Anxiety-like behavior in deaf Ebf1 conditional knockout mice

TL;DR

This study shows that deaf mice with a specific genetic modification display anxiety-like behaviors, supporting the link between hearing loss and mental health issues.

Contribution

The study provides new empirical evidence that hearing loss contributes to anxiety-like behaviors using a genetically modified mouse model.

Findings

KO mice spent significantly less time in the open arms of the elevated plus maze.

KO mice showed increased latency to enter the open arms and reduced acoustic startle response.

There were no significant differences in exploratory behavior or forced swim immobility time.

Abstract

Hearing is critical for communication and clinical evidence suggests that hearing loss can lead to poorer mental well-being including social isolation, anxiety, depression, and cognitive decline. Deaf animal models offer opportunities to investigate the impact of hearing loss on behavioral correlates of mental health while controlling for genetic variability and shared etiological factors such as environmental stressors, aging, and metabolic diseases. We previously showed that conditional deletion of Ebf1 in the inner ear causes deafness. We hypothesize that otic-specific Ebf1 knockout (KO) mice recapitulate neurobehavioral alterations experienced in congenital deafness.

Slc26a9P2A-Cre and Ebf1 floxed mice were crossed previously to generate the otic-specific Ebf1 conditional KO. We measured auditory brainstem response and behavior in groups of otic-specific Ebf1 conditional KO mice and Cre-negative littermate controls (WT). We performed behavioral tests including elevated plus maze, acoustic startle, locomotor activity, forced swim, and Y-maze.

We found that the KO mice spend significantly less time than WT mice in the open arms of the elevated plus maze. KO mice also show significantly increased latency to enter the open arms. Acoustic startle response was decreased in KO mice vs. WT controls and pre-pulse inhibition was also reduced in the KO, consistent with elevated hearing thresholds. We found no significant differences in exploratory behavior, forced swim immobility time, or other behavioral measures examined.

Our findings demonstrate that deaf otic-specific KO mice are indeed a model for the study of effects of hearing loss on clinically relevant behavioral correlates. Our results provide new empirical evidence that hearing loss contributes to anxiety, which is consistent with clinical evidence that anxiety is associated with hearing loss in older adults.

Figures

a](#F2)). No group differences were detected in total arm entries, closed arm entries, and open arm entries in WT vs. KO mice (*p* = 0.7024, 0.4003, and 0.1574, respectively, [Fig. 2c](#F2)), suggesting exploratory behavior is maintained despite differences in arm preference. Furthermore, we found no group differences in total distance traveled in the maze during the trials (*p* = 0.3452, [Fig. 2d](#F2)), suggesting activity levels also were similar in WT and KO mice.

a](#F6)). Testing the difference-in-differences contrast (70–100 dB) in WT and KO showed no significant interaction (estimate = 0.055, SE = 0.83, t(99) = 0.066, *p* = 0.95), indicating that startle growth across 70–100 dB was comparable between genotypes. However, startle response amplitudes in KO mice were significantly lower than those in WT at each sound level ([Fig. 6a](#F6)). This reduction in startle is consistent with elevated ABR thresholds in KO mice (i.e., 73.93±2.58 for click, [Fig. 1](#F1)). Furthermore, startle amplitudes in response to a 100 dB noise burst did not differ signific

Full text

Introduction

Hearing is critical for communication and, early in life, hearing loss has serious risks for children’s language, emotional, social, and cognitive development (Moeller, 2000; Yoshinaga-Itano et al., 1998; Stevenson et al., 2015; Khalid et al., 2025). Later in life, hearing loss is associated with greater odds of anxiety, depression, and cognitive decline. Clinically, meta-analyses estimate 83 % greater odds of anxiety with self-reported and audiometrically confirmed hearing loss (Zhang et al., 2023). A prospective study shows dose-dependent increase in odds of anxiety with hearing loss severity: mild: OR = 1.32 (1.01–1.73); moderate or greater: OR = 1.59 (1.14, 2.22) (Contrera et al., 2017). A similar pattern is reported for depression: a large cross-sectional study demonstrates increase in odds of depression symptoms with self-reported hearing loss severity: mild: OR = 1.29 (1.14–1.47); severe: OR = 1.59 (1.22, 1.87) (Cosh et al., 2018). Meta-analyses which considered additional studies estimate similar levels of risk for depression (Zhang et al., 2023; Lawrence et al., 2020). Beyond emotional well-being, cognitive decline in older adults is also predicted by hearing loss, with a prospective study reporting 22 % greater odds of impairment after hearing loss (Loughrey et al., 2018).

