Blast-induced tinnitus: Animal models

Tinnitus is a condition whereby the affected people develop phantom sounds in the absence of overt acoustic stimulation inside or outside the body. The overall most common tinnitus inducer is acoustic trauma, such as occupational noise (Axelsson and Sandh, 1985), while the most common cause of tinnitus among military personnel and veterans is blast exposure (Dougherty et al., 2013). To the point, studies have shown that blast-related trauma is the “signature injury” in the Iraq and Afghanistan war theaters; chronic tinnitus and hearing loss are the most common auditory-related co-morbidities (Lew et al., 2007), among which nearly 50% of blast-injured service members claim to have tinnitus symptoms (Cave et al., 2007). Thus, blast-induced tinnitus has become an increasingly significant health problem for military personnel and veterans (Taber et al., 2006; Henry et al., 2009). In addition, tinnitus does not solely present as unwanted sounds in the ear or head, but it is often accompanied by anxiety, annoyance, disturbed sleep patterns, irritability, and depression (Hesser et al., 2009; Hebert et al., 2011), creating a significant impact on daily life. Thus far, 3–4 × 10⁶ veterans suffer from tinnitus and up to 1 × 10⁶ seek clinical services. Tinnitus has been rated as the number one service-connected disability affecting military personnel, resulting in nearly two billion dollars in annual disability compensation (VBA, 2013). Despite the adverse health impact and socioeconomic consequences of tinnitus, effective treatments remain limited, per American Academy of Otolaryngology-Head and Neck Surgery (Tunkel et al., 2014), due to limited understanding of the underlying mechanisms. Therefore, animal models are needed to elucidate detailed neural mechanisms underlying blast-induced tinnitus and develop effective medical treatments.

To induce blast-related trauma, animals are subjected to high-pressure shock waves generated by a shock tube (Ouyang et al., 2017). Under anesthesia with isoflurane (0.75–1% in a 2:1 N₂O:O₂ gas mixture), each animal's left ear is exposed to one or more repetitive blast exposures, usually at 15 min intervals. Note, using multiple blasts is a more realistic representation of blast scenarios in actual war theaters. Most animal blasts are unilateral, so prior to exposure, the right ear is occluded with a Mack's earplug, filled with mineral oil, and its pinna is sutured shut. Each animal subject is then placed on supportive netting and secured with a locking device. The commonly used shock tube is custom-built with a maximum working pressure of 100 PSI (∼700 kPa) (ORA, Inc., see Fig. 1). This type of device consists of a pressure chamber connected to a 20′ L 12′ D hollow tube. Parametric changes in shock-wave duration can be made by changing the length of the high-pressure chamber and the peak pressure can be varied by changing the thickness of the Mylar membrane that separates the pressure chamber from the driven tube [Fig. 1(C)].

The experimental setup is configured with instruments to monitor the pressure waveform using piezoelectric sensors: one sensor is placed axial to the blast pressure source (137A22 Free-Field ICP Blast Pressure Senor, PCB Piezotronics), while the other is positioned perpendicular and threaded into the tubing to capture details of the induced pressure wave (1022A06 ICP Dynamic Pressure Sensor, PCB Piezotronics) [Fig. 1(D)]. An analog-to-digital data acquisition system (DASH 8HF, Astro-Med, Inc.) acquires and monitors the collected data. A high-speed video camera (HG100K, Kodak Co.), which captures up to 3000 frames/s, is placed near the open-ended tube to record the effects of the shockwave on the animal's orientation and movement, prior to, during, and after delivery of the pressure wave [Figs. 1(A) and 1(B)]. Figure 1(D) shows a typical blast wave (red color), also termed Friedlander waveform as it follows the Friedlander function (Stuhmiller, 1991), which illustrates the shockwave pressure changes {peaked at 22 PSI [198 dB peak sound pressure level (SPL)] here} that impacts an animal subject. The shock waveform resembles those generated by other blast apparatus although different blast intensities were used for different studies (Ewert et al., 2012; Cho et al., 2013; Liang et al., 2017). Our data showed that most robust and chronic tinnitus is induced at 22 PSI. The shock wave consists of an initial overpressure peak, a subsequent negative pressure phase when pressure waves past the sensor, and a slow recovery phase, and most of the energy is concentrated at frequencies below 5 kHz, with a peak level around 180 dB SPL. The curve in blue color shows the pressure changes in the driver chamber.

It is well known that air- and fluid-filled cavities within the body are highly susceptible to blast-induced trauma (Taber et al., 2006). The ear canals are in an ideal position to serve as the primary entry zone for airborne pressure waves that initiate the traumatic events leading to hearing loss and tinnitus. Because the ear canals are also passive acoustic resonators, they can amplify sound waves at the plane of the eardrum by as much as 20 dB in the 2.0 to 7.0 kHz range (Shaw, 1974), adding to the traumatic effect of blast exposures. While the air-filled conductive mechanism of the middle ear is particularly vulnerable to blasts (Ritenour et al., 2008), direct coupling of middle ear ossicles to the fluid-filled cavities of the inner ear can result in irreversible mechanical damage to delicate sensory structures within the inner ear, including: basilar membrane, inner and outer hair cells, supporting cells, and their homeostatic ionic environment. As a result, damage to the auditory periphery can trigger degeneration of the auditory nerve and initiate a cascade of reactions that induce neurobiochemical and reactive (neuroplastic) changes in auditory pathways, spurring tinnitus and other hearing-related disorders.

At the periphery, the pathophysiology of the auditory system has been well documented in both human and animal studies. Using animal models, blast exposure can cause perforation of tympanic membrane, dislocation or fracture of ossicular chain, and gross cochlear trauma, such as rupture of the basilar membrane (Cho et al., 2013; Liang et al., 2017), loss of inner and outer hair cells (Ewert et al., 2012; Cho et al., 2013), and stereociliary bundle disruption (Niwa et al., 2016), all resulting in elevated hearing thresholds. Interestingly, our animals did not typically experience tympanic membrane rupture. However, tympanic membrane rupture not only depends on blast pressure level, but also on how animal subjects' heads are positioned in relation to the direction of blast shockwaves. The animals in our studies were placed in a holding apparatus in a rostro-cephalic orientation toward the blast source with their heads exposed, and their ear canals were oriented perpendicular to the shock wave front. This may explain why significant tympanic membrane perforation occasionally happened (<1%) and healed within a period of four weeks. In certain cases, the cochlear ultrastructure is relatively intact or ABR hearing threshold is not elevated following blast trauma, yet the number of ribbon synapses between cochlear hair cells and auditory nerve fibers may be significantly reduced (Cho et al., 2013). Such reduction is often accompanied by degraded ABR wave I amplitude. This suggests that military personnel or veterans, who have hidden hearing loss although their hearing threshold is normal, may still experience tinnitus or hyperacusis (Niwa et al., 2016).

