Hearing Rehabilitative Treatment for Older Adults with Comorbid Hearing Loss and Depression: Effects on Depressive Symptoms and Executive Function

Abstract

Objectives:

Recent research has revealed important neural and psychiatric consequences of HL in older adults. This pilot study examined the neural effects of HL and the impact of hearing aids on neuropsychiatric outcomes in major depressive disorder (MDD).

Design:

12-week, double-blind, randomized controlled trial.

Participants/Intervention:

N=25 (≥60 years) with MDD and moderate-profound HL were randomized to receive hearing aids (100% gain targets) or sham hearing aids (flat 30dB HL) in addition to psychiatric treatment-as-usual.

Measurements:

Depressive symptoms (Hamilton Rating Scale for Depression [HRSD]), executive functioning (NIH Toolbox Flanker), integrity of auditory brain areas (structural MRI, diffusion tensor imaging).

Results:

At baseline, worse speech discrimination was associated with auditory cortical thinning (Left anterior transverse temporal gyrus: r=0.755, p=0.012) and lower integrity of the superior longitudinal fasciculus (FA: Left r=0.772, p=0.025, Right r=0.782, p=0.022). After 12-weeks, hearing aids were effective at improving hearing functioning (Hearing Handicap for the Elderly: active −12.47 vs. sham −4.19, t=−2.64, df=18, p=0.016) and immediate memory (active +14.9 vs. sham +5.7, t=2.28, df=16, p=0.037). Moderate improvement was observed for hearing aids on executive functioning but did not reach statistical significance (Flanker: active +4.8 vs. sham − 2.4, t=1.95, df=15, p=0.071). No significant effect on depression was found (HRSD: active −5.50 vs. sham −7.32, t=0.75, df=19, p=0.46).

Conclusions:

HL can affect brain regions important for auditory and cognitive processing, and hearing remediation may have beneficial effects on executive functioning in MDD. Future studies may evaluate whether impairment in cognitive control consequent to HL may be an important risk mechanism for MDD.

OBJECTIVE

While historically considered a benign effect of aging, hearing loss (HL) has recently been linked to neuropsychiatric dysfunction including cognitive decline and depression in older adults. Audiometrically-defined HL has been associated with incident depressive symptoms over longitudinal follow-up in multiple samples (1, 2), among individuals of different ethnicities (3), and even in older adults with subclinical HL (i.e., at a level still considered “normal”) (4). This work demonstrates that any deviation from excellent hearing may increase risk for depressive symptoms and major depressive disorder (MDD) in a “dose-dependent” fashion, such that individuals with at least moderate HL (versus normal hearing) are associated with three times greater rates of clinically significant depressive symptoms. MDD occurring in older adults is associated with high rates of completed suicide (5) and confers risk for adverse medical outcomes, disability, and cognitive decline (6). Despite the important consequences of HL on neuropsychiatric health, it is unfortunate that utilization of hearing aids in older adults is strikingly low (7).

Age-related HL can have adverse neural consequences that may explain its relationship to incident MDD (8). Reduced transmission of auditory stimuli due to peripheral HL decreases input to the primary and secondary auditory cortices (9, 10), resulting in deafferentation-induced atrophy in frontal and temporal brain areas (9-11), including nodes of the cognitive control network (CCN) supporting the executive functions (e.g., anterior cingulate cortex [ACC] and insula) (12). Consistent with this model, observational studies have found associations between HL and impairment in executive functioning that improve with hearing aid use (13). Executive dysfunction is common in older adults with MDD (14), and is a predictor of poor treatment response, early relapse, and recurrence (15). Thus, CCN impairment consequent to HL may be an important risk mechanism for MDD, and improvement in executive functioning occurring with hearing aid treatment may be a mediator of depressive symptom response. However, while HL has been associated with deafferentation-induced atrophy in frontotemporal brain regions and reduced white matter integrity in auditory pathways, discrepant results exist (16, 17) and additional studies are needed to replicate these findings.

