Long Term Changes to Auditory Sensitivity Following Blast Trauma in Mice

Abstract

Blast trauma is a common acoustic/physical insult occurring in modern warfare. Twenty percent of active duty military come into close proximity to explosions and experience mild to severe sensory deficits. The prevalence of such injuries is high but correlating auditory sensitivity changes with the initial insult is difficult because injury and evaluations are often separated by long time periods. Here, auditory sensitivity was measured before and after a traumatic blast in adult CBA/CaJ mice using auditory brainstem responses, distortion production otoacoustic emissions, and behavioral detection of pure tones. These measurements included baseline auditory sensitivity prior to injury in all mice, and again at 3, 30, and 90 days after the blast in the two physiological groups, and daily for up to 90 days in the behavioral group. Mice in all groups experienced an initial deterioration in auditory sensitivity, though physiological measurements showed evidence of recovery that behavioral measurements did not. Amplitudes and latencies of ABR waves may reflect additional changes beyond the peripheral damage shown by the threshold changes and should be explored further. The present work addresses a major gap in the current acoustic trauma literature both in terms of comparing physiological and behavioral methods, as well as measuring the time course of recovery.

1. Introduction

Blasts, like those from the detonation of improvised explosive devices (IEDs), lead to trauma to the brain generally and the auditory system specifically (Lew et al., 2007). This type of injury typically induces primary, secondary, and tertiary damage that includes concussions, hearing loss, tinnitus, and damage to musculoskeletal and soft tissues (Scherer and Schubert, 2009). There are many factors relating to blast exposures that can impact the severity of an individual’s injury outcome, such as their distance from the blast, use of protective devices, and their sensitivity prior to injury, to name just a few (Fausti et al., 2009). Because of the wide variability in injury acquisition, understanding the damage that is directly caused by a blast-induced trauma has been difficult. That is, blast wave exposures may affect both the peripheral and central auditory systems, and these results can be evaluated and disentangled using a variety of different auditory tests. Unfortunately, the number of veterans with service linked sensorineural hearing loss and tinnitus as a direct consequence of blasts has been on the rise, while the development of treatments has not.

Two common ways to physiologically assess hearing problems in veterans and others in audiology clinics include auditory brainstem responses (ABRs) and otoacoustic emissions (OAEs). ABRs track the response of the ear and the brain to sounds and reflect the health of the brainstem, whereas OAEs measure emissions generated in the ear as it responds to sound and reflect the health of the auditory periphery. These assessments are often done in combination with behavioral hearing tests such as the Yes/No paradigm to determine the hearing abilities of a patient, which can then be compared to average data from the general population.

Physiological methods like ABRs and OAEs are desirable for measuring auditory acuity in laboratory animals such as mice because of their relative ease to execute and collect data for multiple stimuli in succession. There are, however, large discrepancies between thresholds obtained from physiological methods in anesthetized subjects and awake and behaving measurements of hearing from many animals, including rodents (Dent et al., 2018). Behavioral thresholds from awake mice can be 5 to 30 dB more sensitive than physiological measurements from anesthetized mice, depending on the strain, age, sex, stimulus, test used, and other factors. Some of these discrepancies may be due to the anesthesia used in the physiological experiments, which could potentially suppress activity of the auditory system. Behavioral tests are conducted in awake and motivated subjects and reflect true perceptual sensitivity. Additionally, with relatively few healthy cells intact, awake and behaving mice are able to detect signals for which no discernible physiological response can be measured (Kobrina and Dent, 2019; Ohlemiller et al., 2010). Within the context of subjects exposed to a blast, this can have huge consequences in the evaluation of the severity of their injury and what course of treatment, if any, they may receive. The development of treatments relying solely on ABRs and DPOAEs without considering the behavioral sensitivity of those subjects could lead to an overestimation of the trauma and a perceived treatment effect when there is none.

The goal of these experiments was to examine changes in ABR and distortion product otoacoustic emission (DPOAE) derived thresholds in anesthetized mice prior to and following blast trauma. Additionally, behavioral thresholds for corresponding tone frequencies were measured in different (awake) mouse subjects to compare the effects of blasts within and across methods. This study is the first of its kind to measure longitudinal data, up to 90 days after a traumatic blast injury, in male and female CBA/CaJ mice using both electrophysiology and behavior. Results show that blast exposures lead to changes in hearing sensitivity that continue to worsen over a 90-day period relative to sham-injured control mice.

2. Materials & methods

2.1. Subjects

Fifty-seven adult CBA/CaJ laboratory mice (Mus musculus) (28 M and 29 F) were used in these experiments. The mice were approximately 426 days old (+/− 36.89 SEM) at the beginning of the experiment. Of the 57 mice, 28 (14 M and 14 F) were tested only in the ABR and DPOAE experiments, and 29 mice (14 M and 15 F) were tested only in the behavioral task. Physiological tests were conducted in anesthetized mice while behavioral tests were conducted in awake mice. As such, different test groups were used for physiology and behavior due to the strain that repeatedly anesthetizing subjects puts on the breathing rates of mice before and after a blast, and due to the extensive time it takes to train a mouse for the behavioral task.

Of the 28 mice tested in the ABR and DPOAE experiments, 22 were exposed to blasts, with 16 mice tested prior to blast exposures (baseline, 5-7 days prior to the blast exposure), and 3, 30, and 90 days after blast exposures. The other 6 mice were tested at baseline and 30 and 90 days after the blast exposure. Additionally, six no-injury physiology control mice (henceforth: sham-blast exposed) were tested at the same time points as the mice exposed to a blast. All behavioral mice were exposed to a blast. Each behavioral subject’s post-blast thresholds were compared to their own baseline thresholds due to variability in baseline thresholds across mouse subjects.

All mice were obtained from our breeding colony maintained at the University at Buffalo, State University of New York and were derived from the Jackson Laboratory’s laboratory mouse supply. Mice were individually housed in standard mouse cages (30 x 19 x 13 cm) with wood shaving bedding and had ad libitum access food. The mice used in the physiology experiments had ad libitum access to water, while the mice in the behavioral experiments were water restricted to approximately 90% of their free watering weights. The lights in the mouse holding room were on a reversed day/night cycle, with the lights turning on at 6 pm and turning off at 6 am. Mice were only tested in these experiments during the dark (awake) portion of this cycle. All procedures were approved by the University at Buffalo’s Institutional Animal Care and Use Committee and complied with the ARRIVE guidelines in accordance with the National Institutes of Health guide for the care and use of laboratory animals.