Although the effects of blast-trauma on peripheral auditory structures are fairly delineated, the impact on the central auditory system is not clearly deciphered. Very recently, Race and colleagues used a rat model to demonstrate severe and persistent central auditory processing deficits, following blast, accompanied by injury spanning from the cochlea to the cortex (Race et al., 2017). This suggests that blast-induced central auditory system dysfunction may complement the peripheral maladaptive neuroplasticity (Race et al., 2017). Furthermore, studies also showed that repetitive blasts induced brain regional and time-related changes in acetylcholinesterase activity (Valiyaveettil et al., 2012b), as well as significant alterations in multiple genes in the brain that are known to be involved in age- or noise-induced hearing impairment in mice (Valiyaveettil et al., 2012a).

Critically, the above studies did not introduce behavioral assessment of tinnitus to directly elucidate the mechanisms underlying tinnitus, yet these findings indirectly help understand the factors that may have potentially contributed to the etiology of blast-induced tinnitus. In order to specifically address the etiology of blast-induced tinnitus, there is a need to introduce measures of behavioral evidence of tinnitus.

It has been a challenging task to reliably measure the psychophysical characteristics of tinnitus manifestation using animal models. Over the years, numerous behavioral paradigms have been developed to achieve this goal by determining the perception and characteristics of tinnitus in animals. These behavioral methods are typically classified into two major types based on induction and measure of behavioral responses. The first method is unconditioned reflex-based gap-detection (Turner et al., 2006; Galazyuk and Hebert, 2015; Marks et al., 2018), in which detection of a silent gap during continuous background sound induces an unconditioned reduction in acoustic startle reflex magnitude; this effect is diminished when the silent gap is obscured by presumed tinnitus perception or comprised gap detection due to tinnitus development. The benefits of this method are that it does not require training and it is generally a short duration test. The second method is an operant-conditioning-based paradigm (Jastreboff et al., 1988; Bauer and Brozoski, 2001; Heffner and Harrington, 2002; Ruttiger et al., 2003; Guitton and Dudai, 2007; Kizawa et al., 2010; Heffner, 2011; Berger et al., 2013; Sederholm and Swedberg, 2013; Stolzberg et al., 2013; Pace et al., 2016). The benefits of the operant conditioning-based method include reduced stress from experimentation (Ruttiger et al., 2003), investigation of the laterality of tinnitus-like behavior (Heffner and Koay, 2005; Heffner, 2011), and simultaneous assessment of tinnitus pitch and chronicity at the individual animal level (Pace et al., 2016; Zuo et al., 2017).

Since its development (Turner et al., 2006), gap-detection has become widely used to assess behavioral evidence of tinnitus in rodents, since food/water deprivation, foot-shocking, and prolonged behavioral conditioning are not needed. This method has been used to assess both onset and lasting tinnitus behavior, following administration of a variety of tinnitus inducers, and it allows characterization of tinnitus pitch, duration, and diagnosis at the individual animal level (Galazyuk and Hebert, 2015). Our lab successfully used the gap-detection method to demonstrate both acute and lasting noise- and tonal-type tinnitus following noise exposure (Zhang et al., 2011a; Luo et al., 2012; Pace and Zhang, 2013), and we were able to do so in individual rats (Pace and Zhang, 2013). Thus, our lab used this method to measure blast-induced tinnitus in a number of studies (Mao et al., 2012; Luo et al., 2014a; Luo et al., 2017; Ouyang et al., 2017).

As shown in Fig. 2, a single 22 PSI blast induced early onset tinnitus across all frequency bands, which shifted toward high frequencies over time (Mao et al., 2012; Luo et al., 2014b). Specifically, the current blast exposure parameters induced tinnitus at the 14–16, 18–20, and 26–28 kHz frequency bands, one month after blast, and at the 26–28 kHz frequency band, three months after blast. The presence of middle frequency tinnitus in our current studies may result from using a higher blast level (22 PSI), compared to our previously used lower blast level (14 PSI) that only triggered temporary high-frequency tinnitus (Mao et al., 2012). Nevertheless, these findings are in line with clinical reports that tinnitus may occur immediately following acoustic trauma (Mrena et al., 2004) and that it tends to eventually shift to high frequency regions (Roberts et al., 2010). The shift of tinnitus frequencies over time illustrates the ongoing neuroplasticity underlying blast-induced tinnitus.

Regarding the gap detection paradigm, several studies report potential confounding factors including reduced startle reflex following acoustic trauma (Longenecker and Galazyuk, 2011; Lobarinas et al., 2015) and the possibility that gap-detection may not be impaired by tinnitus (Fournier and Hebert, 2013). However, the fact that animals' tinnitus, as measured by gap-detection, often mirrors that of human subjects supports the validity of this behavioral method. For example, bimodal sound and somatosensory stimulations have been demonstrated to reduce tinnitus in humans, and the same result has been shown in animals, using the gap detection paradigm (Marks et al., 2018). Furthermore, we have demonstrated in our lab, with the gap-detection method, that electrical stimulation of the auditory cortex (Zhang et al., 2011a) and dorsal cochlear nucleus (Luo et al., 2012) suppresses tinnitus in rats, which also parallels the findings in human studies (Soussi and Otto, 1994; De Ridder et al., 2006). Nonetheless, this method needs more rigorous validation in studies that focus on both animals and humans. With this aim in mind, we have recently developed an optimized conditioned-licking suppression behavioral paradigm, which effectively and efficiently screens for tinnitus behavior. It also allows the characterization of both acute and chronic tinnitus and frequency bandwidths of the phantom sensation, at the individual subject level (Pace et al., 2016). This optimized version will enable us to cross-validate the gap-detection behavioral paradigm and to study mechanisms underlying tinnitus in a more clinically relevant manner and facilitate therapeutic drug and medical device development.

Several lines of evidence indicate that patients with severe tinnitus often experience serious comorbidities such as decreased concentration, increased irritability, anxiety and sleep disorders, and even suicide (Lewis, 2002; Hebert et al., 2012). These neurological disorders are mostly attributed to the reactive aspects of the limbic system. Clinical studies have demonstrated that, in tinnitus patients, many limbic brain regions, such as the amygdala and hippocampus, undergo anatomical and pathophysiological changes (Lockwood et al., 1998; Landgrebe et al., 2009; Schmidt et al., 2013). However, behavioral, electrophysiological, and imaging evidence indicates that the relationship between tinnitus and limbic dysfunction is complex and variable (Stevens et al., 2007; Andersson et al., 2009; Crocetti et al., 2009; Hesser et al., 2009). Given this hazy understanding of how the limbic system is involved with blast-induced tinnitus etiology, there is a need for investigation with animal models.

