Judy R. Dubno

A human cochlear transcriptome is crucial for identifying cell types and marker genes, and for understanding auditory function and disorders like Ménière’s disease (MD), a chronic condition characterized by spontaneous episodic vertigo and fluctuating hearing loss. We generated a comprehensive single-nucleus RNA-seq (snRNA-seq) transcriptome from human cochleae and annotated 21 distinct cochlear cell types. Differential expression analyses combined with single-molecule fluorescence in situ hybridization and immunohistochemistry (smFISH) validated known markers and identified new markers that further refine our understanding of human cochlear cell types. Human-mouse comparison of novel cell type marker genes revealed strong species conservation supporting the use of mouse models in cochlear research. The relevance of these snRNA-seq data was established by integrating them with elevated genetic variant loads from a MD cohort-control analysis of 44,625 individuals from the All of Us Research Program. Our comprehensive human cochlear cellular transcriptome translationally implicates specific cochlear cell types in MD.

Word recognition scores are useful general metrics of speech understanding but provide limited insight about why listeners exhibit difficulties with word recognition tasks. For example, older adults can have the same word recognition score but make different types of errors. Different types of misperceptions for audible speech, such as misunderstanding the final consonant or a whole word, suggest that word recognition errors reflect distinct neural failures in speech processing. Patterns of activity in central auditory and attention-related cortical networks were hypothesized to predict different types of word recognition errors. Trial-level error analyses were performed with data from two neuroimaging experiments where older adults performed word recognition in noise tasks (Data Set 1: N = 45, mean age = 64.71 years; Data Set 2: N = 40, mean age = 68.97 years). Group-level independent component analysis of fMRI data collected during these tasks identified distinct patterns of activity in auditory and cingulo-opercular networks that were associated with initial or final consonant substitution errors. In addition, complete word omission errors occurred with relatively low activity in the auditory network and elevated activity in cingulo-opercular and/or default mode networks. These findings suggest that some errors occur with more cingulo-opercular-related performance monitoring, whereas other errors occur with insufficient signal in auditory cortex and/or task disengagement that has been associated with default-mode network function. Thus, patterns of cortical activity reflect the type of word recognition error and may inform reason(s) for those errors that may not be evident in an overall word recognition score.

This study examined associations between audiometric detection thresholds for older adults with normal and impaired hearing and recognition of degraded speech in speech-modulated noise. Three datasets consisted of 15-16 measures of temporally degraded speech: (1) degraded spectral cues, (2) competing speech-modulated noise, and (3) spectral degradation combined with speech-modulated noise. Speech was spectrally shaped to each listener's pure-tone detection thresholds, and ESTOI-based speech recognition thresholds (SRT_ESTOI) for degraded speech in noise were determined at 50% correct. Individual differences in detection thresholds from 0.25-8 kHz were examined, along with suprathreshold measures of speech modulation detection and glimpsing. Older adults with normal hearing (ONH, N = 20) or with sloping high-frequency hearing loss (OHI, N = 20) were compared to three younger normal hearing control groups (N = 63) to evaluate effects of age, spectral gain, and reduced speech sensation levels. Audiometric detection thresholds below 1 kHz were the strongest predictors of SRT_ESTOI for both groups of older adults. Among OHI listeners, associations of low frequency thresholds with SRT_ESTOI also differed depending on the type of speech distortion. Findings suggest the importance of low-frequency speech cues for glimpsing speech in temporally modulated backgrounds for older adults.

Purpose: Listening workload may partially explain variations in speech recognition that speech audibility alone cannot capture, as increased workload can interfere with speech processing. This study aimed to investigate the association between listening workload and speech recognition in a large sample of the general population with varying ages and hearing thresholds. Method: Data from 635 participants, from a longitudinal cohort study, who completed a modified version of the National Aeronautics and Space Administration Task Load Index (NASA-TLX) following the Northwestern University Auditory Test No. 6 (NU-6; n = 635), Speech Perception in Noise test (SPIN; n = 440), and Staggered Spondaic Word test (SSW; n = 455) were included. Adjusted for age, sex, and race, linear regressions models were used to examine associations between self-reported workload, defined by NASA-TLX domains of performance, effort, and frustration, with five outcomes of speech recognition, measured by the NU-6 (both adjusted and unadjusted for audibility), SPIN low-context (both adjusted and unadjusted for audibility), and SSW total percent scores. Results: In general, lower levels of self-reported workload were associated with better speech recognition scores, including for audibility-adjusted and non–audibility-adjusted speech recognition tasks and after adjustment for demographic factors. Conclusions: Findings suggest that self-reported performance, effort, and frustration ratings were associated with speech recognition in various listening conditions, with influences from reduced speech audibility and demographic factors including age, sex, and race. Thus, self-reported workload may serve as a complementary tool to provide better understanding of listening abilities of individuals with hearing loss. Supplemental Material: https://doi.org/10.23641/asha.32690985

Purpose: Construct validity of the Northwestern University Auditory Test No. 6 (NU-6) monosyllabic word lists or shortened versions have not been examined using the Rasch measurement theory. The study purposes were to test the fit of the Rasch measurement model to monosyllabic word lists and whether short forms and computerized adaptive testing can measure speech recognition. Method: A cross-sectional study design with 50 persons (average age = 71 years; 35 female, 15 male) with mild-to-moderate hearing loss was used to test the fit of NU-6 Word Lists 1 and 2 to the Rasch measurement model. Pearson's correlations quantified the accuracy of person measures from short forms or computer-adaptive testing simulations with measures from the full lists. Average standard error of person measures quantified measurement precision. Results: Word lists were unidimensional, had negligible misfit, and had high person reliability. Nineteen- and 11-word short forms were made per list. Person measures from 19-word short forms had a high linear association with person measures from the full word lists (List 1: r = .92, p < .0001, SE = 0.76; List 2: r = .91, p < .0001, SE = 0.74) compared to moderate association for the 11-word short form (List 1: r = .84, p < .0001, SE = 0.99; List 2: r = .81, p < .0001, SE = 0.97). Person measures from computerized adaptive testing simulation reached a correlation threshold, r > .90, after 15–20 words were administered for both lists. A precision-based stopping rule used an average of 18 (List 1) or 20 (List 2) words. Conclusion: Short forms with 19 words and computerized adaptive testing may accurately and precisely measure speech recognition.