2.2. Stimuli

2.2.1. ABRs and DPOAEs

The stimuli for this task were clicks and pure tones of frequencies spanning the mouse’s normal hearing range (Radziwon et al., 2009). All stimuli were generated by the Tucker Davis Technologies’ (TDT) BioSig and SigGen software. The stimuli for the ABRs were single-channel monophasic clicks (from the TDT root click file) with a spectrum of 0-50 kHz, and 8, 16, 24, and 42 kHz pure tones (from the TDT root tone file). The clicks were 0.1 ms single channel monophasic signals presented at a rate of 21 presentations per second. While a 0.1 ms click would have less spectral energy above 10 kHz than below that frequency, this stimulus is routinely used in studies of mouse auditory brainstem responses and usually elicits a more robust signal than pure tones (e.g., Zhou et al., 2006). The 8, 16, 24, and 42 kHz tones were 5 ms single-channel cosine-squared gated tones presented 21 times per second. In each trial, a stimulus was presented from the speaker 512 times and a mean response wave was generated. The stimuli for the DPOAEs were paired pure tones used to collect DPOAE responses for 8 (7.27 and 8.72 kHz), 16 (14.54 and 17.44 kHz), and 32 (29.09 and 34.88 kHz) kHz. Due to the high frequency limitations of the ER10B+ (TDT) microphone, the tested frequencies in the DPOAE portion are not identical to those used in the ABR task. The paired tones were presented simultaneously and continuously from two different speakers to collect the DPOAE responses until 500 DPOAE responses could be measured and averaged at each intensity. The f2/f1 ratio was set at 1.2. The intensities of f1 and f2 (L1 and L2) varied from 90 to 10 dB SPL in 10 dB steps, with L1 and L2 both being presented at the same intensity to elicit a DPOAE response. The f2/f1 amplitudes and L1/L2 ratios were automatically computed by the TDT software and the DPOAE response of interest was the 2f1-f2 response that was recorded. The noise floor during testing ranged from 0 to −10 dB SPL if taking the absolute highest and lowest spurious noise, but generally was centered at 6 dB SPL and did not vary (+/− less than 1 dB SPL SEM). The DPOAE response needed to be at least 10 dB SPL above the noise floor to be considered present and within that 10-dB range of the noise floor was considered absent to avoid spurious noise as being considered signal. Based on these criteria for the presence and absence of signal, thresholds were determined at the intermediate L1/L2 intensity between the presence and absence of a DPOAE response. Stimuli were calibrated on test days using a TDT calibration microphone (PCB-378C01).

2.2.2. Operant conditioning

The 8, 16, 24, and 42 kHz tones were 500 ms in duration with 50 ms rise/fall ramps, generated in Adobe Audition (Version 8.1.0). Sound pressure levels for all stimuli were calibrated based on their peak-to-peak measurements using a custom Matlab script. An ultrasound recording system (Avisoft, Model USG 116-200) and a microphone (Avisoft Bioacoustics Ultra Sound Gate CM116) were placed at the location of the mouse’s head at the start of a trial to measure the intensity of the stimuli. Calibrations were conducted weekly to ensure that the intensity of the stimuli did not vary throughout the experiments.

2.3. Apparatus

2.3.1. ABR and DPOAE

The apparatus for ABRs and DPOAES was contained within the same sound attenuated booth but was reconfigured at the start of each test to collect the appropriate measurements.

Mice were placed on a heating pad kept at 37°C inside a small sound-attenuated chamber (interior dimensions 55 x 33 x 36 cm) lined with 4 cm thick Sonex sound-attenuating foam (Illbruck Inc., Minneapolis, MN) with a sheet of metal behind it to block electrical noise. The chamber was continuously illuminated by a small lamp with an 8-W white light bulb. The placement of electrodes and condition of the mouse during test sessions was monitored by an overhead web camera (Logitech QuickCam Pro, Model 4000). The test chamber contained one (for ABRs) or two (for DPOAEs) speakers (MF1) configured for open field tests in the ABR experiment, or closed field tests in the DPOAE experiment, with tapered tube tips 1/8” in diameter with PVC tubing cut to 5 cm, and two inline filters connecting to the speaker input and an RCA cable. DPOAEs also had an ER10B+ microphone tip to measure the response, while ABRs were recorded via needle electrodes (ELE-N) connected to a 4-channel preamplifier (RA4PA) and a 4-channel low impedance headstage (RA4LI). The A/D conversion in the RA4PA preamplifier was 25,000 samples/second, and was low pass filtered at 7.5 kHz. The RZ6 oversampled at 200 kHz and was further filtered at 300 and 3000 Hz. For DPOAEs, the speakers were placed 5 cm away from the subject with the PVC tubing on an unimpeded path to the stainless-steel tubes on the microphone probe. The microphone earpiece was tightly situated in the subject’s left or right ear (mice were randomly assigned to the left and right conditions at approximately equal probability). For ABRs, the speaker was placed 6 cm away in an unobstructed path from the mouse for sound to flow openly to the mouse.

The ABR and DPOAE experiments were controlled by a WS4 Windows computer running an Optibit interface on a TDT driver using BioSigRZ. Stimuli were generated by a Multi I/O Processor with an optic port (RZ6-A-P1) and sent to the speaker/s. Digitized data in both experiments were sent back to the RZ6 processor via the ER10B+ (DPOAE) or the preamplifier (ABR).

2.3.2. Operant conditioning

The mice were tested in one of four wire cages (23 x 39 x 15.5 cm), each placed in a sound attenuated Med-Associates chamber (53.5 x 54.5 x 57 cm). The chambers were lined with 4 cm thick Sonex sound-attenuating foam (Illbruck Inc., Minneapolis, MN). The chambers were continuously illuminated by a small lamp with an 8-W white light bulb, and the behavior of the mice during test sessions was monitored by an overhead web camera (Logitech QuickCam Pro, Model 4000). The test cage consisted of an electrostatic speaker (Tucker-Davis Technologies (TDT), Gainesville, FL, Model ES1), a response dipper (Med Associates Model ENV-302M-UP), and two nose poke holes surrounded by infrared sensors (Med Associates Model ENV-254).

