Impacts of hearing history, etiology, vestibular and balance function, and socioeconomic marginalization on developmental outcomes in children with cochlear implants Authors: Melissa Hazen, Sharon L Cushing, Karen A Gordon Published: 2026 ABSTRACT AbstractChildren with bilateral cochlear implants (BCIs) remain vulnerable to developmental challenges relative to typically developing peers (TD), even when access to sound through acoustic or amplified hearing is achieved within the first year of life. This study evaluated contributions of hearing history, etiology, vestibular/balance function, and socioeconomic marginalization to language, working memory, and academic outcomes. Ninety-six children aged 4.65–17.85 years participated: 66 with BCIs (mean [SD] age, 11.54 [3.57]) and 30 typically developing (TD) peers (mean [SD], 11.69 [2.68]). Standardized assessments included the CELF, Dot Matrix, Corsi Block, Digit Span, and WIAT-III subtests. Regression and principal component analysis (PCA) identified predictors of developmental outcomes. Children with BCIs scored significantly lower than TD peers in language (p = 0.003), visuospatial working memory (p = 0.001), math (p < 0.001), and word reading (p = 0.048). PCA identified four components: hearing loss history, auditory experience/resources, social marginalization, and vestibular/balance function. Only auditory experience predicted developmental outcomes across domains (p’s < 0.05). Vestibular and balance function were impaired in the BCI group (p < 0.001) but did not predict language, working memory, or academic scores. Deficits were most pronounced in children with congenital cytomegalovirus, cochleovestibular anomalies, and genetic hearing loss. Results emphasize the importance of sustained early auditory access together with etiology-informed, multidisciplinary support to optimize developmental outcomes in children with BCIs.Subject terms: Health care, Medical research, Neurology, Neuroscience Introduction This study aimed to identify the developmental contributions of hearing history including etiology, vestibular/balance function and socioeconomic marginalization on language, working memory, and academic performance in children with bilateral cochlear implants (BCIs). Cochlear implants (CIs) significantly improve speech perception, language, and academic outcomes in children with severe-to-profound hearing loss by restoring auditory input during critical periods of brain development1,2. Optimal developmental trajectories begin with early auditory access, defined as the point at which a child first receives consistent, functional audibility through acoustic, amplified, or electrical hearing during the first year of life3–9. The Joint Committee on Infant Hearing (JCIH) recommends well-fitted hearing aids by ~ 6 months of age10, as infants who achieve aided audibility within this window demonstrate stronger early language outcomes than those amplified later7,11,12. Early implantation, particularly before 12 months of age7,12–15, further optimizes neuroplasticity and promotes better language outcomes9,16,17. Group level differences in cognitive and academic outcomes are frequently observed, even among children who receive hearing devices early, indicating persistent variability in development1,2,18–25. These challenges likely reflect not only degraded auditory input and increased effort required to process sound21,26–28, but also reduced access to incidental spoken language exposure18,23,29,30, greater reliance on executive resources to support listening and learning31, and cortical reorganization associated with early auditory deprivation and subsequent CI use14,32. Similar cognitive challenges in adults with hearing loss33–35 have been linked to increased dementia risk36. To better understand persistent delays, additional factors that may interact with auditory challenges must be considered. Developmental outcomes in children with hearing loss rehabilitated with CI, are likely influenced by clinical factors beyond age of implantation. These include duration of auditory deprivation, pre-implant auditory experience (e.g., use of hearing aids), exposure to hearing and language, and timing and consistency of habilitation/therapy1,29,30,37–41. Socioeconomic marginalization may further compound these risks by increasing the likelihood of delayed diagnosis, inconsistent rehabilitation, and reduced access to language-rich environments18,23,25,42–47. How socioeconomic factors affect CI outcomes remains unclear. The effects of vestibular impairment (VI) on development after CI in children should be considered. VI is common in children with hearing loss who receive CIs, particularly