Background: One-third of persons age 60 y+ have hearing loss, and hearing loss is a leading preventable risk factor for dementia. We estimated the number of age-associated dementia cases attributable to hearing loss in 2022. Methods: We used DeciBHAL, a validated microsimulation of hearing loss that includes age- and sex-specific annual probabilities of incident hearing loss (0.1%-10.4%) and dementia (0.3%-7.1%). Utility decrements are incorporated yearly, based on hearing loss (−0.13 to −0.31) and dementia severity (−0.04 to −0.42), to calculate quality-adjusted life-years (QALYs). We estimated dementia incidence for persons with and without hearing loss by removing the estimated proportion attributable to hearing loss (adjusted incidence risk ratio, 2.0 [range: 1.5-2.5]). We projected two cohorts: the general US population and a hypothetical US population without hearing loss (counterfactual). We applied model-projected dementia incidence and utility among both cohorts to the 74 190 000 US adults >60 y and without dementia in 2022. Results: Model-projected incident cases of dementia are 412 000/year (males) and 523 000/year (females). In the simulation without hearing loss, dementia cases/year fall to 339 000 for males and 455 000 for females projecting that 141 000 new dementia cases in 2022 would be attributable to hearing loss. In probabilistic sensitivity analysis, 95% of simulations projected the proportion of dementia cases attributable to hearing loss were 11.5%-23.6% for males and 6.7%-18.7% for females. Hearing loss and associated dementia reduced life-time QALYs by 1.38 for females and 1.69 for males. Conclusion: Model-projected estimates support that hearing loss prevention could substantially reduce new dementia cases and should be a priority.

Purpose: This purpose of this study was to determine the association of Type 2 diabetes with rate of hearing change per year. Method: Participants were from the Medical University of South Carolina Longitudinal Cohort Study of Age-related Hearing Loss. History of Type 2 diabetes was self-reported. Outcome measures were individual audiometric thresholds (0.25–8.0 kHz) and pure-tone average (PTA) of thresholds at frequencies 0.5–4.0 kHz, averaged bilaterally. We used linear mixed regression models to determine the association of diabetes with the rate of annual threshold change at each frequency and PTA, separately. Results are presented as regression coefficients with corresponding 95% confidence intervals (CIs). Results: This study included 1,333 participants with a mean age of 63.5 ( SD = 14.5) years and a mean of 4.7 ( SD = 5.5) years of follow-up; 57.7% were female, and 18.1% were of a racial Minority. The prevalence of Type 2 diabetes was 10.4%. Most (85.3%) participants with diabetes reported the use of diabetes medication(s). Participants with and without diabetes showed similar hearing thresholds at the baseline examination, with mean PTAs of 24.5 (95% CI [23.7, 25.2]) dB HL for participants without diabetes and 24.8 (95% CI [22.6, 26.9]) dB HL for participants with diabetes. Moreover, participants with and without diabetes showed similar rates of annual hearing change in models adjusted for age, sex, and race; PTA change was 0.90 (95% CI [0.86, 0.94]) dB for participants without diabetes and 0.87 (95% CI [0.73, 1.02]) dB for participants with diabetes. Conclusions: Diabetes was not consistently associated with baseline hearing thresholds or rate of change. Additional longitudinal studies are needed to clarify relationships of diabetes and hearing over time. Supplemental Material: https://doi.org/10.23641/asha.30826049

A recently described method (rM-NIHL) for diagnosing noise-induced hearing loss (NIHL) sustained during military service for men aged up to 60 years was developed based on the shapes of the audiograms of noise-exposed military personnel and a control group matched in age but without known noise exposure. Here, the applicability of the rM-NIHL method to men aged 61 to 80 years was evaluated, based on a database including 329 military noise-exposed men (ExpDB) and a database including 123 non-exposed men (ContDB). The method had high sensitivity but poor specificity. For further analysis, ExpDB was randomly split into ExpDB1 and ExpDB2, and ContDB was randomly split into ContDB1 and ContDB2. ExpDB1 and ContDB1 were used for the development of new methods and ExpDB2 and ContDB2 were used for evaluation. Modification of the criteria used in the rM-NIHL method led to a better balance between sensitivity and specificity, but the overall diagnostic performance was only moderate. A new method, denoted M-NIHL60+, was developed based on a weighted average of the age-corrected hearing thresholds across all frequencies from 1 to 8 kHz. This method gave a reasonable balance between sensitivity and specificity and gave higher diagnostic performance than the original rM-NIHL method or the rM-NIHL method with modified criteria. Comparison of data for older and younger exposed groups (assessed previously) supported the hypothesis that the rate of change of thresholds with age for noise-exposed individuals is greater for frequencies where the initial NIHL is small than for frequencies where it is large.