The experiments were controlled by Dell Optiplex 580 computers operating TDT modules and software. Stimuli were sent through an RP2 signal processor, an SA1 power amplifier, a PA5 programmable attenuator, an ED1 electrostatic speaker driver, and finally to the speaker. Inputs to and outputs from the testing cages were controlled via RP2 and RX6 processors. Power supplies were used to drive the dippers (Elenco Precision, Wheeling, IL, Model XP-603) and infrared sensors (Elenco Precision, Model XP-650). Custom Matlab and TDT RPvds software programs were used to control the hardware.

2.4. Procedure

2.4.1. Blast exposures

All mice were placed into the blast wave generator described by Newman et al. (2015). Mice were anesthetized using a 50 mg/kg dose of ketamine mixed with 6 mg/kg of xylazine and secured in a 1 cm² wire-mesh cage (20 cm x 7 cm x 6 cm), positioned 2 inches from the mouth of the blast tube expansion chamber at zero degrees azimuth. Blast exposures were conducted inside a 6 x 6.5 ft double-walled sound attenuated booth (Acoustic Systems, Austin, TX) with the blast tube and pressure sensor, and air was supplied by shop air up to 85 psi (586 kPa gauge). The air supply provides filtration to remove traces of oil or particulate matter larger than 40 μm in diameter to prevent damage to the electronic flow control components of the control module. When the experiment was initiated with a user-chosen value of driver pressure (49-52 psi), a pressure transducer (PX182B, Model 100GI, Omega Engineering) sensed the air pressure in the driver section as it was filled with compressed air flow controlled by the electronic air pressure controller (IP610-X30, Omega Engineering). The driver section pressure was displayed by the process monitor (DPI32C24, Omega Engineering), which communicated with the MATLAB code through a serial port. The driver section was allowed to fill at the maximum flow rate until pressure reached 75% of the user-provided driver pressure, after which flow was reduced to 75% of the maximum value to prevent premature diaphragm rupture. Once the driver pressure reached the desired value, MATLAB communicated with the RPvdsEX software to stop airflow with a signal from an RP2 module (TDT) to the electronic air pressure controller. After a programmed delay, another signal was sent to a solenoid assembly attached to a hunting arrow residing within the driver section. The solenoid was energized and the arrow was driven forward, puncturing the brass foil diaphragm in a predictable location producing the blast wave, and returned to its original position by force of a spring. The RP2 module simultaneously coordinated data acquisition from the 137A23 pressure sensor (PCB Piezotronics, Depew, NY) for a period of 2 s at a sampling rate of 100 kHz, capturing important overpressure-related events during blast exposure. Buffers from different sensors were stored in the onboard memory of the RP2 during acquisition, after which the MATLAB code retrieved them for data analysis. For each blast exposure, a high-pressure blast probe (137A23 ICP Pressure Sensor, PCB) positioned at the mouth of the expansion chamber was used to measure the blast pressure level, and this number was recorded and included in analyses. The mean blast intensity was 183.65 dB SPL (+/− .94 SEM) (50.4 kPa). Fifty-one of the fifty-seven mice were exposed to these blasts, while six were exposed to a sham blast (anesthetized and placed into the blast chamber for 30 seconds - the amount of time it takes from start to finish for the chamber to fully pressurize and then detonate in a blast scenario).

2.4.2. ABRs and DPOAEs

Auditory brainstem responses (ABRs) were measured at the same time points, under the same sessions of anesthesia, and in the same chamber as the DPOAEs. DPOAEs were measured prior to ABRs for all mice due to findings that conducting ABRs prior to DPOAEs suppresses DPOAE responses (Mhatre et al., 2010). The DPOAE response is a cochlear emission propagated by the outer hair cells along the basilar membrane to the auditory canal, measured by the microphone as a response with its own amplitude that decreases as the intensity of the tone decreases. DPOAEs were collected 5 days prior to blast or sham exposures and again 3, 30, and 90 days after the blast. The procedures for the DPOAEs are similar to those reported in previous studies (Luebke et al., 2014; Martin et al., 2011; Willott et al., 2006). DPOAEs were measured for three frequencies (8, 16, and 32 kHz) across the mouse’s hearing range. The order of frequencies tested was always from low to high.

Mice remained in the same chamber for the ABRs and the DPOAEs, but the setup was reconfigured for the ABR procedure. Subcutaneous needle electrodes were placed on the bulla of alternating ears (the same ear used for DPOAEs), on the vertex of the skull, and a ground electrode was inserted into the opposite leg of the recorded bulla. The noise floor for each subject was obtained prior to testing by recording from the anesthetized mouse in the booth with no stimuli. After the noise floor was recorded, click thresholds were obtained to verify proper electrode placement and a clearly observable response. After the clicks were presented, each tone frequency was tested from the lowest to the highest frequency (8, 16, 24, and 42 kHz). Once all frequencies of tones were presented (which took approximately 25 minutes), mice were removed from the test chamber and placed on a heating pad until they regained consciousness before being returned to their home cage. The threshold value for any particular stimulus was determined as the intermediate value between an observable ABR or DPOAE response and when that response was no longer visible above the noise floor (e.g., waves present at 50 dB, waves absent at 40 dB, threshold = 45 dB).

2.4.3. Operant conditioning

Mice were trained on a go/no-go operant conditioning procedure with positive reinforcement to detect a single pure tone (8, 16, 24, or 42 kHz). The mice were tested for one hour a day, 6-7 days a week. Each mouse was only tested on one stimulus frequency, with intensities varying according to the psychophysical Method of Constant Stimuli (MOCS). The MOCS is a method of presenting signals both above and below the presumed threshold in a randomized order, and this method has been used previously to obtain behavioral thresholds in our laboratory (e.g., Radziwon et al., 2009; Kobrina & Dent, 2016; Screven & Dent, 2016). After a stable baseline threshold was calculated for each subject, the mouse was blast exposed in the TBI blast chamber. After the blast, thresholds were assessed daily through behavioral testing.