in those with etiologies such as congenital cytomegalovirus (cCMV), cochleovestibular anomalies, and syndromic deafness, but remains an under-recognized comorbidity48–51. Balance deficits occur in 30–70% of children with hearing loss48,52–54 and can delay motor milestones53,55. VI could impact language and cognition by delaying motor development, spatial orientation, or multisensory integration, though evidence is limited. Janky and colleagues (2022) found no differences in language or academic outcomes between children with hearing loss rehabilitated by CIs who did or did not have VI56 whereas Van Hecke et al. (2023) reported broader executive and perceptual-motor deficits, but no differences in language or working memory57. In young adult CI users with VI, Benjamin et al. (2023) found balance and working memory impairments compared to typically developing (TD) children58. As understanding of the mechanisms underlying pediatric hearing loss advances, it is possible to assess whether outcomes of CI differ based on the etiology of deafness59. For example, auditory neuropathy spectrum disorder is associated with poorer speech perception due to disrupted neural transmission60, whereas cochlear ossification secondary to meningitis can complicate electrode insertion and reduce CI effectiveness49. In GJB2-related deafness, the most common genetic hearing loss61, isolated cochlear pathology often predicts strong language and hearing outcomes49,62. By contrast, syndromic etiologies like Waardenburg and Usher syndrome frequently involve vestibular dysfunction and cognitive or motor delays, that might limit hearing and language outcomes46,48,49. Congenital cytomegalovirus (cCMV) can also cause widespread damage to the central and peripheral nervous systems, including white matter abnormalities and cortical disruption, which may compromise not only auditory processing but also working memory, language acquisition, and academic skills50,62–65. Thus, it is important to understand how specific etiologies influence developmental outcomes in children with CIs. The combined influence of hearing history, etiology, vestibular/balance function, and socioeconomic marginalization on developmental outcomes in children with CIs was examined in the present study. Two hypotheses were tested: (1) children with BCIs exhibit age-independent impairments in outcomes of language, working memory, and academic performance, as well as increased vestibular/balance dysfunction; and (2) these challenges relate to hearing history and socioeconomic marginalization, with vestibular/balance dysfunction playing an additive role. Materials & methods This cross-sectional observational study was conducted with approval from the Research Ethic Board (REB 1000079136) at the Hospital for Sick Children and informed consent was obtained in accordance with institutional guidelines. All methods were performed in accordance with the relevant guidelines and regulations. Participants In this study, 100 children aged 4.65 to 17.85 years were initially evaluated, however four typically developing participants were excluded due to inconclusive vestibular results. A total of 96 participants (48 Female: 48 Male) were included in analyses: 30 children with typical development (TD) (mean [SD] age = 11.69 [2.86] years), and 66 children using bilateral cochlear implants (BCI) (mean [SD] age = 11.54 [3.47] years). Among the BCI users, 47 children received simultaneous implantation (mean [SD] age at implantation = 3.08 [3.40] years), while 19 underwent sequential implantation, with mean ages of 3.91 (3.67) years for the first CI and 7.57 (2.91) years for the second. Socioeconomic marginalization was quantified using the Ontario Marginalization Index (ON-MARG)66 linked to postal codes. Inclusion criteria required English proficiency, the ability to complete all assessments, and the absence of significant cognitive or mobility impairments. Use of manual communication (e.g., sign language) as the primary mode of communication was considered exclusionary. Details regarding the support received for hearing and speech-language development were available in all but 3 children through medical chart review, revealing therapy that emphasized spoken communication in all cases; 52/63 (83%) had received this intervention through an auditory–verbal therapist (AVT), and 11/63 (17%) from a speech-language pathologist (SLP). The median age at initiation was 12 months for AVT (IQR: 9.5–36) and 24 months for SLP (IQR: 7–48). Details regarding where this intervention was delivered was available in 60 children. In general, AVT and SLP services were accessed through publicly funded early hearing intervention