Abstract
Introduction:
Hearing loss is associated with accelerated cognitive decline, and auditory rehabilitation via cochlear implantation (CI) may mitigate this trajectory. In the past, the impact of cochlear implantation on different cognitive subdomains has been described. However, verbal fluency (VF), which requires fast semantic retrieval, executive control, and processing speed, and is predictive of dementia risk and overall survival, has been rarely studied and control groups are mostly missing due to ethical reasons. The present study compares long-term VF trajectories in CI recipients and untreated hearing-impaired controls from a large population-based aging study.
Materials and methods:
VF was assessed in 74 CI recipients (M = 65.6 years, SD = 9.1) at pre-operative baseline and 1, 2, 4.5, and up to 9 years post-implantation, and in 383 untreated hearing-impaired participants (M = 72.6 years, SD = 10.0) from the English Longitudinal Study of Ageing (ELSA) across a comparable time frame. Scores were z-standardized within each study to enable cross-cohort comparison. Linear mixed-effects models were used to compare VF trajectories, with age, sex, and education as covariates.
Results:
VF trajectories differed significantly between groups (Time × Study interaction: b = 0.562, p < 0.001). The ELSA cohort showed a steady linear decline over time (b = −0.261, p = 0.001), whereas the CI cohort exhibited an inverted-U trajectory with initial improvement followed by a plateau. After propensity score matching, results remained robust.
Conclusion:
Cochlear implantation is associated with more favorable long-term verbal fluency trajectories compared to untreated hearing loss. These findings add to the growing evidence that auditory rehabilitation may help preserve cognitive function in older adults.
Introduction
Hearing loss is one of the most prevalent chronic conditions in older adults and has been identified as the largest potentially modifiable risk factor for dementia (). The mechanism linking hearing loss to cognitive decline remains debated (; ). Leading hypotheses include cognitive load and effortful listening, social isolation and reduced stimulation, and shared neuropathological pathways (common cause) (; Uchida et al., 2019).
Cochlear implants are increasingly provided to older adults when conventional hearing aids no longer provide sufficient benefit (). In addition to increased speech understanding observed even in older adults, cochlear implant users also experience significant improvements in psychological well-being, health-related quality of life and social participation (; Tang et al., 2024; ; ; Shen et al., 2023; Wang et al., 2024). Recently, growing attention has been directed towards the positive effects of auditory rehabilitation through hearing aids or cochlear implants on various cognitive subdomains, mostly pronounced in attention, inhibition, and working memory (Völter et al., 2022; ; ; ; ; ; ; ; Takkoush et al., 2026; Vandenbroeke et al., 2025; Sarant et al., 2024).
However, evidence on cognitive outcomes post-implantation in the long term remains limited (Vandenbroeke et al., 2025; ; Völter et al., 2023; Takkoush et al., 2026; ) and methodologically heterogeneous, with most studies relying on rather small or uncontrolled pre-post designs (; ; ; Sarant et al., 2019). One major limitation of the existing literature is the lack of appropriate control groups, largely due to ethical constraints that preclude withholding treatment (; ; Völter et al., 2023). Comparing cochlear implant (CI) recipients to individuals with untreated hearing loss in population-based data can help disentangle the natural trajectory of cognitive aging with hearing loss from potential CI-related effects.
Verbal fluency (VF) is a particularly relevant outcome measure: it taps into executive function, semantic memory retrieval, and processing speed, all sensitive to age-related decline while being less dependent on peripheral hearing than other neuropsychological tests (; ). Deterioration in semantic fluency is recognized as a risk factor for dementia (). Recently, longitudinal data from the Berlin Aging Study (N = 516, M = 84.9 years) demonstrated that participants in the upper quartile of verbal fluency performance lived almost 9 years longer than those with lower scores (). VF is routinely collected in large population-based aging studies, such as the English Longitudinal Study of Ageing (ELSA), the Survey of Health, Ageing and Retirement in Europe (SHARE), and the U. S. Health and Retirement Study (HRS), with harmonized longitudinal ageing surveys that allow direct comparison between clinical CI cohorts and population norms (; ; ).
This study uses longitudinal data from a consecutive clinical CI cohort (n = 74) and participants with untreated hearing loss from the English Longitudinal Study of Ageing (ELSA; n = 383) to compare VF trajectories over time.
Study design and participants
Two independent cohorts of older adults with hearing loss were studied: (a) cochlear implant (CI) recipients who underwent surgical intervention and (b) participants of the English Longitudinal Study of Ageing (ELSA) with objectively measured, untreated hearing loss (i.e., no hearing aid and no cochlear implant).
CI cohort
The CI sample comprised 74 adults with postlingual bilateral severe-to-profound sensorineural hearing loss and a mean duration of pre-operative deafness of 22.67 (SD 13.42) years who received a cochlear implant at the Ear, Nose, and Throat Department of the Ruhr-University Bochum, Germany. Cognitive assessments were conducted at five time points: pre-implantation (T1) and at approximately 12 (T2), 24 (T3), 54 (T4), and 99 months (T5) post-implantation, spanning a mean follow-up of 8.25 years. Exclusion criteria were insufficient German language proficiency to complete neuropsychological testing and severe neurologic or psychologic impairments. Preoperative speech perception scores in quiet were 6.67% (SD 12.28) correct monosyllabic words at 65dB and 57.64% (SD 21.83) 1 year after implantation as well as 60.55% (SD 20.86) 2 years, 54.39% (SD 20.04) 4.5 years and 62.32% (SD 21.23) 8.25 years post CI.