Before the mice began testing, they were trained to respond in the operant chamber for a liquid reinforcement of 0.01 ml of Ensure™. Ensure™ is a nutritional supplement with a consistency like chocolate milk, and it was used in an attempt to maximize the number of trials the mice would complete. The training period involved shaping the mouse’s behavior to encourage them to poke into the appropriate holes for reinforcement. The mouse was first trained to locate and drink from the dipper (see Figure 1 for diagram of apparatus). Then the mouse learned to poke their nose into the left poke hole (observation hole) one time to receive a reinforcement from the dipper. Lastly, the mouse was trained to poke twice in the left observation and once in the right report hole to receive a reinforcement. Once the mouse learned the behavioral procedure employed in the detection task, it began training with an acoustic stimulus. The mouse was trained to repeatedly poke to the observation hole until it heard a pure tone, after which they would nose poke to the report hole for the reinforcement. The training stimulus varied across mice to train subjects more rapidly in the tasks; all mice were trained and tested on their final test frequency. Next, catch trials were introduced into the training and the variable waiting interval, or the random time window before target onset, was systematically increased to occur any time between 1 and 4 s. The response window duration was maintained throughout training and testing at 2 s to keep the response to the report poke hole tightly time locked to the stimulus onset. Pokes were registered once a mouse broke the infrared beam with its nose, when initiating trials and when responding to stimuli.

Figure 1.

Schematic of the operant conditioning setup.

During the next phase of the operant experiment, the mouse was trained to report that they detected a high intensity pure tone target using strict criteria for their hit and false alarm rates. For the “go” portion of the training, where mice were expected to ‘go’ to the response poke hole when a stimulus was presented, the hit percentage for all targets in a session was measured (hits/(hits + misses)*100). A hit was recorded when a stimulus was presented from the speaker and the mouse reported that they heard it within the 2 s response window. A miss was recorded when a stimulus was presented from the speaker and the mouse failed to report that they heard the sound within the response window. For the “no-go” portion of the training, where a mouse was expected to ‘not go’ to the report poke hole and withhold that behavioral response in the absence of a stimulus, the false alarm rate for all sham (no signal presented) trials in a session was measured (correct rejections/(correct rejections + false alarms)*100). A correct rejection was recorded when a mouse initiated a trial but no stimulus was presented and they withheld their response. A false alarm was recorded when a mouse moved to the report hole during the 2 s response window. Sham trials were gradually increased throughout training from 0% to 30% of all trials to prepare the mouse for testing and then was maintained at 30% of all trials during testing. Sham trials were never reinforced.

Each mouse was tested on only one tone frequency. During testing, the mouse began a trial by nose poking through the observation nose-poke hole, which initiated a variable waiting interval ranging from 1 to 4 s. After the waiting interval, a single test stimulus (e.g., pure tone of either 8, 16, 24, or 42 kHz) was presented in 70% of all trials. Thirty percent of all trials were shams. When the mouse was able to detect the targets over 80% of the time correctly (hit percentage) with less than a 20% false alarm rate, two or three of the seven targets were attenuated. These targets were attenuated in 3-10 dB steps until the mouse’s performance consistently dropped below 50% for the stimulus with the lowest intensity. The target and sham stimuli were presented using the MOCS, where the pre-determined stimuli at intensities above and below the anticipated threshold were presented in a random order throughout a session with the sham trials. This guaranteed that the mouse received sufficient reinforcement to stay motivated during the session, and ensured that the mouse was under stimulus control. A d’ threshold was determined from at least 200 trials after the quietest stimulus was hit less than 50% of the time and all other stimuli were hit more than 50% of the time. This threshold represented the just noticeable difference for that stimulus. The threshold criterion was a d’ of 1.5, as has been used in many other psychophysical studies conducted in rodents (Cai and Dent, 2020; Kobrina and Dent, 2016; Neilans et al., 2014; Radziwon et al., 2009; Screven and Dent, 2016).

After the threshold was stable (within 5 dB) for several days (3-7 days depending on availability of the blast chamber), the mouse was anesthetized and blast exposed in the chamber. Following the blast, the mouse was given 24 hours to recover and then was returned to operant testing on a daily basis to evaluate thresholds over time. Mice completed between 100-400 trials per day in this experiment and daily thresholds were calculated using every mouse’s individual daily performance. Daily changes were monitored for 3 months following the blast to fully capture any differences in hearing from baseline, and to determine if there were any differences in recovery between physiological responses from anesthetized mice and behavioral responses from awake behaving mice.

2.5. Data analysis

DPOAE, ABR, and operant conditioning data sets were all analyzed separately. A comparison was then conducted between thresholds derived from ABRs and operant conditioning.

2.5.1. ABRs

The ABR data set was sorted by stimulus type (clicks, 8, 16, 24, and 42 kHz tones) and blast-exposure groups (blast exposed vs. sham-blasted). Absolute thresholds were examined to establish changes in ABR thresholds over time. Threshold shifts were then extracted from the absolute threshold data to show effects of the blasts, to compare effects of the blasts across methodologies (ABRs, DPOAEs, and behavioral responses), and across studies (Newman et al., 2015; Smith et al., 2020). Threshold shifts were calculated for every subject for each frequency per observation using the following formula: dB shift = baseline dB – post-blast or post-sham-blast dB. Threshold shifts were used as the main dependent variable in the following analyses and were calculated relative to baseline such that a positive value reflects better sensitivity than baseline and a negative value reflects deficit from baseline.

Given the longitudinal design of this experiment and the unbalanced group composition, we utilized a linear mixed-effects model to examine whether blast exposure and day post-exposure predicted shifts in ABR thresholds across stimuli (LMM, lmer in the lme4 R package) (Bates et al., 2014; R Core Team, 2017). In this model, we examined whether shifts in thresholds (dB) could be predicted by fixed factors of day post-exposure (baseline, 3, 30, and 90), stimulus (click, 8, 16, 24, and 42 kHz tones), blast-exposure group (blast exposed and sham-blast exposed), sex (male and female) and by interactions between day post-exposure, stimulus, and group. The model was compared to the intercept only model (i.e., a model without predictors) for significance. To control for dependencies within our data from sampling each mouse repeatedly, we included a random intercept for mouse identity, sex, and the ear chosen for testing across day post exposure. Post hoc comparisons using Tukey’s method were performed to assess the relationship between day post exposure, stimulus type, and blast-exposure group (emmeans R package) with p values adjusted to the number of family-based comparisons to reduce type 1 error.