services, with care transitioning from early childhood programs to school-based services. Services were provided in community settings for 39/60 children (65%), at home for 9/60 children (15%), through hospital programs for 9/60 children (15%), and through school services for 3/60 children (5%). Information regarding the duration, intensity, and frequency of language intervention was not available in the medical record. Procedures Participants completed standardized outcome measures in three domains: (1) Language, (2) Working memory, and (3) Academics. Language abilities were assessed with the Clinical Evaluation of Language Fundamentals, Fifth Edition (CELF). Auditory-verbal and visuospatial working memory were measured by Digit Span, the Dot Matrix, and the Corsi Block Tapping tests which all assessed participants’ ability to recall and manipulate sequential information. Academic performance was evaluated using subtests from the Wechsler Individual Achievement Test, Third Edition (WIAT-III) math, word reading, and pseudoword reading subtests. Hearing in the typically developing group was confirmed by pure-tone audiometry (250–8000 Hz), tympanometry, and when indicated, acoustic reflex and otoacoustic emission testing. Children in the BCI group were followed clinically at our centre. Vestibular function was evaluated using the video head impulse test (vHIT) for vestibulo-ocular reflex (VOR) and cervical vestibular evoked myogenic potentials (cVEMP) for vestibulo-collic reflex (VCR). A composite score was calculated by summing right and left vHIT gains and cVEMP presence (1) or absence (0), divided by the total number of tests (4). This continuous score provided a more nuanced, though incomplete, measure of vestibular function. Balance was assessed using the Bruininks-Oseretsky Test of Motor Proficiency (BOT) balance subtest67, measuring static and dynamic postural control. Socioeconomic status was measured by the Ontario Marginalization Index (ON-MARG)66 which assesses marginalization across four dimensions: (1) Material Resources, (2) Households & Dwellings, (3) Age & Labour, and (4) Racialized & Newcomer Populations. Scores are ranked into quintiles, with higher scores indicating greater marginalization. Participant ON-MARG scores were determined using Dissemination Area Unique Identifiers (DAUIDs) linked to postal codes via the Postal Code Conversion File (PCCF)68,69. Data analyses All statistical analyses were conducted using R (v4.2.2; R Core Team 2022). Descriptive statistics summarized participant characteristics and task performance. Group and inferential analyses used multivariable linear regression models and principal component analysis (PCA). Linear regression models Regression models examined associations between age, group (BCI vs. TD), etiology, socioeconomic marginalization (ON-MARG), and age at start of AVT or SLP therapy on outcomes (language, working memory, academics). All models included interaction terms (e.g., age × group) and applied Type III ANOVA with Satterthwaite’s approximation. Post hoc contrasts were performed using estimated marginal means (emmeans package), with Tukey correction. Results are reported with 95% confidence intervals. Principal component analysis (PCA) PCA was used to reduce dimensionality and address collinearity among predictors in the BCI group. Variables included: duration of deafness, age at diagnosis, age of first cochlear implantation, inter-implant delay, onset of hearing loss per ear, years of acoustic hearing, balance score (BOT), vestibular function (%), and ON-MARG dimensions (material resources, racialized/newcomer populations, households/dwellings, age/labor force). Variables were standardized to unit variance. ON-MARG variables were reverse-scored so that higher values reflected reduced marginalization. PCA-derived components were entered into regression models to assess associations with developmental outcomes. Results Demographic characteristics of 30 typically developing (mean [SD] age, 11.69 [2.68] years) and 66 BCI users (mean [SD], 11.54 [3.57] years) are summarized in Table 1. ON-MARG scores were comparable between groups (Fig. 1), indicating no systematic differences in socioeconomic status. Table 1Demographic factors: hearing history, vestibular and balance function, and the Ontario marginalization index (ON-MARG). Demographic factors: hearing history, vestibular and balance function, and the Ontario marginalization index (ON-MARG). Age-related performance and group differences across developmental domains in children with bilateral cochlear implants (orange) and typically developing controls (blue). (A) Language, (B) Working Memory, and (C) Academic