While clinical studies identify clear associations, confounding variables including age, genetic variability, nonuniform environmental exposures, and co-morbidities limit the ability to determine the contribution of hearing loss in cognitive decline and mental health disorders. By contrast, deaf animal models offer opportunities to investigate the impact of hearing loss on behavioral correlates of cognitive function and mental health while rigorously controlling for shared etiological factors such as environmental stressors, aging, metabolic diseases, and the timing and severity of hearing loss. If controlled studies substantiate cognitive and emotional consequences of hearing loss, animal models will also be an important tool for investigation of underlying mechanisms across the lifespan.

We previously showed that conditional deletion of Ebf1 in the inner ear causes severe auditory brainstem response deficits associated with cochlear developmental defects (Powers et al., 2024). Because Slc26a9^P2A-Cre^ selectively drives recombination in otic epithelium and the spiral ganglion neurons, but not in the brain, this model enables peripheral auditory-specific deletion without central effects (Urness et al., 2020), providing an opportunity to test how congenital hearing loss affects anxiety, depression, and cognitive function. We hypothesized that otic-specific Ebf1 knockout (KO) mice recapitulate neurobehavioral alterations experienced in congenital deafness, and therefore compared anxiety and depression-related behaviors, as well as cognitive function, in WT and otic specific Ebf1 KO mice.

Methods

Mice.

Otic-specific Ebf1 KO and control mice were generated previously (Powers et al., 2024) by crossing Ebf1^fl/fl^ mice (Vilagos et al., 2012) on a C57BL/6 background (Jackson #028,104) and Slc26a9^P2A-Cre^ mice (MGI #6715,244) on a C57BL/6 background (Urness et al., 2020). A new cohort was generated for this testing by breeding Ebf1^fl/fl^; Slc26a9^P2ACre^-postive male mice and Ebf1^fl/fl^ female mice. To identify Slc26a9^P2A-Cre^-positive; Ebf1^fl/fl^ knockout (KO) mice and Slc26a9^P2A-Cre^-negative; Ebf1^fl/fl^ wildtype (WT) littermate controls, we either performed genotyping on ear punch samples by qPCR as described previously (Powers et al., 2024) or by using the Slc26a9^P2A-Cre^ and Ebf1-floxed qPCR genotyping assays available through Transnetyx (Cordova, TN).

Experimental and control mice were group-housed with same-sex littermates (up to 5 per cage) and provided ad libitum access to food (LabDiet Picolab 5V75; Land O’Lakes, Inc. Arden Hills, MN) and water (Lixit autowatering; Napa, CA). Cages had corn cob bedding and shredded paper nesting material for enrichment. All experiments and procedures were approved by the Institutional Animal Care and Use Committee of the Medical University of South Carolina. Mice were housed in the Medical University of South Carolina Division of Lab Animal Resources.

ABR.

Mice were anesthetized with an intraperitoneal injection of ketamine (90 mg/kg) and xylazine (10 mg/kg). Body temperature was maintained at 37 °C using a DC-powered heating pad and rectal probe (ATC2000, World Precision Instruments; Sarasota, FL).

ABRs were recorded inside of a radiofrequency-shielded acoustic chamber (MAC-2, IAC Acoustics; North Aurora, IL). We placed subcutaneous electrodes (Rhythmlink RLSP483; Columbia, SC) at the cranial vertex for recording, behind the pinna on the stimulus side for reference, and behind the contralateral pinna for grounding. Using an RZ6 system (Tucker Davis Technologies, TDT; Alachua, FL), presented calibrated tone pips (8, 16, 32, and 45 kHz) and broadband clicks from 0–90 dB SPL (5 dB steps).