Pilot work has demonstrated promising benefits of hearing aids for depressive symptoms, cognitive function, and brain health in older adults. For example, naturalistic assessments of neuropsychiatric status before and after hearing treatment show improvement on depressive symptoms (18), global cognition, and memory tasks (19). In addition, recent studies have begun to examine how hearing aids may restore adverse brain changes consequent to HL and increase neural activations and cortical thickness in brain regions important for auditory and language processing (20). While an ongoing open-label (i.e., researchers and participants know which treatment is being administered) randomized controlled trial will definitively test the efficacy of hearing aids vs. a successful aging education intervention on reducing cognitive decline in HL (21), few studies have compared the efficacy of hearing treatment to a control group on depressive symptoms (22). One available study comparing hearing treatment to a wait list control group observed increased self-reported quality-of-life and decreased depressive symptoms post hearing aid prescription (23).

However, open-label and wait list treatments can overestimate treatment effects, and there is a need for rigorously designed research to determine whether hearing remediation is effective for improving depressive symptoms in older adults. In a recent pilot trial of N=13 participants, we found that treatment of HL with hearing aids (versus sham) was associated with greater improvement in depressive symptoms and immediate memory (24). However, the study was limited due to small sample size, incomplete concealment to active and sham hearing aids, and inconsistent use of antidepressant medications. Therefore, larger rigorously controlled experiments with improved blinding methods and comprehensive cognitive, social, and psychiatric assessments are warranted.

In this study, older adults with comorbid HL and MDD were randomized to receive hearing aids fit to their prescription or identically appearing sham hearing aids that provided minimal amplification. Baseline magnetic resonance imaging (MRI), including diffusion tensor imaging (DTI) and structural MRI, were performed in addition to comprehensive neurocognitive testing and psychiatric assessments. Participants received hearing aids vs. sham in addition to psychiatric treatment-as-usual (TAU) for a 12-week clinical trial, at which point psychiatric and neurocognitive assessments were repeated. We hypothesized that baseline neural consequences of HL would be observable on MRI, including atrophy of auditory cortex and decreased integrity of frontostriatal white matter tracts. Further, we hypothesized that treatment with hearing aids, as opposed to sham, would be associated with greater improvement in depressive symptoms and executive functioning.

METHODS

Participants

This study was conducted at the New York State Psychiatric Institute (NYSPI), approved by the NYSPI Institutional Review Board, and registered on Clinicaltrials.gov (NCT03321006). Eligible participants were English-speaking adults aged ≥60 years who met Diagnostic and Statistical Manual 5 (DSM-V) (25) criteria for MDD or persistent depressive disorder of at least six months’ duration, had a 24-item Hamilton Rating Scale for Depression (HRSD) (26) score ≥16, and moderate-profound bilateral HL (pure tone average [PTA] of ≥50dB HL at 2 and 3 kHz in the better hearing ear). Participants were excluded for history of psychosis or bipolar disorder, hearing aid use within the past year, substance abuse/dependence within the past year, probable dementia, Mini-Mental State Examination (MMSE) score ≤24, significant suicidality, current treatment with antipsychotics or mood stabilizers, contraindication to hearing aid placement (e.g., reversible causes of HL such as active ear infections), and acute or unstable medical illness.

Study Assessments

A Structured Clinical Interview Diagnostic for DSM-V (SCID) (27) was performed at baseline to confirm participant eligibility. For audiological assessment, participants were seated in a double walled sound-attenuated booth, and pure tone testing was performed using insert earphones and bone conducted stimuli. Pure tone average (PTA) was measured as the average hearing threshold at 500, 1000, 2000, and 4000Hz in the better hearing ear. Speech reception thresholds (SRT) were obtained in each ear using standard spondee words. Speech discrimination score (SDS) was assessed in each ear using a recorded consonant-vowel nucleus-consonant type word list (25 words) at 40dB SL above the patient’s speech reception threshold. Impaired speech discrimination was defined as an SDS of <80% in the better hearing ear. The Hearing Handicap Inventory for the Elderly Screening Version (HHIE) (28) was used to assess the social and emotional effects of HL. Blinding of the hearing aid as opposed to the sham intervention was assessed by asking participants to guess their treatment assignment at the end of the study.