The EPM test is often used to examine animals' anxiety level following a stressor, so we utilized this method to evaluate rats' anxiety at five weeks post-blast (two days after the last, week-five gap-detection test) with a 5-min trial on the EPM (Ouyang et al., 2017). First, a rat was placed in the center of an elevated maze in the shape of a plus sign to initiate a trial. Importantly, two arms of the maze had walls, regarded as “safe,” while the other two were open and represented an anxiogenic environment. The rat's movements on the maze were recorded with a camcorder and then analyzed offline with Ethovision XT version 6 software (Noldus Information Technology, Leesburg, VA). Interestingly, our blast study showed that there were no significant differences in total-arm entries between any groups, indicating that both groups had similar mobility. In addition, while the tinnitus(+) and tinnitus(−) rat groups spent less time in the open arms and committed less open-arm entries compared to the control group, indicative of elevated anxiety, these results did not reach significance. Only when tinnitus(+) and tinnitus(−) groups were combined into a “blasted” group did they demonstrate significantly less time spent in the open-arms, when compared to controls. This suggests that blast exposure, regardless of tinnitus presence, is linked to limbic-produced anxiety for at least five weeks after trauma (Ouyang et al., 2017). Alternatively, our earlier published noise exposure study in rats showed no significant increase in anxiety for tinnitus(+) or tinnitus(−) groups (Pace and Zhang, 2013), but when evaluating individual animals, the majority of rats that had tinnitus showed the highest anxiety levels. This is in line with another report showing no significant increase in anxiety in rats with noise-induced tinnitus (Zheng et al., 2011b). These findings are reminiscent of previous clinical studies in which only certain tinnitus subjects have significant anxiety (Zoger et al., 2006; Belli et al., 2008; Hesser et al., 2009). What we learned from these studies is (a) when compared to noise-induced trauma, blast-induced trauma may have overwhelmingly influenced the manifestation of tinnitus-linked anxiety; (b) not all animal subjects develop significant anxiety, as is true with human patients; and (c) finer behavioral modalities such as grooming microstructure, sucrose consumption, weight, and other measurements may provide superior sensitivity in examining tinnitus-related emotional distress.

In addition to anxiety and distress measures, there is a need to examine whether blast-induced tinnitus is associated with cognitive impairment. During the MWM test, an animal subject swims through a water tank and uses spatial cues to locate a hidden, underwater platform. If the animal takes longer to find the platform and spends less time in the platform area, then it is deemed to have impaired spatial learning and memory. The test also examines the escape latency and speed to find the target. Specifically, during the escape latency trials, an animal is gently lowered into the water facing the pool wall, at one of four random starting positions. It swims a total of 12 trials, which are divided into three blocks with four trials apiece. Thirty-minute breaks are given between each testing block. During the probe trial (one time), each animal is allowed to swim in the pool for 60 s after the platform is removed. Throughout the trials, a digital camcorder is used to record the swimming trajectories of each animal's movement. To minimize stress accompanying such a lengthy test and increase the accuracy of testing tinnitus-related cognition since it could fluctuate during longer testing periods, a one day MWM approach is used (Frick et al., 2000; Fraticelli-Torres et al., 2010; Vandevord et al., 2012; Pace and Zhang, 2013). Offline data are analyzed with Ethovision software to determine the escape latency and time spent in the target zone.

In the study by Ouyang and colleagues, at five weeks after a single blast exposure, all of the rats went through the MWM test procedure. However, no significant difference in escape latency and time spent in the target zone was found between tinnitus(+), tinnitus(−), and control groups (Ouyang et al., 2017). This indicates that the currently used single blast exposure does not significantly affect the rat's spatial learning and memory, under these conditions (Ouyang et al., 2017). These findings are consistent with our study using noise exposure (Pace and Zhang, 2013), and are supported by previous investigations where no spatial cognitive impairment was found in rats with noise-induced tinnitus (Zheng et al., 2011a). Interestingly, rats with noise-induced tinnitus can develop altered impulse control and social interactions (Zheng et al., 2011b; Zheng et al., 2011c). That is, not all tinnitus(+) animal subjects have cognitive impairment, which reflects the clinical situation, where cognitive dysfunction only occurs among certain tinnitus subjects (Stevens et al., 2007; Andersson et al., 2009). In summary, the main challenge is to develop methods to identify cognitive–emotional dysfunction in animals and to properly extrapolate the findings in animals to humans. Additionally, the variability in blast or noise exposure parameters, and the sensitivity and specificity of the methods for measuring tinnitus-related cognitive–emotional dysfunctions in animal subjects provokes further exploration in methodologies.

To elucidate the neural mechanisms underlying blast-induced tinnitus, investigators have used electrophysiological recordings and imaging to study activity changes in auditory brain structures: mostly in the dorsal cochlear nucleus (DCN), inferior colliculus (IC), and auditory cortex (AC). These studies show a correlative relationship between tinnitus and hyperactivity (increased spontaneous firing), bursting, neuro-synchrony, tonotopic map reorganization, as well as gain changes, which are also frequently studied in noise- or other insult-induced tinnitus.

To conduct in vivo electrophysiology recordings, a craniotomy is performed to expose the left DCN, right IC, and right AC, either individually or simultaneously. The DCN and IC are inserted with multi-channel recording electrode (4 × 4 and 1 × 16 or 2 × 16, 16/32-channel NeuroNexus probes). The primary AC is identified by stereotaxic coordinates, vascular landmarks (the anterior and posterior dorsoventral vessels), and physiological response properties to tone and noise bursts, and a 4 × 4, 16-channel NeuroNexus probe is inserted in the proper location. Prior to insertion, all the recording probes are dipped in 3% Di-I solution (Invitrogen) for post-mortem, histological verification of implanted locations, and they are then are secured to the skull with dental acrylate. During recordings, neural signals including spontaneous and sound-driven unit spikes and local field potentials are preamplified and bandpass filtered. The output is fed into a 40 bit neurophysiology base station (TDT System 3), controlled by an OpenEx software suite, and all the data are stored for offline processing. Frequency tuning curves are obtained to determine the tonotopic representation of the implanted electrodes.

The DCN, as the first relay station along the auditory axis, has been thought to contribute to and/or modulate tinnitus, following intense sound exposure or ototoxic insult (Kaltenbach and McCaslin, 1996; Brozoski et al., 2002; Manzoor et al., 2013). One day after blast exposure at 22 PSI, we found significant increases in spontaneous firing rates (SFR), at all frequency loci of the DCN, in tinnitus(+) rats (Fig. 3). The increased SFR was circumscribed to the middle and high frequency loci, approximately one month after blast exposure, but this hyperactivity was reduced three months after blast exposure (Fig. 3) (Luo et al., 2014b). The fact that the broad frequencies of tinnitus at one day post-blast shifted to a single frequency at one-month post-blast and further transitioned to lower SFR at broad frequencies at three months post-blast demonstrate continued neural plasticity over time, which is in line with another study (Mao et al., 2012). In addition, although SFR was higher in tinnitus(+) rats than tinnitus(−) subjects, one can see that the SFR spiked at one day after blast, decreased and stabilized at one month post-blast, and further decreased by three months after trauma. Comparatively speaking, noise-induced hyperactivity is usually broadly distributed across the DCN, at early time points (five and 14 days), and then subsequently honed to medial positions (30 and 180 days) (Kaltenbach et al., 2000). The explanation of this discrepancy may be attributed to the extreme intensity in the blast study causing drastically more trauma to the inner ear, injuring the entire frequency expansion, whereas noise exposure preferentially injures the high frequency region (Liberman and Beil, 1979; Patterson and Hamernik, 1997). Interestingly, the decreased SFR at three months post-blast was accompanied by decreased P1-N1 amplitude of ABR responses, suggesting that hearing loss may have compromised the neural machinery for hyperactivity. This indicates that hyperactivity may not necessarily be the only contributing factor for tinnitus generation.