ELSA cohort
The comparison sample was drawn from ELSA, an observational nationally representative panel study of community-dwelling adults aged 50 and older in England (Steptoe et al., 2013) including objective hearing measures. We selected participants from five consecutive waves: Wave 7 (fieldwork 2014–2015), Wave 8 (2016–2017), Wave 9 (2018–2019), Wave 10 (2021–2023, with fieldwork partly delayed by the COVID-19 pandemic), and Wave 11 (2023–2024) who met the following inclusion criteria at baseline: (a) objectively measured moderate-to-severe hearing loss, defined as hearing 0–2 of 6 tones on the HearCheck screener in the worse ear; (b) no current hearing aid use; and (c) no cochlear implant. Hearing loss was assessed using the Siemens HearCheck Screener, a handheld device that produces a fixed series of three mid-frequency tones (1 kHz) and three high-frequency tones (3 kHz) at decreasing intensities (). Following the classification used by , hearing acuity was determined based on the worse ear: good (6 tones), mild difficulty (3–5 tones), and moderate-to-severe difficulty (0–2 tones). After applying all inclusion criteria, the ELSA analytic sample comprised 383 adults (206 female [53.8%]; M = 72.6 years, SD = 10.0, range: 50–89). ELSA assessments were nominally biennial; however, the COVID-19 pandemic disrupted the planned fieldwork schedule for Waves 10 and 11 (). Because the publicly released ELSA data top-code age at 90 to protect confidentiality, exact ages are unavailable for participants aged 90 or older; we therefore restricted the ELSA sample to participants aged 50–89 at Wave 7.
Measures
Verbal fluency was assessed in the ELSA cohort using a semantic fluency paradigm in which participants named as many animals as possible within 60 s (Llewellyn and Matthews, 2009). The score was the total number of valid animals named (baseline M = 17.7, SD = 7.5). In the CI cohort, verbal fluency was assessed in a combined letter and category fluency task paradigm adapted from the Chicago Word Fluency Test (Thurstone, 1948) which was developed and validated for use in hearing-impaired older adults (Völter et al., 2017; ), where as many animals as possible starting with a particular letter must be named within 90 s (Völter et al., 2017). The category constraint (animals) engages temporal-semantic retrieval, while the letter constraint adds an executive-frontal filter, so that performance reflects the joint contribution of semantic search and executive selection. Performance was quantified as an inverse efficiency (IE) score, defined as the ratio of processing speed to accuracy (lower IE scores indicate better efficiency). For analysis, IE scores were inverted using vf = (max (IE) + 100) − IE so that higher values indicate better performance. Two re-analyses on raw word counts (on the 90-s window and on the first 60 s of each protocol to match the ELSA response window) were conducted. Baseline word-count statistics for both scoring windows are provided in Supplementary Table S1.
Possible practice effects were minimized by follow-up intervals for cognitive testing spanning at least 1 year and different test versions ().
Covariates
Because the two cohorts were drawn from different populations and differed in baseline demographic characteristics, three covariates were included in all models: age (centered at 70 years), sex (0 = male, 1 = female), and education level. Education was operationalized as a binary variable separating upper secondary education or higher from lower levels. In the CI cohort, participants with Abitur or higher were classified as high education (n = 11, 14.9%); in the ELSA cohort, participants with A-level or university degree were classified as high education (n = 109, 28.5%).
Outcome standardization
Because the two cohorts used different scoring metrics for verbal fluency, direct score comparisons were not possible. To enable cross-study comparison, scores were z-standardized within each study relative to baseline: for each participant, the raw score at each time point was centered on the respective study’s baseline mean and scaled by the baseline standard deviation. This yields scores in baseline SD units (z = 0 at the study-specific baseline mean), allowing trajectory comparisons in terms of relative change.
Statistical analysis
All analyses were conducted in R (R Core Team, 2025) using the lme4 and lmerTest packages (; ).
Time standardization
To compare trajectories across studies with different follow-up durations, time was rescaled to a 0–1 range, where 0 represents the first assessment and 1 represents the last. A slope of, for example, −0.50 would thus indicate a decline of half a baseline standard deviation from the first to the last assessment.
Linear mixed-effects models
Two types of growth models were fitted using restricted maximum likelihood (REML): (a) a linear model estimating a constant rate of change, and (b) a quadratic model adding a time2 term to capture curvilinear trajectories. All models included random intercepts and random slopes for time at the participant level. Cross-study models included a study indicator (0 = ELSA, 1 = CI) and its interaction with time terms to test whether trajectories differed between groups.
Sensitivity analysis
To assess the robustness of the primary findings under demographic matching, a propensity score full matching analysis was conducted.
Results
Sample characteristics
Table 1 presents the baseline demographic characteristics of both cohorts. The CI cohort was younger (M = 65.6 years vs. 72.6 years), included a higher proportion of women (59.5% vs. 53.8%), and had a lower proportion of participants with upper secondary education or higher (14.9% vs. 28.5%). These between-group differences were statistically controlled for in all models.
Table 1
| Variable | CI cohort | ELSA cohort | Test statistic | p |
|---|---|---|---|---|
| (n = 74) | (n = 383) | |||
| Age (years), M (SD) | 65.6 (9.1) | 72.6 (10.0) | t = −5.97 | <0.001 |
| Range | 50–84 | 50–89 | ||
| Sex (female), n (%) | 44 (59.5) | 206 (53.8) | χ2 (1) = 0.59 | 0.441 |
| Education (high)a, n (%) | 11 (14.9) | 109 (28.5) | χ2 (1) = 5.24 | 0.022 |
| Verbal fluency (raw), M (SD) | 214.7 (74.5)b | 17.7 (7.5)c | –d | – |
Baseline demographic and study characteristics.