Experienced readers manually determined thresholds as the lowest sound levels that elicited wave I patterns by visual inspection of the stacked responses across decreasing sound levels in BioSigRz (TDT). Blinding was not feasible in this ABR analysis due to pronounced threshold elevation in KO mice, making them readily identifiable. Additionally, the determination of thresholds in KO mice requires careful inspection of responses at high intensity compared to baseline (0 dB SPL) traces. In otic-specific Ebf1 KO mice, ABR waveforms are tiny, but wave I can usually be identified in this way.

Plots show mean ± SE of ABR thresholds. To measure the effects of genotype, frequency, and sex on hearing sensitivity we fit a linear model using lm(threshold ~ genotype * frequency * sex) in R. No sex differences were found (F = 36.26, p = 0.3901). We then used emmeans (model, pairwise ~ genotype | frequency, adjust = “holm”) to quantify effects of KO stratified by frequency with multiple testing correction across frequencies. ABR testing was performed between P29 and P63 prior to behavioral testing.

Behavioral testing.

After auditory functional tests, mice underwent a battery of neurobehavioral tests: photobeam locomotor activity, y-maze, elevated plus maze, forced swim, acoustic startle, and prepulse inhibition. Forced swim and acoustic startle testing were performed last due to the stressful nature of these tests. Mice were acclimated for at least 2 weeks in a reverse light cycle room (lights off 8 am-8 pm) to allow testing during their active (dark) phase. Staff conducting the behavioral testing were blinded to genotype.

Hearing loss places children at risk for language, emotional, social, and cognitive developmental difficulties (Moeller, 2000; Yoshinaga-Itano et al., 1998; Stevenson et al., 2015; Khalid et al., 2025). After early childhood, anxiety and depression can emerge, typically by young adulthood, representing another window of vulnerability. We conducted testing at 3–5 months of age, which corresponds to approximately 20–30 years old in humans (Dutta and Sengupta, 2016). We chose this mature reproductive stage for study of congenital hearing loss’s effects after neural maturation but before aging in a critical window for mental health risk.

Elevated plus maze.

We placed mice in the center of a plexiglass elevated maze shaped like a “plus” sign. Two arms of the plus maze have open sides and two have tall, opaque walls. Exploratory activity was monitored, tracked, and analyzed using Anymaze (Stoelting; Wood Dale, IL). The duration of the test was 5 min. We tracked arm entries by type, time spent in each arm type, and total arm entries and distance traveled.

Y-maze.

Mice explored a 3-arm Y-maze for 5 min. Exploration was recorded, tracked, and analyzed using video-tracking software (AnyMaze). Spatial working memory is measured as the percentage of correct spontaneous alternations, where a mouse enters the least recently explored of the 3 arms of the maze without re-visiting a recently explored arm. Total distance and total number of arm entries were also tracked to examine overall exploratory behavior and locomotor activity during this testing.

Photobeam locomotor activity.

We placed mice in a standard mouse cage surrounded by an infrared beam array (SDI, San Diego, CA) within a larger opaque container in a dark room. We allowed the mice to explore in the dark for up to 1 hour and measured their activity by recording the number of infrared beam disruptions every 5 min.

Forced swim.

Mice were placed in a container of water (4 L beaker, 23–25 °C) for 6 min and swimming activity was recorded by a video camera. The water was changed between every mouse. Mice alternate between swimming and floating during the test. Videos were analyzed in Anymaze to determine the percentage of time spent immobile and the number of immobile–mobile transitions. The first two minutes of the recording were excluded from analysis to allow for acclimation.

Acoustic startle & prepulse inhibition.