Mechanistically, the increased SFR in the DCN at one-day post-blast may result from loss of peripheral input (Luo et al., 2014b), decreased inhibition (Middleton et al., 2011), reduction in inhibitory glycinergic synaptic transmission in the DCN (Wang et al., 2009; Richardson et al., 2012), and/or upregulation of AMPA receptors in the DCN (Whiting et al., 2009). Reduced peripheral input may cause loss of inhibition—“dis-inhibition”—or compensatory enhancement for the impaired stimulus-driven activity (Schaette and Kempter, 2006). At the cellular level, Kaltenbach and colleagues believe that, via the granule-cartwheel cell circuitry, reduced input from cartwheel cells may diminish inhibitory innervation onto the fusiform cells, elevating their firing rates (Kaltenbach et al., 2000). On the other hand, the reduced SFRs in the DCN of tinnitus(+) rats, at three-months post-blast, may be caused by the extreme high-pressure blast waves-disrupting the basilar membrane, destroying hair and supporting cells (Patterson and Hamernik, 1997; Hoffer et al., 2010), and injuring axons (Kallakuri et al., 2018).

The rationale for studying neural activity changes in the IC following blast trauma is that acoustic trauma may induce onset or lasting hyperactivity in the IC (Bauer et al., 2008; Robertson et al., 2013). To the point, Luo and colleagues showed that a single blast at 22 PSI induced a significant increase of SFRs across all frequency regions, one day after blast (Fig. 4) (Luo et al., 2014a). Specifically, tinnitus(+) rats had increased SFRs, compared to controls, at 2–4, 4–16, and 16–42 kHz. This hyperactivity was honed to 4–16 kHz, in tinnitus(+) and tinnitus(−), when compared to controls at one month post-blast, but only the difference between tinnitus(+) and control rats reached significance. At three months post-blast, increased SFRs occurred in tinnitus(+) and tinnitus(−) groups across all frequency regions, but there was only significance at 4–16 kHz. Again, there was no significant group-wide difference between tinnitus(+) and tinnitus(−) cohorts (Luo et al., 2014a). Generally, these findings are reminiscent of previous studies that showed noise-induced increases of SFRs (Ma and Young, 2006; Bauer et al., 2008). Evidence indicates that the increased SFRs in the IC may result from losses of hair cells from noise exposure (Liberman and Kiang, 1978) and blast exposure (Cho et al., 2013), or peripheral deafferentation-triggered disinhibition (Kaltenbach, 2011). The shift of hyper-SFRs from the middle frequency region, at one month post-blast exposure, to the middle and high frequency regions, at three months post-blast, suggests that hyperactivity in the IC may have become progressively less dependent on afferent input from the cochlea. Thus, adaptive and compensatory plasticity may shift from a lower brainstem level to the mid-brain (Knipper et al., 2013).

In addition to measuring SFRs, bursting activity manifests in a rapid series of spikes separated by quiescent and silent periods and represents an important neural transmission code. It is known to enhance the reliability of synaptic transmission and neuronal selectivity (Swadlow and Gusev, 2001) and to enrich functional values of cortical activity, especially regarding serial synchrony of spontaneous firings in the cortex (Roberts et al., 2010). Bursting also mediates different brain functions such as neural synchronization, neural coding, plasticity, and attention (Kepecs and Lisman, 2003). Regarding tinnitus, increased bursting activity has been found in the DCN (Kaltenbach, 2011; Wu et al., 2016), IC (Bauer et al., 2008), MGB (Llinas et al., 1999), and AC (Norena and Eggermont, 2003), and it has been suggested to be a neural correlate of tinnitus. Supporting this claim, Luo and colleagues found that a single blast at 22 PSI can induce a significant increase in bursting rates at the 2–4 and 4–16 kHz frequency regions, at one day post-blast, and increased bursting at 4–16 kHz, at one month post-blast, for both tinnitus(+) and tinnitus(−) groups (Fig. 5). Again, there were no significant differences between tinnitus(+) and tinnitus(−) rats. At three-months post-blast, increased bursting rates were found at all frequency regions for both tinnitus(+) and tinnitus(−) groups. However, significance was only found in within-group comparisons, and there was no significant difference between tinnitus(+) and tinnitus(−) rats. Although the mechanisms underlying acoustic trauma-induced bursting changes are unclear (Norena and Eggermont, 2003), firing with a bursting pattern could result from a disrupted balance between the excitatory and inhibitory neural responses (Eggermont and Roberts, 2004; Richardson et al., 2012). Meanwhile, bursting shifted from low to middle frequency regions at one day post-blast, to a middle frequency region at one month post-blast, and to an entire frequency region(s) at three months post-blast, illustrating active neuroplasticity involved. Taken together, the results suggest that, in addition to blast-induced tinnitus, blast-related traumatic brain injury (TBI) and its associated neuroplasticity (Elder and Cristian, 2009) may have also contributed to the aforementioned SFRs and bursting activity.

Similarly, noise-induced tinnitus has been associated with increased SFRs in the AC, at both acute (Norena et al., 2003) and chronic stages (Eggermont, 2000; Seki and Eggermont, 2003). Luo and colleagues examined SFRs and bursting activity, in the AC of rats with 22 PSI single blast-induced tinnitus (Figs. 6 and 7). At one day and one month post-blast, there were no significant SFR changes between rats with blast-induced tinnitus and controls. However, at three months post-blast, a significant increase in SFR was found to broadly distribute across all frequency regions, in tinnitus(+) rats, compared to tinnitus(−) or control rats (Fig. 6). Similar to SFR, there were no significant changes in the bursting rate, in any frequency regions, between rats with blast-induced tinnitus and controls, at day one and one month post-blast. However, three months following trauma, bursting rates at all frequency regions increased significantly in tinnitus(+) rats, compared to controls (Fig. 7). Mechanistically, the increased SFRs in rats with blast-induced tinnitus may result from weakened, local GABAergic neurotransmission in the cortex (Eggermont and Roberts, 2004); it is also possible that the hyperactivity generated in the lower auditory brainstem, resonated all the way through the cortex (Schaette and Kempter, 2012). For bursting, the increases could result from lost balance between excitation and inhibition in the AC (Krishnan and Plack, 2009), which may also be carried over from the lower auditory brainstem.