CI, cochlear implant; ELSA, English Longitudinal Study of Ageing. Group comparisons: Welch’s t-test for continuous variables, χ2 test (Yates-corrected) for categorical variables.
CI: educational years >11 (German Abitur); ELSA: qual2 = 1 (A-level or higher).
Inverse efficiency score (transformed; higher = better performance).
Number of animals named in 60 s.
Not compared: different scoring metrics across cohorts (inverse efficiency vs. count).
Cross-study comparison of verbal fluency trajectories
Linear model
The linear mixed-effects model revealed a significant Time × Study interaction (b = 0.562, SE = 0.167, p < 0.001). The main effect of time indicated a significant decline in the ELSA reference group (b = −0.261, SE = 0.080, p = 0.001), while the CI cohort’s trajectory was 0.56 SD more favorable over the same standardized timeframe.
Quadratic model
The quadratic model revealed significant interactions for both the linear (Time × Study: b = 3.636, SE = 0.469, p < 0.001) and the quadratic (Time2 × Study: b = −3.483, SE = 0.501, p < 0.001) components. In the ELSA cohort, the quadratic time term was not significant (b = 0.044, SE = 0.232, p = 0.850). The significant cross-study quadratic interaction reflects the curvilinear trajectory in the CI cohort.
Within-study trajectories
For the CI cohort, the quadratic model provided a better fit (ΔAIC = 29.2 in favor of the quadratic model). The significant positive linear term (b = 3.288, SE = 0.557, p < 0.001) combined with the significant negative quadratic term (b = −3.374, SE = 0.591, p < 0.001) describes an inverted-U trajectory (Figure 1). For the ELSA cohort, the linear model was preferred (ΔAIC = −3.4, indicating no improvement from the quadratic term; quadratic term: b = 0.023, SE = 0.201, p = 0.911). The linear model revealed a significant decline of b = −0.292, SE = 0.076, p < 0.001, corresponding to approximately 0.29 baseline SD units over the standardized follow-up period. This within-study model selection mirrors the significant cross-cohort interaction.
Figure 1
Covariate effects
In the cross-study linear model, higher education was associated with better verbal fluency performance (b = 0.434, SE = 0.096, p < 0.001), and older age was associated with lower performance (b = −0.025 per year, SE = 0.004, p < 0.001). Sex was not significantly associated with verbal fluency in the cross-study model (b = −0.033, SE = 0.084, p = 0.693). Within-study analyses revealed opposing sex effects: in the CI cohort, women performed significantly better than men (b = 0.844, SE = 0.229, p < 0.001), while in the ELSA cohort, women performed significantly worse (b = −0.214, SE = 0.087, p = 0.014). Full model coefficients are presented in Tables 2, 3.
Table 2
| Parameter | Linear model | Quadratic model | ||||
|---|---|---|---|---|---|---|
| b | SE | p | b | SE | p | |
| Intercept | −0.040 | 0.074 | 0.584 | −0.035 | 0.076 | 0.639 |
| Time | −0.261 | 0.080 | 0.001 | −0.305 | 0.216 | 0.158 |
| Time2 | – | – | – | 0.044 | 0.232 | 0.850 |
| Study (CI) | 0.243 | 0.123 | 0.049 | −0.086 | 0.131 | 0.512 |
| Age (centered) | −0.025 | 0.004 | <0.001 | −0.026 | 0.004 | <0.001 |
| Sex (female) | −0.033 | 0.084 | 0.693 | −0.037 | 0.084 | 0.664 |
| Education (high) | 0.434 | 0.096 | <0.001 | 0.434 | 0.096 | <0.001 |
| Time × Study | 0.562 | 0.167 | <0.001 | 3.636 | 0.469 | <0.001 |
| Time2 × Study | – | – | – | −3.483 | 0.501 | <0.001 |
Cross-study mixed-effects models for verbal fluency (N = 457; 1,411 observations).
Time = standardized time (0–1). Study coded 0 = ELSA, 1 = CI. Age centered at 70 years. SE, standard error; b, unstandardized fixed-effect estimate; p, p-value from t-tests with Satterthwaite-approximated degrees of freedom. Statistical significance was evaluated at a two-tailed α of 0.05.
Table 3
| Parameter | CI cohort | ELSA cohort | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Linear model | Quadratic model | Linear model | Quadratic model | |||||||||
| b | SE | p | b | SE | p | b | SE | p | b | SE | p | |
| Intercept | −0.221 | 0.211 | 0.296 | −0.551 | 0.218 | 0.013 | 0.077 | 0.073 | 0.293 | 0.078 | 0.074 | 0.293 |
| Time | 0.290 | 0.198 | 0.145 | 3.288 | 0.557 | <0.001 | −0.292 | 0.076 | <0.001 | −0.311 | 0.186 | 0.095 |
| Time2 | – | – | – | −3.374 | 0.591 | <0.001 | – | – | – | 0.023 | 0.201 | 0.911 |
| Age (centered) | −0.002 | 0.012 | 0.881 | −0.003 | 0.013 | 0.798 | −0.030 | 0.004 | <0.001 | −0.030 | 0.004 | <0.001 |
| Sex (female) | 0.844 | 0.229 | <0.001 | 0.832 | 0.231 | <0.001 | −0.214 | 0.087 | 0.014 | −0.214 | 0.087 | 0.014 |
| Education (high) | 0.469 | 0.311 | 0.136 | 0.513 | 0.314 | 0.107 | 0.400 | 0.096 | <0.001 | 0.400 | 0.096 | <0.001 |
Within-study mixed-effects models for VF (CI cohort; n = 74; 295 observations; ELSA cohort; n = 383; 1,116 observations).