2.5.2. DPOAEs

Similar to the ABR data set, threshold shifts were compared in these analyses. We utilized a linear mixed-effects model to examine whether blast-exposure group and day post-exposure predicted shifts in DPOAE thresholds across stimuli. The model was constructed using the same predictors as the ABR threshold shift data set. Post hoc comparisons using Tukey’s method were performed to assess the relationship between day post exposure, stimulus type, and blast-exposure group with p values adjusted to the number of family-based comparisons to reduce type 1 error.

2.5.3. Operant conditioning

Given the longitudinal design of this study, as well as the repeated sampling for each mouse, a linear mixed effects model was used to analyze the behavioral results. Threshold shifts (dB) were used as the dependent variable to examine whether the blast exposure, days after the blast, sex (male and female), and stimulus (8, 16, 24, and 42 kHz) predicted threshold shifts. In order to rule out the possibility that an uncharacteristic data point on any given day influenced our overall blast effect, a moving window average of the threshold shift across days was calculated. To compute this moving window average, the following formula was used: threshold shift today (dB) = average (today + (today +1) + (today +2) + (today + 3) + (today + 4) + (today + 5) + (today +6)). This smoothing technique has been used previously to reduce spurious day to day variability and show a more accurate reflection of the overall hearing sensitivity in animal subjects (Heinz et al., 2005). To limit the number of critical comparisons, to avoid an inflation of type I error, and to draw parallels between physiology and behavior, key days of baseline (henceforth −3), 3, 30, 60, and 90 days after the blasts were used in this model. The model was compared to the intercept only model for significance with mouse identity as a random intercept. Post hoc comparisons using Tukey’s method were performed to assess the relationships between days after blast exposure, stimulus, and sex with the p values adjusted to the number of family-based comparisons to reduce type I error.

A linear mixed effects model analysis on the false alarm rate before and after blast exposure was also constructed to see if incorrectly responding when no stimulus was present systematically varied with exposure. Mice had 3-7 data points of pre-blast baseline testing sessions and up to 90 sessions after blasts, leading to an extremely unbalanced design. As such, a linear model evaluating false alarm percentage across timepoints (before and after blast), and a post hoc comparison using Tukey’s method comparing these time points was conducted.

2.5.4. ABRs vs operant conditioning

To examine the differences between the effects of a blast on behavioral and physiological assessments of hearing, a model was constructed using raw threshold and threshold shift data sets separately. The comparison of raw thresholds was utilized to determine if behavioral and physiological hearing assessments yield different threshold values. In this model, data collection technique (ABR and operant conditioning), stimulus (8, 16, 24, and 42 kHz tones), and days (−3, 3, 30, 60, and 90 days after the blast) were used as predictor variables for changes in thresholds (dB) after the blasts. Additionally, a separate model was constructed to examine whether threshold shifts (dB) could be predicted using sex (male and female), stimulus (8, 16, 24, and 42 kHz tones), days (−3, 3, 30, 60, and 90 days after the blast), and data collection technique (ABR and operant conditioning). Again, these models were compared to intercept models and Tukey tests were used for post hoc comparisons.

3. Results

3.1. ABRs

The raw threshold data (Figure 2A) revealed that mice were most sensitive at baseline, with the greatest increase in thresholds (i.e., decrease in sensitivity) 3 days after the blast exposures, and showing a gradual recovery thereafter. Three days after exposures to the blasts, the ABR thresholds increased from the baseline by 15 – 26 dB across stimulus types. In contrast, the thresholds decreased by 2 – 6 dB for mice in the sham-blasted group 3 days after the sham blasts (Figure 2B). Thirty days later, thresholds improved by 6.5 dB for some frequencies in mice exposed to a blast. Further improvements (i.e., decreases in thresholds) occurred in blast exposed mice between 30 and 90 days with additional recoveries of up to 5 dB. See Figure 3 for the average waveform generated to click stimuli in sham and blast exposed mice and Figure 4 for the average waveform generated to each of the four frequencies of tones in sham and blast exposed mice.

Figure 2.

(A) Mean ABR absolute thresholds for blast exposed mice before and at three time points after blast exposures and (B) threshold shifts for blast exposed and sham blasted mice. N = 22 for all time points except 3 days after the blasts, where N = 16. Error bars represent between-subject standard errors.

Figure 3.

Mean waveforms for the 90 dB amplitude presentations of the ABR click stimuli in sham-blasted mice (left) and blast exposed mice (right). For blast exposed mice, N = 22 for baseline, 30 days post-blast, and 90 days after the blast, and N =16 for the 3 days post-blast condition. For sham-blasted mice, N = 6 at all time points.

Figure 4.

Mean waveforms for the 90 dB amplitude presentations of the ABR pure tone stimuli in sham blasted mice (left) and blast exposed mice (right). For blast exposed mice, N = 22 for baseline, 30 days post-blast, and 90 days after the blast, and N =16 for the 3 days post-blast condition. For sham-blasted mice, N = 6 at all time points.

The model that best explained the shifts in ABR thresholds used the expression: lmer((shift ~ day*blastgroup + blastgroup*stimulus + stimulus*day + (1 | Id)+(1 | ear)+(1 | sex)) in R. This expression considers the linear modeling (lmer) of a dependent variable of dB threshold shift (shift) as it is explained by interactions between day (day) and blast group (blastgroup), blast group and stimulus (stimulus), and stimulus and day, with random fixed effects of subject identity (Id), ear of threshold measurement (ear), and sex of the subject (sex). This formatting for linear modeling equations will be used for all subsequent expressions. The main effect of blast group was significant (F = 24.18, p < .0001), suggesting that overall there was a significant difference in threshold shifts between blast exposed and sham-blasted groups. The interaction between blast group and stimulus was also significant (F = 2.57, p = .037), suggesting that the relative change in threshold shifts within each stimulus was different in blast exposed mice. This interaction was not significant for sham-blasted mice (p > .05). No significant main effect of day or stimulus was present (p > .05), and no significant interaction between day and blast condition or day and stimulus was present (p > .05). See Table 1 for statistically significant post hoc Tukey comparisons (corrected p = .005).