In addition to SFR and bursting, it has been claimed that reorganization of the tonotopic map, coupled with hyperactivity, may serve as the driving force for tinnitus (Muhlnickel et al., 1998; Eggermont, 2007). Mechanistically, tonotopic map studies may help reveal whether blast exposure diminishes surrounding inhibition, unmasks dormant synapses, and/or generates new axonal sprouting and new connections (Bartels et al., 2007). Masri and colleagues in our group have reported that a single blast at 22 PSI significantly distorts the tonotopic map, in the rat AC at three months post-blast trauma (Fig. 8) (Masri et al., 2018). Indeed, blast exposure is known to cause traumatic brain injury (TBI), as demonstrated by our previous MRI studies (Mao et al., 2012). This indicates that the induced distortion of the tonotopic map may partially result from blast-induced TBI. In addition, TBI is known to cause TNF-α expression (Bermpohl et al., 2007), which can enhance excitatory synapses and weaken inhibitory synapses (Beattie et al., 2002; Stellwagen and Malenka, 2006): thus affecting excitability changes in the brain, as well as increasing the incidence of tinnitus (Eggermont, 2007). Although behavioral evidence of tinnitus was not measured in the aforementioned study, a different experiment (Fig. 3) (Luo et al., 2014b) suggests that the blast-induced tonotopic map reorganization may have also mediated blast-induced tinnitus.

In summary, blast-induced tinnitus is accompanied by early onset of SFRs in the DCN and IC (day one or one month) (Luo et al., 2014a,b), delayed elevation of SFRs in the AC (three months) (Luo et al., 2017), and tonotopic map reorganization (Masri et al., 2018). Chronologically, acute tinnitus may possibly result from hyperactivity and increased bursting predominantly in the lower auditory brainstem, whereas chronic tinnitus may mainly be mediated by higher-level auditory structures. These features may be driven by early onset neuroplasticity transitions which can elevate SFRs, and pathological bursting may become entrained in neurons of a high-level auditory center. However, depending on the types of neural correlates measured, such as neurosynchrony and coherence (Eggermont and Tass, 2015), one may not rule out the possibility that the AC and other high-level centers in the brain could directly mediate acute and/or chronic tinnitus.

Opposite to the highly invasive nature of intraparenchymal electrophysiological recordings, the most advantageous aspect of using an imaging tool for investigations of neurological disorders is its inherent ability to measure activity with an external device. Following the early imaging with single photon emission computed tomography (SPECT) to study tinnitus (Shulman et al., 1995), functional magnetic resonance imaging (fMRI) (Melcher et al., 2009; Husain and Schmidt, 2014; Lanting et al., 2014; Leaver et al., 2016) and positron emission tomography (PET) studies (Lockwood et al., 1998; Plewnia et al., 2007; Lobarinas et al., 2008; Lanting et al., 2009) have shown elevated activation and neural connectivity in both auditory and non-auditory brain structures of tinnitus patients. Diffusion tensor imaging (DTI), commonly used for studying microstructural integrity of white matter, has revealed significant structural correlates of fiber tracking changes with tinnitus, in humans (Crippa et al., 2010; Gunbey et al., 2017). Recently, quantitative electroencephalography/low resolution brain electromagnetic tomography functional brain imaging (QEEG LORETA) are also used to electrophysiologically image tinnitus-related functional connectivity in the brain (Shulman and Goldstein, 2014). In animal studies, manganese-enhanced magnetic resonance imaging (MEMRI) has been used as an activity-dependent and paramagnetic contrast agent to study tinnitus-related hyperactivity (Brozoski et al., 2007; Holt et al., 2010; Cacace et al., 2014). Although imaging tools are frequently used to study blast-related TBI, limited work is available to address mechanisms underlying blast-induced tinnitus. Below are the recent imaging studies investigating neural correlates of blast-induced tinnitus.

In this study, we blasted rats at an average of 14 PSI, which is equivalent to a 180–194 dB peak sound pressure level (Mao et al., 2012). To better simulate the situation in the battlefield, no earplugs were used. Using gap detection acoustic startle reflex behavioral methodology, we found that the single blast induced early onset of tinnitus and central hearing impairment, across a broad frequency range. The frequency of the induced tinnitus later shifted to a higher pitch over time. For the DTI data, we measured axial diffusivity (AD) for axonal integrity, and fractional anisotropy (FA) and apparent diffusion coefficient (ADC) values to reflect microstructural changes. Usually, blast-induced axonal degeneration is revealed by decreased AD and FA values, myelin injury, and ischemia, which are associated with increased ADC (Sotak, 2002; Song et al., 2003).

Following blast, our DTI results showed no significant changes in the DCN or AC. Instead, we found significant changes in the IC and medial geniculate body (MGB). Specifically, there was a significant increase in AD and FA values in the IC, and an increase in the ADC values in the MGB. These changes persisted for about four weeks. The increased AD and anisotropy in the IC and MGB are suggestive of the altered axonal integrity and increased ADC represents compromised myelination. An explanation for the elevated AD and FA values in the IC following blast may be due to active compensatory neuroplasticity, in response to tinnitus, hearing loss, and/or TBI (Chang et al., 2004; Lutz et al., 2007). That is, neuroplastic changes, as a result of blast trauma, may have contributed to the presence of blast-induced tinnitus and hearing impairment (Mao et al., 2012).

Although significant changes were found in the IC and MGB, no significant changes were found in the DCN, AC, and amygdala, following blast (Mao et al., 2012). A possible explanation is that blast-induced neurophysiological changes may have evolved and recovered to a level that is not significantly detectible using DTI, or a much stronger compensatory recovery occurred at the brainstem and cortical levels (Concha et al., 2006; Sidaros et al., 2008), as we used 14 PSI for this experiment, as opposed to the often used 22 PSI in our studies (Mao et al., 2012). Functionally, blast induced behavioral evidence of tinnitus at 14 PSI, which is less severe than that induced by blast at 22 PSI, suggests that the above reported tinnitus-related hyperactivity and bursting increases seen in the AC may not always be identifiable by DTI imaging at the structural level.

Although DTI and other imaging tools are non-invasive, the spatial resolution is limited. To compensate for this drawback, we recently used MEMRI, with cellular resolution, to investigate the neural correlates of blast-induced tinnitus (Ouyang et al., 2017). The working mechanism of MEMRI is that manganese (Mn++) may serve as an activity-dependent paramagnetic contrast agent for magnetic dependent scanning devices (Cacace et al., 2014). Following an intraperitoneal injection and circulatory dissemination, Mn++ enters brain tissues via voltage-gated calcium channels of activated neurons and deposits in the neuropil (Silva et al., 2004; Pautler, 2006). In addition to the drastically enhanced resolution, the MEMRI data are minimally affected by the loud noise from MRI scanner, which can muddy auditory signals. This is because it takes several hours to fully accumulate Mn++ in the brain tissues following injection, so the short time duration of the subsequent MRI scanning session is not enough to alter the already established manganese deposit profile, effectively eliminating the effects of the noise from the scanner.