Time = standardized time (0–1). Age centered at 70 years. SE, standard error; b, unstandardized fixed-effect estimate; p, p-value from t-tests with Satterthwaite-approximated degrees of freedom. Statistical significance was evaluated at a two-tailed α of 0.05.
Attrition
Both cohorts showed attrition over the follow-up period, with lower retention from Wave 10 onward, coinciding with the COVID-19 pandemic. In the CI cohort, retention was high through 24 months (T3: 72 of 74; 97.3%) but declined to 49 (66.2%) at T4 and 28 (37.8%) at T5. In the ELSA cohort, retention decreased from 281 at Wave 8 (73.4%) to 100 at Wave 11 (26.1%). Mixed-effects models used all available observations (1,411 total across both cohorts), thereby reducing the impact of attrition. To assess potential selective attrition, baseline verbal fluency scores were compared between participants retained at the final assessment and those who dropped out earlier. In the ELSA cohort, retained participants had significantly higher baseline scores than dropouts (M = 21.6, SD = 6.0 vs. M = 16.3, SD = 7.5; t = 7.14, p < 0.001, d = 0.75), indicating selective attrition of lower-performing participants. In the CI cohort, no significant difference was found (t = 0.72, p = 0.472, d = 0.16). Per-wave sample sizes and word-count descriptives for both scoring windows are reported in Supplementary Table S2 (90-s window) and Supplementary Table S3 (60-s window).
Sensitivity analysis
Real-time axis
Findings remained stable when the cross-cohort quadratic model was re-fitted on a real-time axis (CI: 0, 1, 2, 4.5, 8.25 years; ELSA: 0, 2, 4, 7, 9 years): the linear (time_years × study_ci: b = 0.345, p < 0.001) and quadratic (time_years2 × study_ci: b = −0.043, p < 0.001) interaction terms remained significant (Supplementary Table S4).
Propensity score matching
To assess whether the primary findings were robust to direct demographic matching, a propensity score full matching analysis was conducted using the MatchIt package (). Propensity scores were estimand via logistic regression with age, sex, and education (binary) as covariates, using the average treatment effect on the treated (ATT) as the estimand. Full matching yielded 73 subclasses retaining all 74 CI and 383 ELSA participants; all post-matching standardized mean differences were below 0.10 (largest: age, SMD = 0.072). The effective sample size (ESS) on the ELSA side was 131.6. Results were consistent with the primary analysis: the linear Time × Study interaction was b = 0.557, SE = 0.191, p = 0.004, and the quadratic interactions remained highly significant (Time × Study: b = 3.438, SE = 0.459, p < 0.001; Time2 × Study: b = −3.344, SE = 0.489, p < 0.001). The direction and significance of all effects were consistent across both approaches, supporting the robustness of the findings.
Discussion
Our study compared verbal fluency trajectories in CI recipients with a control group of individuals with untreated hearing loss over a follow-up period of up to 9 years. The central finding is that VF trajectories differed significantly between the two groups: whereas the ELSA cohort with untreated hearing loss showed a steady linear decline, the CI cohort exhibited an inverted-U trajectory, with initial improvement followed by a gradual return toward baseline levels. Over the full standardized follow-up, the CI cohort’s trajectory was 0.56 SD more favorable than that of the untreated hearing loss group. This is important as VF is significantly linked to longevity, even if the explanation is not clear ().
Verbal fluency (VF) is a particularly complex cognitive ability that requires an interplay of different cognitive abilities (Schmidt et al., 2017). Fluency tasks have been shown to be especially sensitive to prefrontal and frontal-subcortical deficits, to mild cognitive impairment and to dementia diagnosis (). The processing strategies for semantic and phonemic fluency tasks are different, whereas semantic fluency correlates with temporal, phonemic fluency is linked to frontal lobe regions and a decline of VF in aging seems to be more pronounced in semantic than in phonemic fluency ().
The present study is, to our knowledge, the first to compare VF trajectories in CI recipients with a control group of individuals with untreated hearing loss over a follow-up of up to 9 years. It adds further data to the growing body of literature suggesting a positive impact of CI on the cognitive functioning of individuals with hearing loss.
So far, VF has been rarely studied in CI cohorts and with mixed results (; ; ; ; Völter et al., 2022). Huber et al. reported a significant improvement in semantic fluency 12 months after implantation in 21 older CI recipients (aged 60–75 years), with an average gain of 3.71 words (range: 1.65–5.78), but no improvement was found in letter fluency or in a younger subgroup (n = 20, aged 25–58 years) (). In contrast, Baranger et al. reported an improvement only in phonemic fluency 3 months after CI in a mixed cohort of 43 individuals with congenital and acquired hearing loss (). The present study extends these findings by providing long-term data beyond the typical 12-month follow-up and by including an external comparison group, addressing a key methodological limitation of prior work (Takkoush et al., 2026; ; ; Völter et al., 2022).
A plausible pathway linking CI to improved VF is by an increase in working memory through diminished cognitive load as described in detail by the Ease of Language Understanding Model, published by Rönnberg et al. (2013). In the past, improvements in different cognitive subdomains after CI have been observed by various research groups, mostly in attentional processes and working memory (Völter et al., 2022; ; ; ; ; ; ; ; Takkoush et al., 2026; Vandenbroeke et al., 2025; Sarant et al., 2024). However, Huber et al., who recently studied verbal fluency in younger and older hearing impaired 12 months after cochlear implantation showed that improvement in semantic fluency correlated with enhanced working memory after CI, but that working memory did not fully mediate the effect of CI on semantic fluency ().