scores are shown. Left panels depict individual performance by age with regression lines; No significant (group × age) interactions. Right panels show boxplots comparing overall group performance with bars on significant differences. Both groups (TD, BCIs) showed age-related improvements across all outcome measures (Fig. 1). Linear regression models confirmed that performance improved significantly with age in language (β = 0.97; 95% CI, 0.54 to 1.41; P < 0.001), Dot Matrix (β = 0.59; 95% CI, 0.19 to 0.99; P = 0.004), Corsi Blocks (β = 0.61; 95% CI, 0.18 to 1.05; P = 0.006), Digit Span (β = 0.79; 95% CI, 0.39 to 1.19; P < 0.001), mathematics (β = 2.24; 95% CI, 1.68 to 2.80; P < 0.001), word reading (β = 3.73; 95% CI, 2.25 to 5.22; P < 0.001), and pseudoword reading (β = 2.39; 95% CI, 1.07 to 3.71; P = 0.001). Children with bilateral cochlear implants scored significantly lower than typically developing peers in the tests of language (mean difference, − 3.65; 95% CI, − 6.04 to − 1.27; P = 0.003); visuospatial working memory (Dot Matrix: mean difference, − 3.78; 95% CI, − 5.98 to − 1.58; P = 0.001), and academics (mathematics: mean difference, − 5.41; 95% CI, − 8.50 to − 2.33; P < 0.001; word reading: mean difference, − 7.11; 95% CI, − 14.15 to − 0.07; P = 0.048). After accounting for age at test, neither age at start of language intervention (β = −0.01, 95% CI [− 0.07, 0.04], P = 0.648), nor its interaction with hearing loss type (congenital vs. progressive) was significant. No significant (group × age) interactions were observed for any domain, suggesting that impaired scores in the BCI group remained stable across age. No significant group differences were found on the Corsi Blocks test (mean difference, 0.14; 95% CI, − 2.20 to 2.48; P = 0.906) or the Digit Span (mean difference, − 1.56; 95% CI, − 3.77 to 0.65; P = 0.165). Bilateral cochlear implants users tended to have lower Pseudoword reading scores than their typically developing peers although the difference did not quite reach statistical significance (mean difference, − 6.09; 95% CI, − 12.26 to 0.09; P = 0.053). Vestibular (β = -0.31; 95% CI, -0.44 to -0.18; P < 0.001) and balance function (β = -6.41; 95% CI, -8.75 to -4.07; P < 0.001) were significantly poorer in children with BCIs than the TD group. A positive but weak association between vestibular function and balance performance on the BOT (Fig. 2A) was found for children with bilateral cochlear implants (β = 9.04; 95% CI, 5.50 to 12.58; P < 0.001; R² = 0.28). Balance deficits in the bilateral cochlear implant group were most pronounced in children with Usher syndrome (β = −10.47; 95% CI, − 16.99 to − 3.96; P < 0.001), cCMV/infectious etiologies (β = −8.90; 95% CI, − 15.42 to − 2.39; P = 0.002), and Waardenburg syndrome (β = −10.90; 95% CI, − 17.09 to − 4.71; P = 0.013) (Fig. 2B). Children with these etiologies of hearing loss also showed reduced vestibulo-ocular reflex (VOR) and vestibulo-collic reflex (VCR) function (Fig. 2C, D, respectively). Associations between vestibular function, balance, and etiology. (A) Scaled balance scores (BOT) vs. composite vestibular function (%) for typically developing children (blue) and BCI users (orange). Dashed lines indicate normal thresholds. (B) Balance performance by etiology. Dashed horizontal line marks the clinical threshold for normal balance (BOT ≥ 10). Individual data points are overlaid and shaped by genetic subtype. (C) Vestibulo-ocular reflex (VOR) function (%) and (D) Vestibulo-collic reflex (VCR) function (%) across etiological groups, ordered by balance scores (worst to best). A principal component analysis (PCA) was conducted to reduce dimensionality across multiple potential predictors of developmental outcomes. The PCA identified four components: PC1 (Hearing Loss History), PC2 (Hearing & Resources), PC3 (Social Marginalization), and PC4 (Vestibular & Balance) (Fig. 3A). Hearing loss history variables included duration of deafness (from onset of hearing loss to first consistent auditory access ≤ 40 dB HL through hearing devices), age at diagnosis, onset of hearing loss (congenital, early childhood, or progressive), age at implantation, inter-implant delay (for sequential implantation), and cumulative years in sound (total years with auditory access ≤ 40 dB HL through acoustic hearing, hearing aids, or cochlear implants). Socioeconomic marginalization was captured using the Ontario Marginalization Index (ON-MARG), which comprises four dimensions: Material Resources (economic disadvantage such as income and housing affordability), Households and Dwellings (residential stability and household structure), Age and Labour Force Structure (workforce engagement