In this study, rats were blasted in their left ear at 14 PSI, and we found that eight of the 13 blasted rats developed tinnitus at five weeks post-blast (Ouyang et al., 2017). The rats with blast-induced tinnitus had a bilateral increase in Mn++ activity in both auditory and certain limbic structures, compared to age-matched controls. More specifically, compared to the control group, rats with blast induced tinnitus manifested an increase in Mn++ activity in the DCN, IC, and AC, bilaterally, and in the right ventral cochlear nucleus and left MGB (Fig. 9) (Ouyang et al., 2017). However, we did not find significant differences in Mn++ activity between tinnitus(+) and tinnitus(−) groups, suggesting that MEMRI does not have the sensitivity to explicitly separate tinnitus-related Mn++ activity from that caused by blast-trauma and hearing impairment. Nevertheless, the current MEMRI data demonstrated hyperactivity in several auditory centers, which are in line with those electrophysiological findings, as described in the above sessions and some fMRI studies in humans (Smits et al., 2007; Lanting et al., 2008). This supports the hypothesis that increased Mn++ activity in these auditory brain structures could represent blast-induced tinnitus-related neural activity. Although these animal subjects were subjected to unilateral blast exposure, Mn++ activity manifested bilaterally in the auditory brain structures. The bilateral Mn++ activity may be attributed to the fact that the IC and AC receive bilateral afferent input (Brunso-Bechtold et al., 1981; Moore et al., 1998). The bilateral increase in Mn++ activity seen in the DCNs may be due to cross-talk and/or broad compressive effects from the blast shock waves on the whole brain. Furthermore, cross-board disinhibition and enhanced excitatory processes may also be involved in the bilateral Mn++ activity, even though the blast insult occurred unilaterally (Brozoski et al., 2012b).

The motivation of studying limbic brain structures is that tinnitus is accompanied by a number of limbic dysfunctions such as increased anxiety, irritability, disturbed sleep patterns, and depression. It has been thought that limbic structures may mediate attentional, cognitive, and emotional processes, as well memory storage of the tinnitus signals (Kraus and Canlon, 2012).

Based on the above context, many limbic structures are worthy of investigations concerning the mechanisms of the reactive aspects of tinnitus. Recently, we targeted the amygdala, which is thought to be a pivotal structure linking tinnitus to stress, emotion, anxiety, and memory (Cacace, 2004; Hui et al., 2006). It responds to environmental disturbances, ranging from emotional processing of anxiety (Wu et al., 2007) and memory (Roozendaal et al., 2009) to the orchestration of body homeostasis (Aggleton, 1993) and sensorimotor gating (Decker et al., 1995). The amygdala forms circuitry connections with auditory structures, such as the AC, to mediate auditory fear conditioning, learning, and memory (Poremba and Gabriel, 1997). Evidence also indicates that the amygdala integrates inflammatory information to coordinate behavioral and autonomic responses (Engler et al., 2011). However, it is unclear how the amygdala and its interactions with other auditory structures influence tinnitus and associated limbic dysfunctions.

To investigate this unknown, simultaneous electrophysiological recordings were conducted in the amygdala and AC of rats with and without noise-induced tinnitus (10 kHz, 105 dB SPL, 3 h duration) (Zhang et al., 2016). We found that rats with noise-induced tinnitus manifested significant hyperactivity and neurosynchrony, when compared to the tinnitus(−) and control groups (Figs. 10 and 11). The results are consistent with the findings in a study using a model of salicylate-induced tinnitus (Chen et al., 2012). The enhanced neuronal firing in the amygdala may be the consequence of reduced GABAergic input from the inter-inhibitory neurons onto the principal neurons in the basolateral nucleus of the amygdala. The reduced inhibition in the amygdala could be derived from increased AMPA and NMDA activity, both originating in the MGB and AC (Doyere et al., 2003), which receive direct impact from noise trauma (Zhang et al., 2016). In addition to the enhanced activity within the amygdala, correlogram ratio values suggest that tinnitus is associated with increased amygdala-AC temporal coupling, especially in the high-frequency region (Zhang et al., 2016). This neuronal evidence demonstrates direct interactions between a limbic center with an auditory center, in the context of tinnitus. One wonders whether such findings may be obtained from using a blast-induced tinnitus model as well.

Using MEMRI, we found that, following blast exposure, rats with tinnitus and overall increased anxiety demonstrated increased Mn++ activity in the superficial/cortical-like amygdala (AMG_S), the deep/basolateral amygdala (AMG_D), and the nucleus accumbens core (NA_C), compared to the controls (Fig. 12) (Ouyang et al., 2017). Similar to our findings with the central auditory system, we did not find significant differences between tinnitus(+) and tinnitus(−) groups, although tinnitus(+) rats did tend to have slightly higher Mn++ activity. These results suggest that the increased Mn++ in limbic structures may have played a role in the reactive processes, responding to tinnitus development. Mechanistically, blast-related TBI is closely associated with fear-related neural activation in the amygdala (Pieper et al., 2019), as well as increased inflammation and apoptotic processes in the nucleus accumbens (Sajja et al., 2013). The nucleus accumbens has abundant connectivity with the auditory cortices (Maudoux et al., 2012) and undergoes tinnitus-related hyperactivity (Leaver et al., 2011). Thus, the behavioral evidence of blast-induced tinnitus may be the consequence of trauma-driven hyperactivity, reflected in elevated Mn++ activity in both the amygdala and nucleus accumbens. This is consistent with their involvement in fear-conditioning and aversive responses, respectively. The reason for not being able to reveal significant differences in Mn++ activity between tinnitus(+) and tinnitus(−) groups could be that blast exposure not only induces tinnitus but also hearing loss and TBI, all of which involve neuroplastic processes. One possible approach to increase the sensitivity of MEMRI, sufficient to differentiate Mn++ activity specifically related to tinnitus or to hearing loss, is to use different levels of blast exposure in order to create a tinnitus(+) alone (without hearing loss) group or a hearing loss alone (without tinnitus) group. Furthermore, detailed investigations are needed by using different blast parameters, quantifying peripheral and central histopathology, and separating blasted animals with and without cognitive–emotional impairment.

Although numerous drugs and compounds have been investigated, none of them have proved to be effective in treating tinnitus (Langguth and Elgoyhen, 2012). Below are recent studies targeting oxidative stress, nitric oxide (NO)-cyclic guanosine monophosphate (cGMP), and tumor-necrosis factor α (TNF-α) pathways, in order to treat blast-induced tinnitus.