The 24-item HRSD was performed at every study visit. Secondary outcomes included the Social Adjustment Scale Self-Report (29) and the Short Physical Performance Battery (SPPB) (30) to provide measures of gait, balance, and lower extremity strength. The Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) (31) was utilized to examine baseline and endpoint attention, memory, language, and visuospatial/constructional ability (32). Because the RBANS requires instructions and some stimuli to be presented auditorily, the RBANS for Hearing Impaired Populations (RBANS-H) was employed in this study, all of whose components are presented in written format on an external computer monitor simultaneous with oral administration (33). Coverage of executive functioning was augmented with select tasks from the NIH Toolbox (Dimensional Change Card Sort Test [DCCST] and the Flanker Inhibitory Control and Attention Test [Flanker]) (34). Baseline DTI and structural MRI were performed on all eligible participants.

MRI Data Acquisition

Imaging data were acquired on a GE Discovery MR750 3.0 Tesla whole body scanner (GE Medical Systems, Waukesha, Wisconsin). Sequences included a sagittal MPRAGE with TI=1060ms, flip angle=8°, voxel size=1x1x1mm, and 176 contiguous slices for T1 structural imaging, and for DTI a multi-band spin-echo echo planar imaging (SE-EPI) sequence with the following parameters: 96 gradient directions (6 for b=500s/mm2, 15 for b=1000s/mm2, 15 for b=2000s/mm2, 60 for b=3000s/mm2, and 6 b=0 images, respectively), TE/TR=minimum/5000 ms, FOV=24×24 cm, matrix=140×140. 81 slices, slice thickness =1.7mm, voxel size= 1.7x1.7x1.7mm, scan time is 8.5mins.

MRI Data Preprocessing and Analysis

T1 images were processed using previously published automated segmentation processes identifying cortical and subcortical regions using FreeSurfer, followed by visual inspection of segmentation (35, 36). Regions within the auditory cortex constituted T1 regions-of-interest (ROIs): cortical thickness of anterior transverse temporal gyrus and superior temporal gyrus (lateral, planum polare, and planum temporale) (37). These were selected due to previous evidence that HL is associated with deafferentation-induced atrophy of primary and secondary auditory cortices in older adults (16, 17).

DTI ROIs included the inferior fronto-occipital fasciculus, inferior longitudinal fasciculus, superior longitudinal fasciculus (SLF), and uncinate fasciculus. These tracts are important for processing of auditory information (11) and also have been shown to underlie constituent executive functions in older adults (38). DTI data were processed using FMRIB Software Library (FSL) version 6.0.1 (Oxford, UK) (39). After movement and distortion correction using FSL Eddy (40), extraction of a brain mask using the Brain Extraction Tool (41), a diffusion tensor model was fitted using FSL DTIFIT for each voxel and used to generate fractional anisotropy (FA) images. TBSS (42) were applied to the FA images by aligning the images to a 1x1x1mm standard space, creating a mean FA image, and projecting all particpants’ FA images onto the mean FA skeleton which was then used for statistical analyses. The Johns Hopkins University (JHU) white matter tractography atlas (43, 44) was used to quantify mean FA for the selected ROIs. Nonlinear warps and skeleton projection were also applied to radial diffusivity (RD) images for statistical analyses.

Hearing aid intervention and psychiatric treatment as usual (TAU)

Hearing aids were Audeo B-R 90 devices manufactured by Phonak (Aurora, IL). Active and sham hearing aids were identical in appearance, battery use, and data logging capability. At Week 0, hearing aids were programmed to 100% gain targets, while sham hearing aids were programmed to a flat 30dB HL. We chose a 30dB HL gain for the sham hearing aid device based on results from our earlier pilot study that revealed inadequate concealment of treatment allocation when the devices were programmed to a 10dB HL gain (24). We hypothesized that providing more amplification to the sham HA group may improve blinding, though the degree of amplification was balanced against dilution of a signal for active vs. sham treatment. A 30dB HL gain was selected in order to result in a small but noticeable volume increase without substantively improving the participant’s ability to discriminate speech. To avoid worsening hearing in the sham group due to the occlusion effect of hearing aids in the ear canal, all participants randomized to sham hearing aids were fit with open domes.