We placed mice in a darkened, sound-attenuated, startle reflex testing chamber (Med Associates) with 65-dB white noise, allowing them to acclimate for 2 min. We measured startle responses to randomly presented calibrated startle stimuli (70–100 dB SPL) via a load cell platform for 400 ms from the startle onset. To measure prepulse inhibition (PPI), some startle stimuli were preceded 50 ms by a 70 dB prepulse tone. PPI was calculated as PPI = ((NB100 - CS) / NB100) * 100, where NB100 is the startle level in response to 100 dB noise burst and CS is the startle response to 100 dB noise burst when preceded by the prepulse.

Statistical testing and plots.

Plots show mean ± SE. For behavioral measures examined at a single time point, we fit a linear model using lm () in R of the following form: behavior ~ genotype * sex * test-specific variables. Additional variables were included where appropriate based on specific features of each test such as arm types in elevated plus maze, immobility and mobility in forced swim, and sound levels in acoustic startle responses. Percentage measures were transformed using Yeo-Johnson transformation to stabilize variance (Yeo and Johnson, 2000). This transformation handles 0 and negative values which is a limitation of arcsine-square root transformation. For paired measures, we performed linear mixed-effect modeling using lme4 in R including between-subjects fixed effects (genotype, sex) and within-subject fixed effects (time points, frequencies). Mouse ID was included as a random intercept to account for within-subject correlation of repeated measures. No sex differences in behavioral responses were found and sex was therefore excluded from subsequent analysis. For pairwise contrasts of KO and WT, we used emmeans(model, pairwise ~ between subjects factors | within-subject factors, adjust = “holm”), with multiple testing correction across timepoints. We used a linear mixed-effect model for startle amplitude with genotype and sound level as fixed effects and a random intercept for mouse ID. We then constructed difference-in-differences contrasts to test whether magnitude of the genotype effect differed specifically between 70 dB and 100 dB.

Results

ABR thresholds.

As in our previous cohort (Powers et al., 2024), otic-specific Ebf1 KO mice exhibited significantly elevated hearing thresholds compared to WT controls across all frequencies tested (Fig. 1). KO mean hearing thresholds exceeded 70 dB for all stimuli, with elevations of 35.1, 27.4, 37.6, 32.4, and 20.7 dB for click, 8, 16, 32, and 45 kHz, respectively (p < 0.0001 for all).

Elevated plus maze.

Avoidance of open arms in the elevated plus maze is a spontaneous approach-avoidance behavior in mice (Pellow et al., 1985). We found that KO mice spent 63.45 % less time in the open arms of the elevated plus maze compared to WT mice, which is equivalent to a 2.74-fold reduction (p = 0.0065, Fig. 2a). In addition, KO mean latency to enter the open arms was increased by 41.28±19 s, which is 9.76-fold longer than that in WT controls (p = 0.0344, Fig. 2b).

Both WT and KO mice demonstrated the typical preference for the closed arms of the maze, spending 80.85 % and 93.83 % more time in the closed arms, respectively (p < 0.0001, Fig. 2a). No group differences were detected in total arm entries, closed arm entries, and open arm entries in WT vs. KO mice (p = 0.7024, 0.4003, and 0.1574, respectively, Fig. 2c), suggesting exploratory behavior is maintained despite differences in arm preference. Furthermore, we found no group differences in total distance traveled in the maze during the trials (p = 0.3452, Fig. 2d), suggesting activity levels also were similar in WT and KO mice.

Y-maze.

In a 3-armed Y-shaped maze, mice readily explore but typically avoid entering the arms they have most recently entered, demonstrating intact spatial working memory function (Dennis and Sollenberger, 1934). Reduced correct spontaneous alternations are indicative of a deficit in working memory. Correct spontaneous alternation percent vs. WT did not differ significantly (53.65 % vs. 58.89 %, respectively, p = 0.1787, Fig. 3). No group differences were detected in total number of arm entries (p = 0.7151) or the total distance traveled (p = 0.5669), suggesting that KO mice maintained similar levels of exploratory behavior and locomotor activity in this testing.

Photobeam activity assay.