A complementary mechanism is through increased social participation (), given that effective communication is central to social activities () and that reduced social engagement is itself a dementia risk factor (Sommerlad et al., 2023).
Because social participation was not measured in either cohort, this mechanism remains a hypothesis to be tested in future studies with direct social-engagement assessment (; ; Sommerlad et al., 2023).
The opposing sex effects across cohorts call for caution. Possible explanations include differential referral patterns into the CI cohort, cross-national differences in verbal-fluency performance, and task-format effects (letter-constrained vs. unconstrained fluency). The present data cannot distinguish among these accounts; we regard this finding as exploratory.
Limitations
This study has four limitations. First, the two cohorts used different verbal fluency tasks (CI: letter-constrained category fluency, 90 s; ELSA: unconstrained semantic fluency, 60 s) and different hearing assessments (audiometric testing vs. HearCheck screener; ). Cross-cohort inference therefore rests on the shape of the longitudinal trajectories rather than on absolute score levels. Letter and category fluency share executive-frontal and semantic-temporal substrates and correlate in older adults (Shao et al., 2014; ), and a 60-s reanalysis of the CI protocol reproduces the cross-cohort divergence in trajectory shape.
Second, residual confounding cannot be excluded. While age, sex, and education were measured comparably and controlled for in all models, other variables linked to cognitive trajectories (e.g., comorbidity burden, depressive symptoms, physical activity, social engagement) could not be harmonized across cohorts. Within-cohort z-standardization likewise cannot fully eliminate cross-national differences between the German CI cohort and the English ELSA cohort (educational systems, healthcare access, language-specific verbal fluency norms). We therefore read our findings as describing an observed association rather than a controlled CI effect estimate.
Third, both cohorts showed attrition. In the CI cohort, the SD of word counts narrowed from 3.18 at T2 to 1.70 at T5, indicating selective retention of higher-performing participants. Such attrition can attenuate late-wave estimates but cannot generate the early rise from T1 to T3, where attrition was minimal.
Finally, because this study is observational, we cannot draw causal inferences; factors associated with the decision to seek implantation may have shaped the CI cohort’s trajectory.
To summarize
Considering that hearing loss is associated with a faster cognitive decline, the observation that VF improves after implantation is promising. As it is unethical to deny a cochlear implantation to a severely hearing-impaired person for many years, a fictive control group such as the ELSA cohort might be a good option adding another view to the ongoing discussion on the impact of cochlear implantation on cognition in the long-term. However, there are limitations, and the present data cannot resolve the question whether the cognitive booster effect after implantation might help to prevent or at least to delay dementia. Multicenter studies with larger sample sizes and a long-term follow-up are mandatory to underline this important finding in the context of healthy aging.
Statements
Data availability statement
The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author.
Ethics statement
The studies involving humans were approved by the Ethics Commission Westfalen-Lippe (No. 2025-845-f-S). The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.
Author contributions
CV: Conceptualization, Project administration, Supervision, Writing – original draft, Writing – review & editing. LB: Data curation, Writing – original draft, Writing – review & editing. SD: Resources, Supervision, Writing – review & editing. STK: Conceptualization, Formal analysis, Methodology, Software, Validation, Visualization, Writing – original draft, Writing – review & editing.
Funding
The author(s) declared that financial support was received for this work and/or its publication.We appreciate the support by the DFG Open Access Publication Funds of the Ruhr-University Bochum.
Acknowledgments
We are thankful to Michael Falkenstein, ALA Institute, Bochum, Germany, for providing the ALAcog test instrument and to Ludger Blanke, ALA Institute, for technical support. Furthermore, we thank all the patients, the staff of the cochlear implant center Ruhrgebiet that participated in the present study.
Conflict of interest
CV and SD have received reimbursement of scientific meeting participation fees and funding for research projects that they initiated, from MED-EL©.
The remaining author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that Generative AI was used in the creation of this manuscript. During the preparation of this manuscript, the authors used Claude (Claude Opus 4.8; Anthropic) for language editing and to assist with writing and reviewing analysis code. The authors reviewed all AI-assisted output and take full responsibility for the content and integrity of this work.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
Supplementary material
The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fnins.2026.1854922/full#supplementary-material
References
1
AminiA. E.NaplesJ. G.HwaT.LarrowD. C.CampbellF. M.QiuM.et al. (2023). Emerging relations among cognitive constructs and cochlear implant outcomes: a systematic review and meta-analysis. Otolaryngol. Head Neck Surg.169, 792–810. doi: 10.1002/ohn.344,
2
AmuntsJ.CamilleriJ. A.EickhoffS. B.PatilK. R.HeimS.von PolierG. G.et al. (2021). Comprehensive verbal fluency features predict executive function performance. Sci. Rep.11:6929. doi: 10.1038/s41598-021-85981-1,
3
AndriesE.BosmansJ.EngelborghsS.CrasP.VandervekenO. M.LammersM. J. W.et al. (2023). Evaluation of cognitive functioning before and after cochlear implantation in adults aged 55 years and older at risk for mild cognitive impairment. JAMA Otolaryngol. Head Neck Surg.149, 310–316. doi: 10.1001/jamaoto.2022.5046,