Real ear measures were performed to verify fitting, education and counseling was provided regarding the use of the hearing aid, and participants were informed that a minimum of 8 hours/day hearing aid usage was required to stay in the study. After hearing aid fitting, individuals in the active and sham hearing aid groups underwent the exact same audiological procedures. Compliance with hearing aids was measured (hours/day) using data log technology built into the hearing aids, and all participants had follow-up audiology appointments at Weeks 2, 6, and 12, which served to verify fitting and provide counseling. Participants in both groups were counseled on their HL as well as proper use of the HA in order to achieve a high level of comfort with the devices.

During the study period, participants chose to start a study medication (escitalopram or duloxetine) or be continued on the medications they were already taking prior to study enrollment. For participants who chose to start a study medication, escitalopram 10mg per day or duloxetine 30mg per day were started following baseline screening. Two medication choices were allowed so that study participation could be offered to individuals who had previously taken one medication and either not responded or not tolerated it. If participants did not meet remission criteria (HRSD≤10) after 4 weeks, the dose of study medication was increased to escitalopram 20mg or duloxetine 60mg for the remainder of the study.

Data Analysis

Participants were randomized to receive hearing aid treatment vs. sham, and analysis of covariance (ANCOVA) was used to analyze the between-group differences (i.e., active-sham effect) on hearing, psychiatric, functional, and cognitive outcomes. For each measure (i.e., HHIE), the outcome modeled was the change from baseline (Week 0) to the end of study (Week 12) with predictors including baseline HHIE, group (active vs. sham), and covariates of age and education. Encompassing all participants (including those randomized to sham), this model provides baseline adjusted estimates of change within group, i.e. the difference from pre- to post-randomization for the active group, the difference from pre- to post-randomization for the Sham group, and the randomization effect between groups (active-sham effect). Effect sizes for the outcome measures were computed as Cohen’s d, using the standard deviation of the baseline measure. Differences in the number of hours of hearing aid usage between groups (active vs. sham) by time (Week 2, 6, 12) was analyzed using an analysis of variance (ANOVA) by including the predictors of group, time, and group by time interaction.

Lastly, association analyses were performed between audiometric measures (PTA and SDS) and mean FA and RD of the specified DTI ROIs using partial correlations and adjusting for age and gender as covariates. The same association analyses were performed between audiometric measures and cortical thickness of our T1 ROIs, adjusting for age, gender, and intracranial volume. Given the small sample size, the exploratory analyses did not control for multiple comparisons.

RESULTS

Participant Disposition and Characteristics

As shown in Figure 1, N=25 participants were randomized and N=23 participants were available for analyses of hearing aid treatment. As some participants were not scanning-eligible, only N=14 received baseline MRI scans. Of the N=25 participants randomized, N=20 participants started on study medications (either escitalopram or duloxetine) and N=5 participants were continued on the antidepressants they were already taking prior to study enrollment (either bupropion, sertraline, or venlafaxine). Demographic and hearing characteristics for participants are provided in Table 1.

Figure 1.

CONSORT flow diagram

Table 1.

Demographic and hearing characteristics of study participants.

	Sham (n=14)		Active (n=11)
	n	Mean (SD) or %	n	Mean (SD) or %
Sex
Male	4	28.6	4	36.4
Female	10	71.4	7	63.6
Race
Asian	1	7.7	0	0
Black/African-American	2	14.3	0	0
White	11	84.6	11	100
Ethnicity
Not Hispanic/Latino	14	100	11	100
Age	14	75 (6.5)	11	72.9 (9.2)
Education (years)	10	16.7 (2.1)	10	18.1 (4.1)
Marital Status
Single	1	7.1	3	27.3
Divorced/separated	3	21.4	3	27.3
Married/living with partner	4	28.6	2	18.2
Widowed	6	42.8	3	27.3
Employment Status
Employed full time	1	7.1	1	10.0
Employed part time	0	0	1	10.0
Retired	11	78.6	7	70.0
Other/Missing	2	14.2	1	10.0
Pure Tone Average (PTA)	14	44.3 (7.7)	11	38.8 (5)
Speech Discrimination Score (SDS, %)	14	85.4 (14.1)	11	90.9 (11.6)
Hearing Handicap Inventory for the Elderly (HHIE)	14	29.4 (8.3)	11	27.5 (7.6)
Mini-Mental State Examination (MMSE)	14	28.4 (1.2)	11	29.5 (1)