Photobeam activity systems quantify spontaneous locomotor activity of mice in an open field as photobeam breaks. Mice avoid illuminated open spaces and activity in such conditions is influenced by anxiety. To minimize the influence of anxiety on activity and better isolate exploratory behavior, we conducted the photobeam test in a dark box. Under low anxiogenic (dark) conditions, WT and KO mice exhibited similar rates of habituation during the test and there were no significant differences in number of beam breaks at any time points (Fig. 4), indicating comparable exploratory drive, locomotor function, and habituation to novelty. There were no significant differences in total beam breaks between WT and KO mice (p = 0.8309, Fig. 4b).

Forced swim.

Immobility in the forced swim test is a despair-like state characteristic of depression (Porsolt et al., 1977). The amount of time spent immobile during the forced swim test did not differ between groups (Fig. 5). Both groups spent significantly more time immobile than mobile, which indicates that the test was effective in inducing a state of behavioral despair: WT mice spent 2.04-fold more time immobile and KO mice spent 1.70-fold more time immobile than mobile (p < 0.001 and p = 0.0060, respectively). While overall swimming duration during the trials was similar across groups, the latency to immobility was increased 2.01-fold in KO relative to WT (p = 0.0250).

Acoustic startle response & prepulse inhibition.

Acoustic startle response testing measures movements in response to sudden noise burst and can assess hearing sensitivity, motor reflexes, and sensorimotor gating, which is indicated by prepulse inhibition of the startle (Hoffman and Ison, 1980; Longenecker et al., 2016). Both WT and KO mice showed startle responses that increased with sound intensity (Fig. 6a). Testing the difference-in-differences contrast (70–100 dB) in WT and KO showed no significant interaction (estimate = 0.055, SE = 0.83, t(99) = 0.066, p = 0.95), indicating that startle growth across 70–100 dB was comparable between genotypes. However, startle response amplitudes in KO mice were significantly lower than those in WT at each sound level (Fig. 6a). This reduction in startle is consistent with elevated ABR thresholds in KO mice (i.e., 73.93±2.58 for click, Fig. 1). Furthermore, startle amplitudes in response to a 100 dB noise burst did not differ significantly with vs. without prepulse Fig. 6b, perhaps reflecting variable prepulse audibility among KO mice. By contrast, WT mice showed 32–36 % reduction in startle amplitude when the 100 dB noise burst was preceded by prepulse tones (Fig. 6b). Percent prepulse inhibition (% PPI, defined as the percent reduction in startle amplitude in response to 100 dB noise burst when preceded by a 70 dB prepulse tone) was significantly reduced in KO mice relative to WT mice (2.01-, 0.95-, 0.92-fold reductions; p = 0.0117, 0.0046, and 0.0130 for 4, 12, and 20 kHz pre-pulse tones, respectively, Fig. 6c). Reductions in startle and PPI likely reflect reduced audibility in KO mice (Fig. 1), rather than a change in central sensorimotor-gating.

Discussion

Our findings show that otic-specific Ebf1 knockout (KO) mice exhibit some neurobehavioral alterations associated with congenital deafness, particularly anxiety-like behaviors. KO mice also showed reduced acoustic startle but no significant differences in tests for despair-like behavior or spatial working memory. We controlled for genetic, environmental, and systemic variables that often confound clinical and experimental studies of hearing loss’s effects on mental health by using a conditional genetic strategy that selectively targets the auditory periphery in age-matched littermates on an inbred background. These findings have implications for understanding hearing loss’s effects on mental health and cognitive decline.

Anxiety-like behavior.