4
BarangerM.ManeraV.SérignacC.DerreumauxA.CancianE.VandersteenC.et al. (2023). Evaluation of the cognitive function of adults with severe hearing loss pre- and post-cochlear implantation using verbal fluency testing. J. Clin. Med.12:3792. doi: 10.3390/jcm12113792,
5
BassoM. R.BornsteinR. A.LangJ. M. (1999). Practice effects on commonly used measures of executive function across twelve months. Clin. Neuropsychol.13, 283–292. doi: 10.1076/clin.13.3.283.1743
6
BatesD.MächlerM.BolkerB.WalkerS. (2015). Fitting linear mixed-effects models using lme4. J. Stat. Softw.67, 1–48. doi: 10.18637/jss.v067.i01
7
BoisvertI.ReisM.AuA.CowanR.DowellR. C. (2020). Cochlear implantation outcomes in adults: a scoping review. PLoS One15:e0232421. doi: 10.1371/journal.pone.0232421,
8
CalvinoM.Sánchez-CuadradoI.GavilánJ.Gutiérrez-RevillaM. A.PoloR.LassalettaL. (2022). Effect of cochlear implantation on cognitive decline and quality of life in younger and older adults with severe-to-profound hearing loss. Eur. Arch. Otorrinolaringol.279, 4745–4759. doi: 10.1007/s00405-022-07253-6,
9
CosettiM. K.PinkstonJ. B.FloresJ. M.CosettiM.PinkstonJ.FloresJ.et al. (2016). Neurocognitive testing and cochlear implantation: insights into performance in older adults. Clin. Interv. Aging11, 603–613. doi: 10.2147/cia.s100255,
10
CudaD.ManriqueM.RamosÁ.MarxM.BovoR.KhnifesR.et al. (2024). Improving quality of life in the elderly: hearing loss treatment with cochlear implants. BMC Geriatr.24:16. doi: 10.1186/s12877-023-04642-2,
11
DavisA.SmithP.FergusonM.StephensD.GianopoulosI. (2007). Acceptability, benefit and costs of early screening for hearing disability: a study of potential screening tests and models. Health Technol. Assess.11, 1–294. doi: 10.3310/hta11420,
12
DawesP. (2019). Hearing interventions to prevent dementia. HNO67, 165–171. doi: 10.1007/s00106-019-0617-7,
13
DawesP.VölterC. (2023). Do hearing loss interventions prevent dementia?Z. Gerontol. Geriatr.56, 261–268. doi: 10.1007/s00391-023-02178-z,
14
DorchiesF.MuchembledC.AdamkiewiczC.GodefroyO.RousselM. (2024). Investigating the cognitive architecture of verbal fluency: evidence from an interference design on 487 controls. Front. Psychol.15:1441023. doi: 10.3389/fpsyg.2024.1441023,
15
García-HerranzS.Díaz-MardomingoM. C.VeneroC.PeraitaH. (2020). Accuracy of verbal fluency tests in the discrimination of mild cognitive impairment and probable Alzheimer's disease in older Spanish monolingual individuals. Neuropsychol. Dev. Cogn. B Aging Neuropsychol. Cogn.27, 826–840. doi: 10.1080/13825585.2019.1698710,
16
GhislettaP.AicheleS.GerstorfD.CarolloA.LindenbergerU. (2025). Verbal fluency selectively predicts survival in old and very old age. Psychol. Sci.36, 87–101. doi: 10.1177/09567976241311923,
17
GkotzamanisV.KoliopanosG.Sanchez-NiuboA.OlayaB.CaballeroF. F.Ayuso-MateosJ. L.et al. (2023). Determinants of verbal fluency trajectories among older adults from the English longitudinal study of aging. Appl. Neuropsychol. Adult30, 110–119. doi: 10.1080/23279095.2021.1913739,
18
GordonJ. K.YoungM.GarciaC. (2018). Why do older adults have difficulty with semantic fluency?Neuropsychol. Dev. Cogn. B Aging Neuropsychol. Cogn.25, 803–828. doi: 10.1080/13825585.2017.1374328
19
GötzeL.SheikhF.HaubitzI.FalkensteinM.TimmesfeldN.VölterC. (2024). Evaluation of a non-auditory neurocognitive test battery in hearing-impaired according to age. Eur. Arch. Otorrinolaringol., 281, 2941–2949. doi: 10.1007/s00405-023-08408-9
20
GurgelR. K.DuffK.FosterN. L.UranoK. A.deTorresA. (2022). Evaluating the impact of cochlear implantation on cognitive function in older adults. Laryngoscope132, S1–S15. doi: 10.1002/lary.29933,
21
GustavsonD. E.ElmanJ. A.PanizzonM. S.FranzC. E.ZuberJ.Sanderson-CiminoM.et al. (2020). Association of baseline semantic fluency and progression to mild cognitive impairment in middle-aged men. Neurology95, e973–e983. doi: 10.1212/wnl.0000000000010130,
22
HenryJ. D.CrawfordJ. R. (2004). A meta-analytic review of verbal fluency performance following focal cortical lesions. Neuropsychology18, 284–295. doi: 10.1037/0894-4105.18.2.284,
23
HoD.ImaiK.KingG.HoD. E.StuartE. A. (2011). MatchIt: nonparametric preprocessing for parametric causal inference. J. Stat. Softw.42, 1–28. doi: 10.18637/jss.v042.i08
24
HuberM.IllgA.ReuterL.WeitgasserL. (2025). After cochlear implantation in older adults, enhanced working memory does not fully mediate the relationship between CI and improved semantic fluency. Front. Aging Neurosci.17:1701934. doi: 10.3389/fnagi.2025.1701934,
25
HuberM.RoeschS.PletzerB.LukaschykJ.Lesinski-SchiedatA.IllgA. (2021). Can cochlear implantation in older adults reverse cognitive decline due to hearing loss?Ear Hear.42, 1560–1576. doi: 10.1097/aud.0000000000001049,
26
IssingC.BaumannU.PantelJ.StöverT. (2021). Impact of hearing rehabilitation using cochlear implants on cognitive function in older patients. Otol. Neurotol.42, 1136–1141. doi: 10.1097/mao.0000000000003153,
27
IssingC.LothA. G.SakmenK. D.GuchlernerL.HelbigS.BaumannU.et al. (2024). Cochlear implant therapy improves the quality of life and social participation in the elderly: a prospective long-term evaluation. Eur. Arch. Otorrinolaringol.281, 3453–3460. doi: 10.1007/s00405-023-08443-6,
28
JohnsonB. R.DillonM. T.ThompsonN. J.RichterM. E.OvertonA. B.RoothM. A.et al. (2025). Benefits of cochlear implantation for older adults with asymmetric hearing loss. Laryngoscope135, 352–360. doi: 10.1002/lary.31718
29
KnopkeS.SchubertA.HäusslerS. M.GräbelS.SzczepekA. J.OlzeH. (2021). Improvement of working memory and processing speed in patients over 70 with bilateral hearing impairment following unilateral cochlear implantation. J. Clin. Med.10:3421. doi: 10.3390/jcm10153421,
30
KuznetsovaA.BrockhoffP. B.ChristensenR. H. B. (2017). lmerTest package: tests in linear mixed effects models. J. Stat. Softw.82, 1–26. doi: 10.18637/jss.v082.i13
31
LinF. R. (2020). Hear and now. Cerebrum2020:cer-07.