Oxidative stress is known to be associated with a variety of acoustic trauma-related pathophysiological processes, including mitochondrial injury, activation of cell death pathways, and activation of mediators of inflammation (Fetoni et al., 2019), resulting in secondary injury to cochlear hair cells and supporting cells (Patterson and Hamernik, 1997). Thus, antioxidants may be used to block these signaling pathways to treat hearing-related disorders. Along this line of treatment development, Kopke and colleagues chose NHPN-1010, a combination of N-acetylcysteine (NAC) and 2,4-disulfonyl α-phenyl tertiary butyl nitrone (Kopke et al., 2019). The former may restore mitochondrial electron transfer, energy coupling capacity, and calcium uptake activity following brain trauma, thus attenuating inflammatory responses in the injured rat brain (Xiong et al., 1999; Chen et al., 2008). The latter is a free radical spin-trapping agent that inhibits iNOS, decreases glutamate excitotoxicity, mitigates brain injury-related tissue loss, and improves cognitive function following blast injury (Du et al., 2017). Using a rat model of repetitive blasts, this group has found that NHPN-1010 significantly reduced ABR threshold shifts and DPOAE level shifts, along with significant reduction in cochlear inner and outer hair cell losses over a period of up to 21 days monitored (Ewert et al., 2012). Recently, this group reported that NHPN-1010 reduced behavioral and neuropathological evidence of blast-induced tinnitus, including normalization of ABR wave V/I amplitude ratios, ribbon synapse densities, and expression patterns of Arc/Arg3.1 and other tinnitus-related biomarkers in the auditory pathway. The data suggests that this composite chemical compound may be an effective candidate for treating blast-induced tinnitus (Kopke et al., 2019).

Phosphodiesterase-5 (PDE-5) inhibitors possess neuroprotective potential by improving blood flow and reducing apoptosis (Zhang et al., 2006; Li et al., 2007). Also, it may prevent cGMP degradation and enhance the NO-cGMP signaling. Functionally, sildenafil—a PDE-5 inhibitor—treatment can facilitate hearing recovery in guinea pigs following noise trauma (Zhang et al., 2011b), and vardenafil—another PDE-5 inhibitor—significantly mitigates noise-induced hearing loss (Jaumann et al., 2012). Clinically relevant, adverse effects of sildenafil on hearing have also been demonstrated (Okuyucu et al., 2009). Nonetheless, whether sildenafil treatment may have therapeutic effects on blast-induced TBI and its associated tinnitus and hearing loss remains unclear.

To achieve this goal, our lab subjected rats to three consecutive blast exposures at 22 PSI under anesthesia and then orally administered daily, 10 mg/kg sildenafil for seven days (Mahmood et al., 2014). Our results demonstrated that sildenafil did not prevent blast-induced early onset of tinnitus and hearing impairment, but instead, sildenafil significantly reduced hearing threshold shift during the first week after blast trauma and suppressed high-frequency tinnitus during the three to six weeks following blast trauma (Fig. 13). In addition, sildenafil treatment was able to reduce hyperacusis-like behavior. In summary, sildenafil yielded a therapeutic effect on blast-induced tinnitus and audiological impairment but in a time-dependent manner (Mahmood et al., 2014). Thus, it is possible that other treatment regimens such as higher dosage or combination with an alternative treatment modality could yield higher efficacy.

In terms of time course, very little is known about the differences between the generator for acute tinnitus and that for chronic tinnitus. Although sildenafil did not suppress tinnitus immediately post-blast, it may have prevented the maladaptive plasticity that is responsible for long-lasting chronic tinnitus, starting from the third week of treatment (Mahmood et al., 2014). The early recovery of hearing impairment and delayed tinnitus suppression by sildenafil may result from its ability to improve blood flow to the cochlea and peripheral auditory system (Mahmood et al., 2014), as a PDE-5 inhibitor increases vasodilation and activates the NO-cGMP pathway (Charriaut-Marlangue et al., 2014; Olmestig et al., 2017). That is, the sildenafil-induced tinnitus suppression and hearing recovery may have received protective effects from activation of the protein kinase B pathway, endothelial nitric oxide synthase, and inhibition of apoptotic process that eventually improve neuronal cell survival and functional recovery (Noshita et al., 2002).

In addition to the traditional notion that TNF-α mainly mediates cellular inflammatory responses (Namas et al., 2009), acute and chronic TNF-α signaling is also involved in homeostatic plasticity in the brain (Eyre and Baune, 2012), which is recognized as a potential mechanism underlying tinnitus (Yang et al., 2011; Schaette and Kempter, 2012; Yang and Bao, 2013) and associated limbic dysfunction (Weinberg et al., 2013). Specifically, TNF-α may be activated by sensory deprivation (Kaneko et al., 2008), which subsequently strengthens excitatory synapses (Stellwagen and Malenka, 2006), weakens inhibitory synapses (Stellwagen and Malenka, 2006), and thus possibly contributes to tinnitus etiology (Yang et al., 2011; Brozoski et al., 2012a; Yang and Bao, 2013). Cohesive with this logic, blocking TNF-α has been reported to benefit autoimmune cochleovestibular disorders, including hearing loss, vertigo, aural fullness, and tinnitus (Rahman et al., 2001). This rationalizes the need to examine whether targeting TNF-α may produce therapeutic effects on blast-induced tinnitus.

Our group recently investigated the role of the TNF-α pathway in tinnitus and found that neuroinflammation is actively involved in noise-induced tinnitus by mediating synaptic balance in the mouse auditory cortex (Wang et al., 2019). Using our blast rodent model, we studied the hippocampus, as tinnitus patients often suffer from cognitive and memory impairments (Hallam et al., 2004), and the hippocampus is a frequently affected structure in TBI patients (Kotapka et al., 1994). It responds to auditory stimuli and communicates with auditory nuclei (Liberman et al., 2009) and plays a critical role in acquisition and retention of memory (Liberman et al., 2009). Furthermore, chronic hippocampus overexpression of TNF-α causes neuroinflammation in the forms of microglial activation and blood-brain barrier compromises (Weinberg et al., 2013).

In a recent study, adult mice were anesthetized and subjected to a single blast exposure at 22 PSI. We found that 3,6′-dithiothalidomide (dTT), a TNF-α inhibitor, significantly blocked blast-induced morphological changes in microglial cells and prevented inflammatory responses in the hippocampus (Wang et al., 2018). Electrophysiologically, dTT prevented blast-induced enhancement of excitatory synaptic transmission in the CA1 region, as measured by reduced frequency of the miniature excitatory postsynaptic currents (mEPSCs) (Fig. 14). In parallel, dTT also alleviated blast-induced reduction in inhibitory synaptic transmission in CA1 pyramidal neurons, as measured by the recovered frequency of blast-reduced miniature inhibitory postsynaptic currents (mIPSCs) (Fig. 15) (Wang et al., 2018). The mediation of cellular excitability, as demonstrated by these findings, suggest that TNF-α may play an important role in tinnitus etiology.

Taken together, drug development by targeting both auditory and limbic structures may be a promising approach for an effective treatment for blast-induced tinnitus.