and age-related demographic vulnerability), and Racialized and Newcomer Populations (proportion of recent immigrants and racialized communities). Associations between principal components and developmental outcomes. (A) Variable loadings from principal component analysis (PCA) of hearing history, vestibular/balance function, and socioeconomic marginalization. Size of circles indicates strength of variable contribution to each principal component (PC). (B) Associations between PC scores and testing age. (C–E) Associations between each PC and standardized scores for language (C), working memory (D), and academic achievement (E). Asterisks indicate statistically significant associations (*P < 0.05, **P < 0.01, ***P < 0.001). Regression lines are shown only for significant effects. PC1 = Hearing Loss History, PC2 = Hearing Experience and Resources, PC3 = Social Marginalization (ON-MARG), PC4 = Vestibular and Balance Function. PC2 (hearing experience and resources) was most strongly associated with age (β = 2.10; 95% CI, 1.75 to 2.46; P < 0.001), whereas weak or no associations were found for other components (PC1: β = 0.34; 95% CI, 0.09 to 0.60; P = 0.009; PC3: β = 0.40; 95% CI, 0.05 to 0.74; P = 0.03; PC4: β = -0.25; 95% CI, -0.59 to 0.09; P = 0.15) (Fig. 3B). The strong link between age and auditory experience likely reflects early auditory access and increased time in sound with age in the present cohort of children with BCI. Only PC2 predicted language performance (β = 2.30; 95% CI, 0.86 to 3.75; P = 0.0019) (Fig. 3C). Working memory outcomes were predicted by PC2 (Digit Span verbal working memory: β = 1.27; 95% CI, 0.14 to 2.40; P = 0.029) (Fig. 3D). Academics (Fig. 3E) were also predicted by PC2 (Math: β = 4.48; 95% CI, 2.53 to 6.42; P < 0.001; Word reading: β = 7.49; 95% CI, 3.33 to 11.65; P < 0.001; Pseudoword reading: β = 4.56; 95% CI, 1.11 to 8.01; P = 0.011). Further correlational analyses examined whether distinct effects of resources captured by the ON-MARG indices of Household & Dwelling and Material Resources were related to outcomes; these analyses reveal weak but significant associations in the BCI group between specific developmental tests: (1) Material Resources to Math (r = 0.27, P = 0.03) and to Dot Matrix (r = 0.26, P = 0.02) scores and (2) Household & Dwelling scores to Math (r = 0.35, P = 0.005) and to Pseudoword reading (r = 0.23, P = 0.04). Weak to absent associations were found between developmental outcomes and PC1, PC3 or PC4 (PC1 to math: β = 1.57; 95% CI, 0.17 to 2.96; P = 0.029, PC3 to word reading (β = 4.41; 95% CI, 0.35 to 8.47; P = 0.034). Effects of etiology as a marker of broader comorbidities on cognitive and academic outcomes were assessed with linear regression models. Age distributions across etiologies are plotted in Fig. 4A. Figure 4B illustrates the effects of etiology across language, working memory, and academics, ordered from worst to best balance (consistent with Fig. 2B-D). Developmental outcomes by etiology*. (A) Age distribution across etiologies. Boxplots show raw scores for (B) Language, (C) Working Memory, and (D) Academic performance. Groups are ordered from worst to best balance performance on the BOT. Shown tests reflect those most impacted by etiology; less affected outcomes (Corsi Blocks, Digit Span, Pseudoword Reading) are excluded. Shaded blue region represents the typically developing group’s mean ± 1 SD. Asterisks indicate significant differences relative to typically developing controls after adjusting for age. *Genetic: GJB2 (15), LOXHD1 (3), MYO15A (3), MYO6A (1), DFN3 (1), CLRN1 (1). Regression analyses accounting for age revealed that deficits in language, visuospatial working memory, math, and reading in the BCI group (Fig. 1) varied by etiology. Children with cCMV and non-syndromic genetic hearing loss had deficits in all measures except reading (Fig. 4). Scores were lower by ~ 4–10 points in children with cCMV (language: β = −4.82; 95% CI, − 9.57 to − 0.07; P = 0.047; dot matrix: β = −4.38; 95% CI, − 8.56 to − 0.20; P = 0.040; math: β = −9.57; 95% CI, − 15.88 to − 3.26; P = 0.003), and by ~ 4–6 points in those with non-syndromic genetic hearing loss (language: β = −3.88; 95% CI, − 7.03 to − 0.73; P = 0.016; dot matrix: β = −3.73; 95% CI, − 6.47 to − 0.99; P = 0.008; math: β = −6.17; 95% CI, − 10.05 to − 2.29; P = 0.002). Children with IP2 cochlear anomalies showed deficits in language (β = −4.70; 95% CI, − 8.97 to − 0.43; P = 0.032) and dot matrix (β = −6.31; 95% CI, − 10.07 to − 2.55; P = 0.001). The small Waardenburg group (n = 3) had deficits in language (β = −7.50; 95% CI, − 14.29 to − 0.72; P = 0.031) and reading (β = −25.73; 95% CI, − 44.11 to − 