Table 1.

ABR Blast Group*Stimulus

Group Contrast	Stimulus Contrast	p value
Blast vs Blast	Click – 42 kHz	.0035
	16 kHz – 24 kHz	.0012
	24 kHz - Click	.0039
Blast vs Sham	8 kHz	<.0001
	42 kHz	.0002

^*For simplicity only relevant statistically significant comparisons are represented by this table. Comparisons are within and between groups across stimuli for ABR threshold shifts (corrected p = .005).

In summary, the blast exposed mice experienced increases in thresholds relative to sham-blasted mice at the highest and lowest frequencies tested. In mice exposed to a blast, ABR responses for frequencies at the mouse’s peak sensitivity (16 and 24 kHz) recovered more than they did at the other frequencies tested over the course of 90 days after the blast exposures. Threshold shifts from sham-blast exposed mice did not significantly differ across any time points, as expected.

3.2. DPOAEs

The raw threshold data (Figure 5A) reveal that thresholds are the lowest at baseline, similar to the ABR results. An initial increase in thresholds occurred three days after the blasts, which slowly decreased over time at 8 kHz, but not the other frequencies, over the 90 days (Figure 5B). The model analyses were conducted on the threshold shift data to draw comparisons between ABRs and DPOAEs. Three days after exposure to the blasts, DPOAE thresholds increased from baseline by 11 – 28 dB, but decreased from .8 – 3 dB for sham-blasted mice. Thirty days after exposure to the blasts, thresholds decreased, but were still elevated relative to baseline. Further improvements occurred between 30 and 90 days, with additional recovery of up to 2 dB for blast exposed mice, while sham-blasted mice worsened by as much as 6 dB between baseline and 90 days.

Figure 5.

(A) Mean DPOAE thresholds over time for blast exposed mice and (B) mean threshold shifts for blast exposed and sham blast exposed mice over time. N = 22 for all time points except 3 days after the blast, where N = 16. Error bars represent between-subject standard errors of the mean and asterisks denote statistical significance (*=p < .01).

The model with the greatest fit for analyzing DPOAE threshold shifts was: lmer((shift ~ day*sham + sham*stimulus + stimulus*day + (1 | Id)+(1 | ear)+(1 | sex))). The main effect of blast group was significant (F = 8.40, p = .0059), suggesting that overall there was a significant difference in DPOAE threshold shifts between blast exposed and sham-blasted groups. The interaction between blast groups and stimulus was also significant (F = 3.34, p = .037), meaning that the relative change in threshold shifts within each blast group changed differently across the stimuli tested. A significant main effect of day (F = 5.78, p = .017) but not stimulus (p > .05) was present, and no significant interaction between day and blast group or day and stimulus was present (p > .05). See Table 2 for statistically significant post hoc Tukey comparisons (corrected p = .008). Overall, for blast exposed mice, threshold shifts at the three frequencies all significantly differed from each other, a finding that was not observed in the sham-blasted mice. Blast exposed and sham-blasted mice only differed from each other at 16 kHz.

Table 2.

DPOAE Group Comparisons

Group Contrast	Stimulus Contrast	p value
Blast vs Blast	8 kHz – 16 kHz	.0011
	16 kHz – 32 kHz	<.0001
Blast vs Sham	16 kHz	.0007

^*For simplicity only relevant statistically significant comparisons are represented by this table. Comparisons were conducted between blast exposed and sham-blasted mice for all stimuli, comparing within-blast conditions across stimuli (e.g., 8 kHz compared to 16 kHz for blast exposed mice), and across blast conditions within stimuli (e.g., comparing 32 kHz for blast exposed and sham-blasted mice) (corrected p = .008).

3.3. Operant conditioning

The daily smoothed threshold data for individual mice (Figure 6) reveals that thresholds are the lowest at baseline, similar to the ABR results. An initial increase in thresholds occurred immediately after blast exposure, which changed progressively over time at all four frequencies tested, but specifically improved at 16 kHz (Figure 7). After exposure to a blast, thresholds increased on average by 20 dB within 90 days. The thresholds for 8, 24, and 42 kHz stimuli permanently increased in most mice. The detection thresholds for 16 kHz pure tones shifted after the blasts and partially recovered over the 90-day window.

Figure 6.

Smoothed thresholds (dB) for individual mice (abbreviated in legend) across days for 8, 16, 24, and 42 kHz pure tones.

Figure 7.

Smoothed mean thresholds for all mice for pure tone stimuli across the testing period.

To evaluate our hypothesis that blasts would lead to changes in detection thresholds, we assessed the interaction between days, sex, and stimulus using the following model: lmer((smoothed threshold shift (dB) ~ sex*day*stimulus + (1| Id))). We found a significant main effect of day (F = 237.97, p < .0001), and interactions between sex and day (F = 30.48, p < .0001), day and stimulus (F = 34.29, p < .0001), and a three-way interaction between sex, day, and stimulus (F = 38.84, p < .0001) (Figure 8). All other comparisons were not statistically significant (p > .05). See Table 3 for significant post hoc Tukey comparisons (corrected p = .0013).

Figure 8.

Mean threshold shifts (dB) for 8, 16, 24, and 42 kHz pure tones across 90 days tested. Error bars represent standard errors of the mean and asterisks denote statistical significance (*=p < .01).

Table 3.

Behavior Comparisons

Sex	Stimulus	Day Contrasts	p value
Male	8	−3/3/30/60/90	<.0001
	24	−3/3/30/60/90	<.0001
	42	−3/3/30/60/90	<.0001
Female	16	−3/3/30/60/90	<.0001
	24	−3/3/30/60/90	<.0001

^*For simplicity only relevant statistically significant comparisons are represented in this table. Day comparisons are all contrasts (−3/3, −3/3, −3/30 and so on for all day by day comparisons). Post hoc comparisons were conducted for the three-way interaction between sex, day, and stimulus (corrected p = .0013).

Mean false alarm percentages before blast exposure were 9.8 +/− 2.28 SEM and after blast exposures were 16.2 +/− 1.99 SEM. To evaluate whether these false alarm rates were significantly different, we ran a linear mixed effects model predicting false alarm rate from timepoint using the model formula: lmer((False Alarm ~ timepoint + (1| Id))). We found a significant main effect of timepoint (F = 29.54, p < .0001). A Tukey’s comparison revealed that false alarm rates before and after blasts were significantly different (p < .0001).