Baseline Correlations between Neuroimaging Measures and Hearing Functioning

As shown in Figure 2, impairment on pure tone testing and speech discrimination were significantly associated with cortical thinning in the primary auditory cortex (L anterior transverse temporal gyrus, PTA [r=−0.642, df=8, p=0.0445], SDS [r=0.755, df=8, p=0.0116]). Impairment in SDS at baseline also was significantly associated with decreased integrity (i.e., decreased FA and increased RD) of the SLF bilaterally (FA: L SLF r=0.772, df=6, p=0.0249, R SLF r=0.782, df=6, p=0.0219; RD: L SLF r=−0.763, df=6, p=0.0277, L SLF [temporal pole] r=- 0.782, df=6, p=0.0218, R SLF r=−0.726, df=6, p=0.0415).

Figure 2.

Associations between hearing capacities and brain structure. A. Left anterior transverse temporal gyrus (yellow). B. Right superior longitudinal fasciculus [SLF] (blue) and left SLF (red). C. Table summarizing significant associations between hearing variables and T1 MRI and DTI auditory regiors-of-interest. Abbreviations: PTA = pure tone average; SDS = speech discrimination score; r = partial correlation coefficient; FA = fractional anisotropy; RD = radial diffusivity.

Hearing Aid Treatment Effects

As shown in Table 2, after adjusting for age, education and baseline scores, a significant improvement in HHIE was noted for the hearing aid compared to the sham group (active −12.47 vs. sham −4.19, ES=1.05, t=−2.64, df=18, p=0.016). No significant differences were found between hearing aid and sham groups on depressive symptoms (HRSD: active −5.50 vs. sham - 7.32, ES=0.38, t=0.75, df=19, p=0.464). Hearing aid treatment as opposed to sham was associated with a significant improvement on RBANS Immediate Memory (active +14.9 vs. sham +5.7, ES=0.62, t=2.28, df=16, p=0.037). Hearing aid treatment was also associated with improvement on the NIH Toolbox Flanker, though the results failed to achieve statistical significance (active +4.9 vs. sham −2.4, ES=0.63, t=1.95, df=15, p=0.071). No other significant between-group differences were noted on cognitive tasks, nor was there differential improvement noted in social functioning as measured by the SAS-SR or physical functioning as measured by the SPPB. Due to the limited sample size, after adjusting for multiple testing, the p-values were not significant.

Table 2:

Hearing aid treatment effects on hearing, psychiatric, functional, and cognitive outcomes.