7.35; P = 0.007). Children with unknown etiology showed deficits in math (β = −6.39; 95% CI, − 10.80 to − 1.97; P = 0.005) and reading (β = −10.94; 95% CI, − 20.66 to − 1.22; P = 0.028). No significant differences were observed in the Ushers group across any domain. Discussion This study evaluated developmental outcomes in children with BCIs, focusing on how hearing history including etiology, vestibular/balance function and socioeconomic marginalization, relate to performance in language, working memory, and academics. Despite age-related gains and early hearing access (Table 1), children with BCIs showed persistent deficits across these domains compared to typically developing peers. Children with BCIs show persistent developmental delays Children with BCIs demonstrated significantly poorer outcomes in language, visuospatial working memory, and academic skills (Fig. 1), aligning with reports that early access to sound through hearing devices does not always close developmental gaps relative to typically developing peers3,31. These delays remained stable across ages (Fig. 1A), supporting longitudinal studies which found that developmental gaps in children with hearing loss at 3 years of age persist over time and remain at 9 years of age8,43. Persistent variability in language and working-memory outcomes reflects access to sound through identification of hearing loss and device fitting as well as how well complex linguistic and cognitive skills are supported across development18,25,37,39,70,71. Even with early implantation, CI users commonly exhibit weaknesses in phonological processing, vocabulary, grammar, and verbal working memory that constrain reading and academic achievement, reflecting the downstream effects of limited auditory access on spoken language development8,25,38,72,73. Language exposure and mode of communication are reported to influence language, cognitive, and academic outcomes in children with hearing loss40,72,74,75. In the present study, all participants received intervention emphasizing listening and spoken English, and children who used manual communication as their primary mode of communication were excluded. Nevertheless, these children with cochlear implants experienced differences in the duration and cumulative language exposure and learning opportunities relative to their typically developing peers, as the period of hearing loss prior to identification and amplification limits early access to auditory language input. Further, although age at onset of AVT or SLP therapy did not significantly effect outcomes in the present study, variability in language exposure and intervention remains an important contextual factor to consider. Data on exposure to additional languages (spoken or manual) were not collected, limiting direct examination of communication modality or multilingualism. However, evidence increasingly suggests that neither multilingual exposure nor the inclusion of sign language inherently hinders developmental outcomes after CI when language access is early, consistent, and fully accessible29,40,41. Rather, outcomes of hearing70,74, spoken language29,30,75–78, academics (reading, writing)75,77,79, and memory71,78,80,81 appear more strongly associated with the timing, quality, and stability of language input. Vestibular and balance function are not strongly linked to cognitive abilities Poor balance was common in children with BCI and weakly associated with vestibular impairment (Fig. 2A). This finding was further supported by clustering of balance and vestibular results on PCA (PC4; Fig. 3A). Yet, PC4 was not linked to language, working memory, or academic outcomes, aligning with prior findings that vestibular loss primarily affects motor milestones rather than higher-order skills53,82. This challenges the idea of an additive cognitive burden from vestibular loss and suggests that predictors of developmental variability in children with CIs need to consider a broader set of factors that might be impeding their overall health and development. Auditory experience is essential for development in children with BCIs In the present cohort of children with BCI, auditory experience with age emerged as the strongest predictor of the measured developmental outcomes. Higher PC2 scores, reflecting greater time in sound and lower marginalization, were associated with better language, verbal working memory, and academic performance (Fig. 3C–E). By contrast, hearing loss history (PC1) and vestibular/balance function (PC4) showed minimal associations with the developmental outcomes measured. Social marginalization (PC3) also demonstrated limited influence, which may reflect healthcare systems, like