Regarding anxiety, we found that the KO mice spend significantly less time than WT mice in the open arms of the elevated plus maze (Fig. 2). KO mice also show significantly increased latency to enter the open arms. Previous reports give mixed results on elevated plus maze testing in rodents with hearing loss. Similar to our findings, Ucn null mice were also characterized by elevated hearing thresholds and were likewise shown to spend less time in the open arms of the elevated plus maze relative to wildtype controls (Vetter et al., 2002). Also similar to our findings, open arm time was reduced in a model of salicylate-induced hearing loss and tinnitus (Kim et al., 2025). Furthermore, nonsignificant reductions in open arm time and entries were reported in blast-exposed rats with elevated compound action potential thresholds (Manohar et al., 2020). By contrast, deaf Cacna1d and Cldn14 null mice showed preference for open arms, but under testing conditions that limit comparison: testing was performed in the light phase, when mice are less active, and under low red light, which may have limited mice’ ability to perceive the height of the maze, attenuating the anxiogenic stimulus of this testing (Busquet et al., 2010). It is important to note that global knockout, salicylate, and acoustic trauma models have central effects or stressor confounds, hampering interpretations about the role of peripheral deficits on behavioral differences. Clinical studies show greater odds of anxiety with hearing loss (Zhang et al., 2023; Contrera et al., 2017). By using an otic-specific conditional KO approach, we can attribute anxiety-like behavior in the elevated plus maze more confidently to peripheral hearing loss.

Cognitive function.

Hearing loss is the greatest modifiable risk factor for dementia (Livingston et al., 2020), however, important etiological questions remain due to the limited number and scope of controlled studies on the cognitive effects of hearing loss. A review of 20 animal studies examining cognitive behavioral endpoints after hearing loss found cognitive impairment in 30 % of the studies (Jagersma et al., 2025).

One study using a cognitively demanding 8-arm maze and identified more working memory errors (i.e., re-entries into the previously entered arm) following hair cell ablation (Qian and Ricci, 2020). Another study found significant spatial memory deficits in blast-exposed rats using the Morris water maze (Manohar et al., 2020). A recent study reports slight reduction in object recognition in noise-exposed rats with mild hearing loss (Jagersma et al., 2024).

In our study using the 3-arm Y-maze, KO mice showed a trend toward reduced correct spontaneous alternations (Fig. 3); however, this difference was not statistically significant. This is consistent with the neutral findings reported in most previous studies of cognitive function in animals with hearing loss (Jagersma et al., 2025). Furthermore, photobeam activity testing showed KO exploratory behavior is intact (Fig. 4), which serves as an additional indicator of spatial memory and cognitive function.

Altogether, these findings suggest that Y-maze and open field exploratory activity tests may lack sufficient sensitivity to detect subtle cognitive effects of hearing loss. While hearing loss appears unlikely to cause major spatial working memory deficits by young adulthood, it may promote anxiety-related behaviors. However, a contribution of hearing loss to cognitive decline at later ages in life cannot be ruled out. In addition to the age at testing, the age of hearing loss onset may have influenced the results. In the study by Qian and Ricci (Qian and Ricci, 2020), mice that underwent hair cell ablation in adulthood had more working memory errors than those ablated at an immature stage, suggesting a late critical period whereas KO mice in our study may have adapted to hearing deficits during a developmental window of heightened compensatory neuroplasticity.

Because aging is the greatest factor in cognitive decline, more experiments are needed to determine whether early hearing loss contributes to cognitive decline later in life. However, an extended time gap between hearing loss and the emergence of cognitive decline reduces the plausibility of a direct causal relationship. The association may be mediated by intermediate processes which accumulate over time. For example, hearing loss can lead to social isolation, a major risk factor for cognitive decline, in addition to increased cognitive load for listening and anxiety, both of which impair attention and memory (Wingfield et al., 2005; Mogg et al., 2015; Najmi et al., 2012; Wall and Messier, 2000). Over time, these factors may deplete cognitive reserve, reducing protection against neurodegenerative pathology (Livingston et al., 2024).

The C57BL/6 background used in this study shows accelerated age-related hearing loss and OHC degeneration caused by the Cdh23^Ahl^ mutation (Johnson et al., 1997; Martin et al., 2007). To enable long term studies and to isolate the Cdh23^Ahl^ effects, we are backcrossing the KO strain onto a stable-hearing background. Future studies should also incorporate testing that varies in complexity, retention duration (e.g., working vs. long-term memory), and targets additional cognitive and memory domains such as attention, executive function, and recognition memory. Such approaches will be essential for identifying specific domains of vulnerability and for clarifying the mechanisms through which hearing loss affects mental health and function throughout life.