Hearing
Assessment		Sham Group			Active Group			Between Group Difference
	Time	n	Adjusted Mean	SE	n	Adjusted Mean	SE	Mean Differencea	p- value	Effect Size b	t	df
Hearing Handicap Inventory for the Elderly (HHIE)	Week 0	14	28.56		11	28.56
	Week 12	12	24.37	2. 11	11	16.09	2. 15	−8.28	0.016	1.05	2.64	18
Psychiatric
Outcome		Sham Group			Active Group			Between Group Difference
	Time	n	Adjusted Mean	SE	n	Adjusted Mean	SE	Mean Difference	p- value	Effect Size	t	df
24-item Hamilton Rating Scale for Depression (HRSD)	Week 0	14	19.36			19.36
	Week 12	11	12.04	1.72	11	13.86	1.67	1.81	0.464	0.38	0.75	19
Social Adjustment Scale Self-Report (SAS-SR)	Week 0	14	2.33		11	2.33
	Week 12	11	1.96	0.12	11	2.10	0.13	0.15	0.427	0.41	0.81	17
Functional
Short Physical Performance Battery	Week 0	14	9.24		11	9.24
	Week 12	11	9.25	0.50	10	10.17	0.54	0.93	0.236	0.34	1.23	16
Cognitive
RBANS
Immediate Memory	Week 0	14	105.42		11	105.42
	Week 12	11	111.13	2.57	10	120.28	2.71	9.16	0.037	0.62	2.28	16
Delayed Memory	Week 0	14	102.58		11	102.58
	Week 12	11	108.55	2.00	10	108.1	2.11	−0.40	0.902	0.03	−0.12	16
Attention	Week 0	14	104.54		11	104.54
	Week 12	11	101.17	3.86	10	104.70	3.98	3.53	0.556	0.23	0.60	16
Language	Week 0	14	105.71		11	105.71
	Week 12	11	108.35	3.73	10	103.98	4.06	−4.36	0.456	0.25	−0.76	16
Visuospatial/Const ructional	Week 0	14	92.04		11	92.04
	Week 12	11	84.69	3.28	10	90.23	3.51	5.53	0.280	0.35	1.12	16
NIH Toolbox
FLANKER	Week 0	14	87.83		11	87.83
	Week 12	10	85.43	2.53	10	92.70	2.63	7.28	0.071	0.63	1.95	15
DCCST	Week 0	14	107.75		11	107.7
	Week 12	10	103.41	3.92	10	110.29	4.06	6.88	0.263	0.36	1.16	15

^aMean difference reflects the difference in adjusted mean scores between active and sham groups at Week 12, such that positive numbers denote higher

Week 12 scores for active (vs. sham) and negative numbers denote higher Week 12 scores for sham (vs. active).

^bEffect size computed as Cohen’s d, using the standard deviation of the baseline measure.

Two total participants dropped out of the study (14.2%, in the sham group), including one participant who died due to medical illness unrelated to study participation. Mean hearing aid usage (hours/day) did not significantly differ between groups at the Week 2, 6 or 12 time points (F(2, 48)=0.34, p=0.711). Four participants had less than the minimum prescribed 8 hours/day mean hearing aid usage. Among those assigned to sham, 92.3% correctly guessed they were receiving sham. Among those receiving active treatment, 54.5% correctly guessed they were receiving active hearing aids and 45.5% guessed sham.

CONCLUSIONS

In this study, older adults with comorbid HL and depression were associated with significant cortical thinning and lower integrity of white matter tracts in brain regions responsible for executive functioning and speech processing. Hearing aid treatment was effective in improving hearing functioning and significantly reducing HL-related disability, but no significant effect of treatment on the primary depression outcome was observed. Hearing aids, as opposed to sham, were associated with significant improvements in immediate memory, a constituent executive function. A moderate effect size improvement was observed for hearing aids on executive functioning, but this finding did not reach statistical significance. Blinding for the study was incomplete: whereas adequate concealment of treatment allocation was achieved for the active treatment group, more than 90% in the sham group correctly guessed their assignment.

The associations between HL and brain structure observed at baseline are consistent with prior findings of deafferentation-induced atrophy consequent to diminished input to central auditory regions (10). Decreased brain volumes and cortical thinning associated with HL have been demonstrated in frontotemporal regions, particularly those implicated in auditory and language processing (9). Similarly, reduced integrity of the SLF in hearing-impaired individuals in consistent with this tract’s role in language-related processing, including auditory comprehension and phonological processing (45). Interestingly, diminished SLF integrity has been observed in depressed older adults compared to age-matched controls (46), and we have reported positive correlations between SLF integrity and executive functioning performance (38). Thus, the SLF presents itself as an intriguing mechanistic link between hearing changes and the development of depression and executive dysfunction that merits further study.