that in Canada, that are able to foster early identification and intervention through universal newborn hearing screening, timely referral, and standardized pediatric CI pathways5,6,8,10,25,37,38,83–87. Recent work demonstrates that socioeconomic status accounts for more variance in spoken language communication mode and outcomes than home language or bilingual exposure, highlighting the importance of access to services, consistency of rehabilitation, and family resources47. Meta-analytic and longitudinal studies across the world confirm that age at implantation accounts for only part of the variability in language outcomes, with educational environments, therapy intensity, incidental language exposure, and broader cognitive-linguistic scaffolding contributing substantially to long-term development3,7,8,17,18,24,30,37–40. These findings suggest that some children may not receive sufficient opportunities to develop higher-order language and working-memory skills, even when early auditory access is achieved, potentially contributing to variability observed in CI outcomes. Variable outcomes were similarly shown in the present study and were primarily predicted by auditory experience. Although the clinical program from which participants were drawn provides integrated multidisciplinary care including auditory-verbal therapy, speech-language services, social work support, and school-based accommodations, individual differences in learning needs and developmental trajectories remain. Thus, even within well-resourced settings, early implantation alone is not sufficient to fully prevent later delays18. Etiology significantly contributes to developmental variability in children with BCIs Developmental outcomes varied by etiology of hearing loss (Fig. 4). Comparisons were possible due to increasing identification of childhood deafness etiologies59,88–90 and were known in all but 17.71% of the BCI group. Most children had non-syndromic genetic hearing loss (n = 23; 35%), commonly GJB2 mutations (15/23), consistent with prior reports91,92. Children with non-syndromic genetic etiologies showed impairments across all three domains, challenging suggestions of minimal developmental risk in these children93,94. Children with cCMV (11%) showed deficits across all domains, likely reflecting cCMV-related injury to white matter and cortical regions essential for cognition62,95–97. Balance was frequently affected, consistent with known vestibular impact63,64; however, present results (Fig. 3) question the additive impact on developmental outcomes. Cochlear anomalies (IP2; 14%) were associated with poorer language, potentially stemming from degraded auditory input98,99, and poorer visuospatial working memory than typically developing peers, perhaps reflecting cognitive co-morbidities100. Smaller Usher (n = 7) and Waardenburg (n = 3) groups matched expected prevalence90,91,101. Low language and reading performance occurred in 2 of the 3 children with Waardenburg, consistent with mixed outcomes in these children91,102,103. Outcomes in children with Ushers were variable but strong, possibly reflecting early identification and intervention due to known balance (Fig. 2) and vision loss risks104. These findings emphasize that hearing loss often includes developmental comorbidities. The type of insult (genetic, infectious, or structural) may affect neural systems necessary for multisensory and cognitive integration32,105–107 with implications for prognosis and intervention in children with hearing loss. Limitations and future directions This study focused on language, working memory, and academic skills but did not assess other cognitive domains such as executive function, attention, or processing speed. Clinical vestibular assessments were limited to vHIT and cVEMP, omitting oVEMP and additional semicircular canal testing, potentially underestimating vestibular contributions. Socioeconomic data was area-based, which may not accurately capture individual differences such as parental education or language exposure. Although all children with cochlear implants received therapy emphasizing listening and spoken language, detailed information on the duration, intensity, and frequency of intervention was not consistently available, as services were primarily delivered through community or school-based providers and not systematically documented in hospital records. In addition, broader measures of language exposure (e.g., spoken and/or manual languages at home, school, etc.) were not collected. Yet, socioeconomic status appears to be stronger determinant of spoken language skills than home language or bilingualism in children with cochlear implants47. COVID-19-related