3.4. ABRs vs behavior

Mean thresholds from the smoothed behavioral results functions were 2 - 50 dB lower than the ABR thresholds across pure tone stimuli (Figure 9). Further, the threshold shift data revealed that there was a higher degree of variability between mice in the behavioral thresholds than in the ABR thresholds (Figure 10). A linear model comparing raw (unsmoothed) thresholds obtained using electrophysiological methods (ABRs) and our behavioral technique (operant conditioning procedures) was utilized in order to compare these two assessments of auditory function after a blast exposure. To test the hypothesis that the specific data collection technique (ABRs and operant conditioning) has a significant impact on our understanding of blast exposure effects, we used the model: lmer((threshold (dB) ~ data collection technique*stimulus*day + (1 | Id))). We found a significant main effect of data collection technique (F = 45.54, p < .0001), stimulus (F = 43.80, p < .0001), and day (F = 91.12, p < .0001). We also found significant interactions between data collection technique and stimulus (F = 3.01, p = .038), data collection technique and day (F = 8.50, p = .004), stimulus and day (F = 4.79, p = .0008), and a significant three-way interaction between data collection technique, stimulus, and day (F = 9.34, p < .0001). See Table 4 for significant post hoc Tukey comparisons (corrected p = .0013).

Figure 9.

Mean thresholds derived from individual smoothed threshold functions across 90 days tested, measured using operant conditioning (left) and ABRs (right). Error bars represent between-subject standard errors of the mean and asterisks denote statistical significance (*=p < .01).

Figure 10.

Threshold shifts across 90 days tested for behavioral thresholds (left) and ABR thresholds (right). Error bars represent between-subject standard errors.

Table 4.

ABR vs Behavior Threshold Comparisons

Day Contrasts	Stimuli Contrasts	p value
−3/3/30/90		<.0001
	24 kHz	.009
	42 kHz	<.0001
−3/3/30/90	42 kHz	.0007

^*For simplicity only relevant statistically significant comparisons are represented in this table. Thresholds measured by ABR and by behavior are compared at different days and stimuli in the 3 way interaction between data collection technique stimulus and day. Day comparisons listed as −3/3/30/90 include all individual comparisons (e.g., −3/3, −3/3, −3, 30 and so on) and stimuli comparisons without day comparisons are significant across the two data collection methods irrespective of day (corrected p = .0013).

In order to compare the true effects of the blasts (day) across sex as they interacted with data collection techniques and stimuli, we used the following model: lmer((threshold shift (dB) ~ data collection technique*sex*stimulus*day + (1 | Id))). Significant main effects of stimulus (F = 6.45, p < .0001) and day (F = 85.38, p < .0001) emerged. Two-way interactions between data collection technique and stimulus (F = 5.61, p = .002), data collection technique and day (F = 4.22, p = .04), sex and day (F = 31.04, p < .0001), and stimulus and day (F = 3.08, p = .03) were significant. In addition, significant three-way interactions between data collection technique, stimulus, and day (F = 12.67, p < .0001), as well as between sex, stimulus, and day (F = 7.08, p < .0001) were also observed. Finally, a significant four-way interaction between data collection technique, sex, stimulus, and day (F = 4.90, p = .002) was present.

A post hoc analysis examining the four-way interaction between data collection technique, sex, stimulus, and days was conducted. The p value was adjusted for 64 comparisons (p = .0008 for significance). Several comparisons within ABR and behavioral thresholds were significantly different; however, when comparing across these methods with the four-way interaction term no significant differences were observed (p > .0008). No critical comparisons at any time points, for any frequencies, for either sex, across the data collection techniques significantly differed (p > .0008).

4. Discussion

Nearly 20% of active duty American soldiers in Iraq and Afghanistan have been exposed to traumatic blasts (Terrio et al., 2009; Warden, 2006; Warden et al., 2005). There are numerous damaging effects of these blasts on cognitive processing, mobility, and perception (Bogdanova and Verfaellie, 2012). Changes in hearing sensitivity, deterioration in speech perception, and tinnitus are major disabilities that service members develop as a result of traumatic blasts (Lew et al., 2007; Oleksiak et al., 2012). In order to develop diagnostic and treatment strategies for these conditions and to improve post-blast outcomes in service members, animal models must be developed. In these experiments, the CBA/CaJ mouse was used as a model for studying the effects of TBI on the auditory system. The effects of blasts on auditory brainstem function and cochlear health as well as the behavioral expression of hearing loss for pure tones was evaluated and comparisons between these methods were also quantified. Together, the results from these experiments show that auditory function deteriorates after a blast exposure in mice similarly to humans. Unlike humans, the mouse shows a long-lasting change in behavioral sensitivity to pure tones, indicative of peripheral damage. Veterans with traumatic blast injury have peripheral audiometric thresholds that recover after acute injury.

Although we suggest increases in thresholds to acoustic stimuli in mice to be caused by peripheral (middle and inner ear) damage, blast traumas in humans are likely caused by both peripheral and central (brainstem and brain) auditory system trauma, leading to decreased sensitivity to auditory stimuli (e.g., DePalma et al., 2005). Differences in the ABR and DPOAE waveforms (see supplemental tables 1-5 for preliminary analyses) between blast exposed and sham blast exposed mice suggest a need for anatomical and physiological studies to clarify the source of the trauma in blast exposed mice.

Physiological measurements of auditory sensitivity are common in both humans and animals because they are fast and require no subject training. These measures can also be obtained independently of cognitive functioning because one does not need to be awake to respond. The experimental design of several other physiological studies that examined hearing loss following brain injury were used here in a mouse blast-exposure model (Amanipour et al., 2016a, 2016b; Newman et al., 2015, Smith et al., 2020). In this study, mice experienced an initial shift in thresholds after the blast, that surprisingly, was on a trajectory for recovery across the entire 90-day period tested. This finding is striking in comparison to other studies that examine the effects of TBI on auditory function which found permanent changes to sensitivity (e.g., Amanipour et al., 2016a; Smith et al., 2020). These studies, however, only traced effects of trauma for 14 days. This recovery has implications for mouse and mammalian hearing and communication after blasts. The findings of this experiment highlight the importance of studying blast-exposure effects on auditory function using a longitudinal design.