Blast exposure not only damages the microstructures within the cochlea, but also impacts both auditory and non-auditory brain structures anatomically, biochemically, and neurophysiologically, which in turn affect the psychophysical properties of tinnitus and other related hearing disorders. Behaviorally, the available studies have consistently demonstrated that blast exposure induces tinnitus in rodents, and specifically, high-intensity blasts induce chronic tinnitus; whereas low-intensity blasts induce acute tinnitus. The tinnitus pitch tends to shift to high-frequencies over time, illustrating ongoing neuroplasticity following blast-trauma. Beyond auditory testing, limbic dysfunctions are mainly measured using methods to assess anxiety and cognitive impairment. No significant correlation between measures of blast-induced tinnitus and those of limbic dysfunctions appeared on the group level, although, individually, some tinnitus(+) animals tend to have strong limbic functional impairment. This is in line with the clinical findings that not all tinnitus patients suffer from anxiety and cognitive impairment. However, finer behavioral modalities such as grooming microstructure, sucrose consumption, weight, and other measurements may provide superior sensitivity in differentiating and specifying tinnitus-related emotional distress and other limbic dysfunctions. Figure 16 illustrates the blast-induced trauma and summarizes the overall results of blast-induced tinnitus.

Electrophysiologically, blast-induced tinnitus is accompanied by early onset of hyperactivity in the auditory brainstem and midbrain and delayed hyperactivity in the cortex. The delayed cortical hyperactivity may result from active neuroplasticity, as revealed by significant tonotopic map reorganization following blast impact. Looking further, acute tinnitus may be predominantly associated with hyperactivity and increased bursting in the lower auditory brainstem, whereas chronic tinnitus may be more mediated by higher-level auditory brain structures. The hyperactivity in the midbrain becomes progressively less dependent on afferent input from the cochlea, indicating that adaptive and compensatory plasticity may shift from a lower brainstem level to a higher-level structure, over time. However, it is possible that the initial onset tinnitus may also be mediated by cross-center interconnectivity. In addition, the decreased spontaneous firing rates in the DCN at three months post-blast were accompanied by decreased P1-N1 amplitude of ABR responses, suggesting that significant hearing loss may have compromised the neural machinery for hyperactivity. This implicates that hyperactivity may not necessarily be the only contributing factor for tinnitus generation. Furthermore, simultaneous electrophysiological recordings in the amygdala and AC showed hyperactivity and synchronous coupling within and between both structures in rats with tinnitus, demonstrating direct interactions between a limbic center with an auditory center. As tinnitus generation is believed to be a product of integrated global neural networks, through cortical connectivity with other structures, further pursuits in cross-center interconnectivity will speed up identification of neural signatures of blast-induced tinnitus and its related limbic dysfunctions and will also help find ways to better extrapolate scientific findings from animal models to explanations of clinical problems.

Considering imaging, DTI studies showed significant increases in axial diffusivity and fractional anisotropy values in the IC and the apparent diffusion coefficient values in the MGB, as a result of compensatory neuroplasticity in response to blast-induced tinnitus, hearing loss, and/or TBI. However, tinnitus-related hyperactivity and bursting increases, obtained electrophysiologically, may not always be identifiable by DTI imaging at the structural level. MEMRI studies demonstrated an increase in Mn++ activity in both auditory and limbic structures of rats with blast-induced tinnitus, which are generally consistent with electrophysiological findings. However, the failure to reveal significant differences in Mn++ activity between tinnitus(+) and tinnitus(−) groups may be because blast exposure induces tinnitus, hearing loss, and TBI, all of which undergo neuroplastic processes. Thus, there is a need to use different levels of blast exposures to generate tinnitus(+) alone and hearing loss alone groups to differentiate tinnitus-specific activity from hearing loss- or TBI-specific activity.

When comparing the mechanisms underlying noise-induced tinnitus with those underlying blast-induced tinnitus, one can see that both share many common features: mainly the overall increase in spontaneous firing rates, bursting, neurosynchrony, microstructural integrity revealed by DTI, Mn++ accumulation, and increased excitatory synaptic transmission. However, one can also see their differences, as blast-induced tinnitus is associated with larger variability in the data gleaned across subjects and in time course. For instance, the extreme physical force of the blast shockwaves may not only damage the cavity-filled cochlea and then hearing, but may directly cause TBI and induce drastic neuroplasticity by pressure compression (Povlishock et al., 1992; Elder and Cristian, 2009). That is, noise trauma-induced tinnitus mainly stems from direct peripheral deafferentation at the cochlea, and its etiology then develops along the ascending auditory pathways. The blast trauma-induced tinnitus, on the other hand, results from direct impact on both the peripheral and central auditory systems and the simultaneous TBI, which may account for the less-uniform neural manifestations of blast-induced tinnitus.

Therapeutically, efforts have been made to identify a drug or compound to effectively treat blast-induced tinnitus and its related limbic dysfunctions. The approach to target oxidative stress appears to be promising, as the chemical compounds are already FDA-approved and show a certain degree of efficacy in initial experiments. The studies targeting nitric oxide (NO)-cyclic guanosine monophosphate (cGMP) and tumor-necrosis factor α (TNF-α) pathways are merely a start. Despite the therapeutic effects on blast-induced tinnitus and hearing impairment that we found with sildenafil, this therapy may possess its limitations. Further work is needed to determine if sildenafil at a different regimen, such as varied dosage and administration prior to- vs post-blast impact, would yield greater and longer-lasting reduction of tinnitus and hearing impairment. Targeting the TNF-α pathway may not only help change the traditional view that it predominantly mediates the inflammatory process, but also broaden our understanding of its important role in many aspects of neural plasticity and blast-induced tinnitus. Ultimately, further pursuit of study by targeting the TNF-α pathway is a promising approach for development of an effective pharmacological tool for treating blast-induced tinnitus and its related limbic dysfunctions.

This work was supported by the Department of Defense (Grant Award Nos. W81XWH-11-2-0031 and W81XWH-15-1-0357). The author thanks MD/Ph.D. student Ethan Firestone for his assistance in the preparation of Fig. 16.

AC

Auditory cortex

ABR

Auditory brainstem response

AD

Axial diffusivity

ADC

Apparent diffusion coefficient

AMG

Amygdala

cGMP

Cyclic guanosine monophosphate

dTT

3,6′-Dithiothalidomide

DCN

Dorsal cochlear nucleus

DTI

Diffusion tensor imaging

EPM

Elevated plus maze

FA

Fractional anisotropy

IC

Inferior colliculus

MGB

Medial geniculate body

MRI

Magnetic resonance imaging

fMRI

Functional magnetic resonance imaging

mEPSC

Miniature excitatory postsynaptic current

mIPSC

Miniature inhibitory postsynaptic current

MWM

Morris water maze

MEMRI

Manganese-enhanced magnetic resonance imaging

NAc

Nucleus accumbens

NAC

n-Acetylcysteine

NO

Nitric oxide

PET

Positron emission tomography

PFC

Prefrontal cortex

PPI

Prepulse inhibition

SFR

Spontaneous firing rate

TBI

Traumatic brain injury

TNF-α

Tumor necrosis factor α