. Thoutenhoofd E. (2006). Cochlear Implanted Pupils in Scottish Schools: 4-Year School Attainment Data (2000–2004). The Journal of Deaf Studies and Deaf Education, 11(2), 171–188. doi: 10.1093/deafed/enj029 [DOI] [PubMed] [Google Scholar]
  29. Uhler K, Warner-Czyz A, Gifford R, & PMSTB Working Group. (2017). Pediatric Minimum Speech Test Battery. Journal of the American Academy of Audiology, 28(3), 232–247. doi: 10.3766/jaaa.15123 [DOI] [PubMed] [Google Scholar]
  30. Van der Kant A, Vermeulen A, De Raeve L, & Schreuder R (2010). Reading Comprehension of Flemish Deaf Children in Belgium: Sources of Variability in Reading Comprehension after Cochlear Implantation. Deafness & Education International, 12(2), 77–98. doi: 10.1179/146431510X12626982043769 [DOI] [Google Scholar]
  31. Wade-Woolley L, Wood C, Chan J, & Weidman S (2021). Prosodic Competence as the Missing Component of Reading Processes Across Languages: Theory, Evidence and Future Research. Scientific Studies of Reading, 26(2), 165–181. 10.1080/10888438.2021.1995390 [DOI] [Google Scholar]
  32. Wang Y, Sibaii F, Lee K, Gill MJ, & Hatch JL (2021). Meta-Analytic Findings on Reading in Children With Cochlear Implants. Journal of Deaf Studies and Deaf Education, 26(3), 336–350. doi: 10.1093/deafed/enab010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Wechsler D. (2003). Wechsler Intelligence Scale for Children - Fourth Edition (WISC-IV). San Antonio, TX: Pearson. [Google Scholar]
  34. Wechsler D. (2011). Wechsler abbreviated scale of intelligence - second edition (WASI-II). San Antonio, TX: NCS Pearson. [Google Scholar]
  35. Wenrich KA, Davidson LS, & Uchanski RM (2017). Segmental and suprasegmental perception in children using hearing aids. Journal of the American Academy of Audiology, 28(10), 901–912. doi: 10.3766/jaaa.16105 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Woodcock RW (2011). Woodcock reading mastery tests - Third edition (WRMT III). Bloomington, MN: Pearson. [Google Scholar]
  37. Zhang J, & McBride-Chang C (2010). Auditory sensitivity, speech perception, and reading development and impairment. Educational Psychology Review, 22(3), 323–338. doi: 10.1007/s10648-010-9137-4 [DOI] [Google Scholar]

RESOURCES