Depression-like behavior.

Clinical studies demonstrate that odds of depression increases with severity of hearing loss (Cosh et al., 2018). Immobility in the forced chamber swim, in contrast to swimming and struggle to escape, is a passive and depression-like state of despair (Porsolt et al., 1977). While we found no difference in WT and KO immobility time in the forced swim test, we observed greater latency to immobility in the KO mice relative to that in the WT group (Fig. 5). We speculate that the prolonged period of swimming observed in the KO is a fear-driven response reflecting heightened anxiety in this group, rather than reduced despair. Similarly, a previous study reported higher latency to immobility in an “anxious” group of rats which displayed low-open arm time in elevated plus maze testing relative to that in groups with medium- and high-open arm time (Estanislau et al., 2011). Also consistent with our findings, deaf Cldn14 null mice showed no difference in immobility during forced swim testing compared to wildtype controls (Busquet et al., 2010). Immobility was increased after salicylate (Kim et al., 2025), however, this treatment has both central and peripheral effects. Our findings indicate that hearing loss does not affect the acute despair-like immobility state in the forced swim test or alter locomotor and exploratory activity. However, future studies should assess chronic depression-related characteristics such as anhedonia as well as feeding and sleep disturbances.

Acoustic startle and sensorimotor gating.

Our findings showed decreased startle response and a significant reduction in prepulse inhibition in KO mice compared to WT mice (Fig. 6), however, these differences likely reflect diminished hearing sensitivity (Fig. 1), rather than altered central sensorimotor-gating or reactivity to the loud sound. Future studies should evaluate startle and sensorimotor-gating in response to non-acoustic stimuli.

Conclusions and future directions.

At the young adult stage, otic-specific Ebf1 KO mice with hearing loss display greater anxiety-like avoidance of open elevated spaces in our testing, but no differences in despair-like behavior or spatial working memory. Our findings reinforce that deaf mice are a useful model to study the effects of hearing loss on behavioral correlates of mental health and provide opportunities for mechanistic study.

Mechanistically, we hypothesize that peripheral auditory impairments disrupt pathways linking auditory- and emotion-related regions. Studies in animals and humans with hearing loss show that emotionally-significant sounds such as social vocalizations, warnings, and environmental threats stimulate the amygdala and hippocampus (Kraus and Canlon, 2012). In mice, we speculate that the inability to hear warnings and other vocalizations might reduce situational awareness, resulting in surprise encounters, territoriality, fighting, and social isolation. In addition, ambient sound is important for spatial awareness, postural control, and balance (Gandemer et al., 2017; Kanegaonkar et al., 2012; Zarei et al., 2022; Termoz, 2004), and hearing loss may otherwise add to anxiety while navigating the elevated plus maze.

Future priorities include comparing anxiety-like behaviors in models with differing levels and types of hearing dysfunction and studying the stability of anxiety phenotypes across the lifespan. The present findings indicate that congenital hearing loss alone is sufficient for anxiety in young adulthood. Assessing anxiety at later stages could clarify whether prolonged hearing loss confers cumulative risk for anxiety. Inducing hearing loss at different life stages could help to define boundaries for age-dependent susceptibility to hearing loss-induced anxiety. Also, avoidance of heights and open spaces is innate and potentially different in mechanism from anxieties related to conditioned/learned fears. Future studies should evaluate other aspects of emotional behavior following hearing loss such as sucrose preference, light-dark preference, social interaction testing, and non-acoustic fear conditioning and inhibition.

Our findings highlight that hearing loss is sufficient to promote anxiety-like behavior. Hearing loss is common, particularly in older adults, however, most adults with hearing impairment (~86 %) do not use hearing aids (Hoffman et al., 2017; Chien and Lin, 2012). Future clinical studies are therefore needed to determine which interventions are most effective for management of hearing loss-induced anxiety and to identify the groups at the most risk.

Supplementary Material

Appendix

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.heares.2025.109510.