32
LivingstonG.HuntleyJ.LiuK. Y.CostafredaS. G.SelbækG.AlladiS.et al. (2024). Dementia prevention, intervention, and care: 2024 report of the lancet standing commission. Lancet404, 572–628. doi: 10.1016/s0140-6736(24)01296-0
33
LlewellynD. J.MatthewsF. E.. (2009). Increasing levels of semantic verbal fluency in elderly English adults. Aging, Neuropsychology, and Cognition, 16, 433–445.
34
LloydL.TaylorK.TsantaniM.EnglefieldL.WilliamsR.AllenJ.. The Dynamics of The Dynamics of Ageing: The 2023/2024 English Longitudinal Study of Ageing (Wave 11) (2023).
35
MertenN.DawesP.MunroK. J.BrenowitzW. D. (2026). Hearing impairment and cognitive decline: alternative explanations to causality. J. Alzheimer's Dis110, S25–S29. doi: 10.1177/13872877251410502,
36
MertensG.AndriesE.ClaesA. J.TopsakalV.Van de HeyningP.Van RompaeyV.et al. (2021). Cognitive improvement after cochlear implantation in older adults with severe or profound hearing impairment: a prospective, longitudinal, controlled, multicenter study. Ear Hear.42, 606–614. doi: 10.1097/aud.0000000000000962,
37
MosnierI.BelminJ.CudaD.Manrique HuarteR.MarxM.Ramos MaciasA.et al. (2024). Cognitive processing speed improvement after cochlear implantation. Front. Aging Neurosci.16:1444330. doi: 10.3389/fnagi.2024.1444330,
38
MosnierI.VanierA.BonnardD.Lina‐GranadeG.TruyE.BordureP.et al. (2018). Long-term cognitive prognosis of profoundly deaf older adults after hearing rehabilitation using cochlear implants. J. Am. Geriatr. Soc.66, 1553–1561. doi: 10.1111/jgs.15445,
39
NicholsE.GrossA. L.KobayashiL. C.LangaK. M.LeeJ. (2026). The measurement of cognition in large-scale cross-national surveys: lessons from the health and retirement study international network of studies and the harmonized cognitive assessment protocol. J. Gerontol. B Psychol. Sci. Soc. Sci.81, S18–S30. doi: 10.1093/geronb/gbaf200,
40
OhtaY.ImaiT.MaekawaY.MorihanaT.OsakiY.SatoT.et al. (2022). The effect of cochlear implants on cognitive function in older adults: a prospective, longitudinal 2-year follow-up study. Auris Nasus Larynx49, 360–367. doi: 10.1016/j.anl.2021.09.006,
41
Pichora-FullerM. K.MickP.ReedM. (2015). Hearing, cognition, and healthy aging: social and public health implications of the links between age-related declines in hearing and cognition. Semin. Hear.36, 122–139. doi: 10.1055/s-0035-1555116,
42
PowellD.OhE. S.ReedN.PowellD. S.ReedN. S.LinF. R.et al. (2021). Hearing loss and cognition: what we know and where we need to go. Front. Aging Neurosci.13:769405. doi: 10.3389/fnagi.2021.769405,
43
R Core Team. (2025). A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/,
44
RayJ.PopliG.FellG. (2018). Association of cognition and age-related hearing impairment in the English longitudinal study of ageing. JAMA Otolaryngol. Head Neck Surg.144, 876–882. doi: 10.1001/jamaoto.2018.1656,
45
RehanS.PhillipsN. A. (2025). Psychosocial function in mild cognitive impairment: social participation is associated with cognitive performance in multiple domains. J. Appl. Gerontol.44, 1629–1640. doi: 10.1177/07334648241311661,
46
RikosN.LinardakisM.SmpokosE.SpiridakiE.SymvoulakisE. K.TsiligianniI.et al. (2024). Assessment of cognitive function in European adults aged 50+in relation to their handgrip strength and physical inactivity: the SHARE study during 2019-2020. J. Res. Health Sci.24:e00611. doi: 10.34172/jrhs.2024.146,
47
RönnbergJ.LunnerT.ZekveldA.SörqvistP.DanielssonH.LyxellB.et al. (2013). The ease of language understanding (ELU) model: theoretical, empirical, and clinical advances. Front. Syst. Neurosci.7:31. doi: 10.3389/fnsys.2013.00031,
48
SarantJ. Z.BusbyP. A.SchembriA. J.BriggsR. J. S.MastersC. L.HarrisD. C. (2024). COCHLEA: longitudinal cognitive performance of older adults with hearing loss and cochlear implants at 4.5-year follow-up. Brain Sci.14:1279. doi: 10.3390/brainsci14121279,
49
SarantJ.HarrisD.BusbyP.MaruffP.SchembriA.DowellR.et al. (2019). The effect of cochlear implants on cognitive function in older adults: initial baseline and 18-month follow up results for a prospective international longitudinal study. Front. Neurosci.13:789. doi: 10.3389/fnins.2019.00789,