Contrary to our hypotheses, hearing aid treatment was not associated with improved depressive symptoms relative to sham. One possibility is that improvement of hearing functioning is not relevant to MDD severity, while another is that it our study was small and underpowered. It may be the case that observation periods longer than 12 weeks would be useful in detecting the effects of hearing aids on depressive symptoms. As it can often take months to maximize the rehabilitative effects of hearing aids, the study period may not have provided sufficient time for hearing aid adjustment, especially as their effects on depression may be moderated by salutary behavioral changes (e.g., increased social engagement, decreased isolation and loneliness) that take place on a longer time scale. It is also possible that antidepressant treatment may have resulted in the improvement of depressive symptoms across treatment arms and therefore placed a ceiling on further improvement attributable to hearing aids. Lastly, defining patient groups based on PTA may be insufficiently precise to classify the specific subgroups of hearing-impaired older adults at greatest risk for depressive and cognitive problems. For example, functional assessments of speech discrimination in quiet and noise may identify individuals who may benefit the most from hearing rehabilitation. Larger future studies designed to comprehensively understand the hearing, cognitive, and social/interpersonal characteristics of older adults at greatest risk for MDD may facilitate targeting vulnerable individuals with treatments tailored to the type of dysfunction present.

Even in this small pilot study, there was a suggested signal of effect for hearing aids on immediate memory and response inhibition. This is an intuitive result, since as hearing capacity declines, older adults must expend increasing effort to discriminate speech in the presence of background noise, such as a social event in a noisy venue. At a neural level, this increased effort is reflected in greater engagement of frontal brain systems and the CCN during listening, and it is hypothesized that taxing the CCN through effortful listening diminishes the resources available for higher level cognitive processing and may trigger a cascade of executive dysfunction and MDD (47). Such a model is consistent with our above finding of an association between worse speech discrimination scores and lower SLF integrity. In addition, we have found in a previous analysis of the National Alzheimer’s Coordinating Center Uniform Data Set that older adults with more severe and longer lasting HL were more likely to have executive dysfunction after controlling for multiple covariates, including APOE genotype, than those without HL (48). Executive dysfunction (i.e., lower performance on Trail Making Test [TMT] Part B after controlling for TMT Part A) was associated with increased longitudinal risk for dementia. Consistent with our observed signal of an effect of hearing aids on improved executive functioning, enhancing hearing ability may reduce the effort and time required to decode speech and free up cognitive resources. Given the links between executive dysfunction and MDD in older adults, we hypothesize that improvement of executive functioning may be important step in relieving depressive symptoms. However, observation periods longer than 12 weeks may be required to observe whether executive functioning occurring with hearing aid treatment may be a mediator of depressive symptom response. Larger studies incorporating comprehensive neurocognitive assessments of each cognitive domain, perhaps especially executive functioning, are needed to follow up this finding.

These findings must be considered in light of several important limitations, the most notable of which was the small sample size, which precluded detection of small to medium treatment effect sizes on the order we observed. Given the small sample size, the exploratory analyses did not control for multiple comparisons and larger studies are needed to replicate our findings given the possibility of Type I error. In addition, as the study only included participants with moderate-profound HL and MMSE above 24, results may not be generalizable to individuals with only mild HL or dementia. Lastly, incomplete blinding was achieved with the use of sham hearing aids, and methods for improving the concealment of active vs. sham hearing aids should be investigated. While the active hearing aid group was adequately blinded, the inadequate treatment concealment in the sham hearing aid group may have led to negative expectations of treatment and have resulted in a nocebo effect (e.g., worse hearing, cognitive, and depressive outcomes). Mitigating this limitation is the fact that the active condition was blinded in this study (i.e., participants assigned to active hearing aids correctly guessed their group at roughly 50/50), minimizing a potential increase in the placebo effect component of response attributable to unblinding in this group.

DATA STATEMENT

The data has not been previously presented orally or by poster at scientific meetings.

Highlights.

• This pilot study aimed to examine the neural effects of hearing loss and the impact of hearing aids on neuropsychiatric outcomes among older adults with comorbid major depressive disorder (MDD) and age-related hearing loss.

• At baseline, hearing loss was associated with significant cortical thinning and lower integrity of white matter tracts in brain regions responsible for executive functioning and speech processing. Hearing aid treatment over the 12-week study was effective at reducing hearing loss-related disability, but no effect on depression was observed. There was a suggested signal of effect for hearing aids on immediate memory and response inhibition, both constituent executive functions.

• Hearing loss can have significant neural and psychiatric consequences in older adults, and larger studies may evaluate whether hearing aids may improve cognitive and depressive symptoms in older adults with MDD.