disruptions, including testing delays and masking, may have influenced performance, particularly in language tasks. Although testing sessions were comparable to typical school-day academic demands (1–2 h), fatigue or boredom were not formally measured and may have contributed to performance variability. Small sample sizes within etiologic subgroups reduced power to detect subtle effects. Future work should aim to capture a broader range of developmental influences by incorporating measures of reaction time, listening effort, socioeconomic, therapeutic, home-environment, and language exposure to better characterize environmental and lifestyle factors that shape research outcomes and are clinically informative for understanding functional outcomes and guiding follow up care. Future work could also build on recent dual-task paradigms used in adults with hearing and vestibular loss108 to examine cognitive–motor interference in children with bilateral cochlear implants. Dual-task designs, such as combining balance or gait tasks with simultaneous memory or language demands, might reveal limitations to attentional resources affecting concurrent dynamic tasks with implications for real-world functioning in classrooms and playgrounds, where cognitive and motor demands frequently overlap. Longitudinal and etiologically stratified studies could also provide added power. Conclusion Despite age-related gains, children with BCIs show persistent deficits in language, working memory, and academics. Auditory experience and underlying etiology associated with hearing loss were the primary predictors of developmental outcomes, whereas vestibular/balance function and available socioeconomic indices were not. Findings emphasize the importance of early auditory access along with etiology-informed, multidisciplinary care to optimize developmental outcomes in children with hearing loss. Acknowledgements We acknowledge the support of the SickKids Foundation Research Endowments and extend our sincere thanks to the children and families who participated in this study. We thank R.Z.A and H.B for their advice and input regarding R programming for data analyses, A.B for formatting assistance, and J.N for conducting a medical chart review. Author contributions M.H contributed to study conception and design, led data collection, conducted primary analyses and interpretation, and wrote the manuscript. S.L.C contributed to the study design, supported data interpretation, and provided critical manuscript revisions. K.A.G conceived the study, oversaw study design, guided analytical planning and interpretation, and contributed to drafting and revising the manuscript. Funding sources SickKids Foundation and the SickKids Clinician Scientist Training Program. Data availability Deidentified data supporting the findings of this study are available upon reasonable request. Access to the data is contingent upon approval from an appropriate ethical review board to safeguard the privacy and confidentiality of the research participants. Declarations Competing interests S.L.C and K.A.G have received honoraria as members of the Speakers’ Bureau for Cochlear Corp. K.A.G have received honoraria as members of the Speakers’ Bureau for Med-El, and Advanced Bionics. They have also acted as consultants for Eli Lilly, Akouos (Eli Lilly), and Decibel (Regeneron). S.L.C has received royalties from Plural Publishing for editing the Manual of Pediatric Balance Disorders and Nova Publishing for editing K.J. Lee’s Essential Pediatric Otolaryngology Volume 1 and 2. S.L.C also maintains a Sponsored Research Agreement with Cochlear Americas. K.A.G holds the Cochlear Americas Chair in Auditory Development and the Bastable-Potts Health Clinician Scientist in Hearing Impairment position and serves as a lecturer at Salus, Drexel University. M.H acts as a sub-investigator for Akouos (Eli Lilly). Ethical approval This study was conducted in accordance with ethical standards and approved by the independent Research Ethics Board at The Hospital for Sick Children (REB 1000079136). All participants provided informed consent or assent, with parental or guardian consent obtained for minors, as required by the ethical review process. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Melissa Hazen, Email: melissa.hazen@sickkids.ca. Karen A. Gordon, Email: karen.gordon@utoronto.ca References Associated Data Data Availability Statement Deidentified data supporting the findings of this study are available upon reasonable request. Access to the data is contingent upon approval from an appropriate ethical review board to safeguard the privacy and confidentiality of the research participants.