50
SchmidtC. S. M.SchumacherL. V.RömerP.LeonhartR.BeumeL.MartinM.et al. (2017). Are semantic and phonological fluency based on the same or distinct sets of cognitive processes? Insights from factor analyses in healthy adults and stroke patients. Neuropsychologia99, 148–155. doi: 10.1016/j.neuropsychologia.2017.02.019,
51
ShaoZ.JanseE.VisserK.MeyerA. S. (2014). What do verbal fluency tasks measure? Predictors of verbal fluency performance in older adults. Front. Psychol.5:772. doi: 10.3389/fpsyg.2014.00772,
52
ShenS.SayyidZ.AndresenN.CarverC.DunhamR.MarsigliaD.et al. (2023). Longitudinal auditory benefit for elderly patients after cochlear implant for bilateral hearing loss, including those meeting expanded Centers for Medicare & Medicaid Services criteria. Otol. Neurotol.44, 866–872. doi: 10.1097/MAO.0000000000003983,
53
SommerladA.KivimäkiM.LarsonE. B.RöhrS.ShiraiK.Singh-ManouxA.et al. (2023). Social participation and risk of developing dementia. Nat. Aging3, 532–545. doi: 10.1038/s43587-023-00387-0,
54
SteptoeA.BreezeE.BanksJ.NazrooJ. (2013). Cohort profile: the English longitudinal study of ageing. Int. J. Epidemiol.42, 1640–1648. doi: 10.1093/ije/dys168,
55
TakkoushS.DuffK.FosterN. L.PatelNSGurgelRK. (2026). Evaluating the impact of Cochlear implantation on cognitive outcomes in older adults: a 5-year follow-up. Otol. Neurotol.47, e283-e289. doi: 10.1097/MAO.0000000000004770
56
TangD.TranY.LoC.LeeJ. N.TurnerJ.McAlpineD.et al. (2024). The benefits of cochlear implantation for adults: a systematic umbrella review. Ear Hear.45, 801–807. doi: 10.1097/aud.0000000000001473,
57
ThurstoneL. L. (1948). Primary mental abilities. Science108:585.
58
UchidaY.SugiuraS.NishitaY.SajiN.SoneM.UedaH. (2019). Age-related hearing loss and cognitive decline - the potential mechanisms linking the two. Auris Nasus Larynx46, 1–9. doi: 10.1016/j.anl.2018.08.010,
59
VandenbroekeT.AndriesE.LammersM. J.Van de HeyningP.Hofkens-Van den BrandtA.VandervekenO.et al. (2025). Cognitive changes up to 4 years after cochlear implantation in older adults: a prospective longitudinal study using the RBANS-H. Ear Hear.46, 361–370. doi: 10.1097/aud.0000000000001583,
60
VölterC.GötzeL.BajewskiM.DazertS.ThomasJ. P. (2022). Cognition and cognitive reserve in cochlear implant recipients. Front. Aging Neurosci.14:838214. doi: 10.3389/fnagi.2022.838214,
61
VölterC.GötzeL.DazertS.ThomasJ. P.KaminS. T. (2023). Longitudinal trajectories of memory among middle-aged and older people with hearing loss: the influence of cochlear implant use on cognitive functioning. Front. Aging Neurosci.15:1220184. doi: 10.3389/fnagi.2023.1220184,
62
VölterC.GötzeL.FalkensteinM.DazertS.ThomasJ. P. (2017). Application of a computer-based neurocognitive assessment battery in the elderly with and without hearing loss. Clin. Interv. Aging12, 1681–1690. doi: 10.2147/cia.s142541,
63
WangQ.KapolowiczM. R.LiJ.-N.JiF.ShenW. D.WangF. Y.et al. (2024). A prospective cohort study of cochlear implantation as a treatment for tinnitus in post-lingually deafened individuals. Commun. Med. (Lond.)4:274. doi: 10.1038/s43856-024-00692-8,
Summary
Keywords
cochlear implantation, dementia, ELSA, hearing loss, verbal fluency
Citation
Völter C, Bode L, Dazert S and Kamin ST (2026) Verbal fluency after cochlear implantation: a longitudinal comparison with untreated hearing loss in the ELSA cohort. Front. Neurosci. 20:1854922. doi: 10.3389/fnins.2026.1854922
Received
13 April 2026
Revised
29 May 2026
Accepted
08 June 2026
Published
15 July 2026
Volume
20 - 2026
Edited by
Terrin N. Tamati, The Ohio State University, United States
Reviewed by
Keith N. Darrow, Worcester State University, United States
Hema Nagaraj, All India Institute of Speech and Hearing (AIISH), India
Kristina Bowdrie, Case Western Reserve University, United States
Updates
Copyright
© 2026 Völter, Bode, Dazert, and Kamin.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Christiane Völter, christiane.voelter@rub.de
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.