This study is the first of its kind to examine longitudinal changes to auditory sensitivity in mice before and after blast trauma using both physiological and behavioral methodologies. Here, we opted to use both methods because there are known significant differences in sensitivity obtained by these two types of threshold collection mechanisms (e.g., Dent et al., 2018), and because both of these methods are used in humans for assessing auditory function. When animals are anesthetized, as is usually necessary to measure ABRs (see Schrode et al., 2017 for an exception), there is a suppression in the magnitude of the response of the auditory system to sounds of intensities which awake and behaving animals have been shown to detect. We opted for a longitudinal study to probe the longstanding belief in the laboratory that trauma to major structures involved in auditory perception, including hair cells, cochlear nerve fibers, and auditory nerve synapses are permanently and irreversibly affected within the first seven days of an insult (e.g., Kujawa and Liberman, 2009). Longitudinal studies do not typically exceed 14 days when investigating acoustic trauma, but especially so in the case of blasts (e.g., Smith et al., 2020). Here, we opted to test up to 3 months after the insult to understand, in greater detail, the trajectory of hearing loss and potential recovery following traumatic blasts.

4.1. ABRs

We first examined the effects of blast trauma on ABR thresholds. Overall findings reveal an initial decrease in sensitivity immediately following exposure to a blast, which then gradually improves over 90 days. The shapes of the ABR waves change and do not show a recovery back towards baseline in a similar way to thresholds. Complex analyses of the waveforms were not included due to the marked changes to wave 1 of the ABR waveforms, which subsequently affects all later stages of the ascending auditory pathway. Some preliminary work in humans with mild TBI suggests that the ABR amplitudes are decreased and the latencies are increased after injury (Kraus et al., 2016), which is observed here in the mouse ABR waveforms. This is further evidence that changes to the ABR waveforms, beyond threshold shifts, may be useful in the future as an injury marker. More work must be conducted in humans and animals to disentangle these influences and to determine if waveform feature changes could be used as a secondary injury detection/severity indicator.

4.2. DPOAEs

We next evaluated DPOAE thresholds to several tone frequencies spanning the normal audibility range of mice. Here, we found post-blast hearing deficits in the peak sensitivity range of hearing, not in the lower and higher frequency ranges. These changes may be limited due to a ceiling effect where pre-sham-blasted mice may already hear poorly at high frequencies and therefore a shift cannot be detected. The blast exposed mice may also experience a swift recovery of hearing at lower frequencies to sham blasted levels. Thus, our only conclusive finding in the case of the otoacoustic emissions is that the mid-frequency region of the cochlea where peak sensitivity (best hearing) is present is susceptible to blasts in a way that other parts of the cochlea are not. A more thorough investigation of several more locations along the cochlea could help to disentangle findings from low to middle to high frequency hearing in a way that we were unable to do, since we tested only three frequencies. Additionally, it is important to note that the tympanic membranes did not rupture, and are thus not responsible for the changes that were observed in the measured DPOAEs.

4.3. Operant conditioning

We also evaluated changes to auditory sensitivity in trained awake and behaving mice before and for 90 days after the blasts. We found a relationship between the sex of the animal, the stimulus frequency they were tested on, and the days after the blast. We saw a similar trend across all mice of an increase in thresholds from baseline after exposure to the blast. Mice that were tested on 16 kHz recovered from the blast trauma differently than mice tested on all other tone frequencies, which may be due to the resilience of males at this frequency of tones. That is, males may have had an increased susceptibility to the blast exposure relative to females at all frequencies except 16 kHz.

It is important to note that there were slight differences in the false alarm rates of mice before and after blast exposures (increasing from 9.8 to 16.2%). The false alarm rates of the mice were maintained at well below the 30 percent criteria, so we know that the animals were engaging in the task appropriately. However, it is still notable that there was a significant difference in the likelihood that mice would respond in the absence of a signal after their blast exposure. One possible reason for this could be auditory uncertainty, where mice become more unsure of whether they hear a sound after exposure to a blast, especially in instances where signals are being presented at intensities close to thresholds. One possible explanation for this auditory uncertainty is that the mice may have been experiencing tinnitus, similar to humans exposed to a blast injury. If mice experienced tinnitus after the blast, the increased uncertainty could be reflected in the higher false alarm rates after the blast. While this was not explicitly measured in our study, we believe that the animals remaining under stimulus control, as indicated by the low overall false alarm rates while participating in the operant experiments, is indicative that they were reliably responding to the test stimulus and not an internal percept of sound even if they were experiencing one. Future research should behaviorally investigate the development of tinnitus in animals after exposure to a blast.

4.4. ABRs vs behavior

Finally, our last critical comparison involved assessing differences in the auditory sensitivity of mice across behaviorally and physiologically derived thresholds. While major differences were observed between the raw ABR and behaviorally obtained thresholds (as expected based on previous findings, e.g., Dent et al., 2018), surprisingly, the threshold shifts did not differ significantly between results obtained through operant conditioning and ABRs. This finding is significant to the field, not only for blast research, but for other studies examining hearing loss that seek to reconcile the major differences between trained awake and behaving subjects and untrained anesthetized passive subjects (e.g., Kobrina et al., 2019). Here, the lack of difference between the threshold shifts obtained through ABRs and operant conditioning provide a promising look to the future for all kinds of acoustic trauma research. Researchers should consider examining changes in both the threshold (dB) and change to threshold (difference from baseline) as a more complete way of understanding the overall changes in sensitivity caused by acoustic trauma.

5. Conclusions

In this study, we explored the short- and long-term effects of exposure to a blast on the auditory system. The recovery of thresholds derived using ABR and DPOAE methods in mice may shed some light on why hearing issues such as understanding speech in noise go undetected in humans. While behaviorally derived thresholds were lower and permanently affected, the changes in thresholds after the blasts were similar between behavioral and physiological tests. Overall, the results here highlight the need for longer term measurements of auditory function following acoustic trauma in humans, and underscore the usefulness of utilizing an animal model for those time-consuming measurements.