Traditionally, tinnitus was considered to be a pure otological disorder, but modern theories have challenged this and proposed that tinnitus is a group of illnesses with multiple physiological mechanisms.

The most common attribute of tinnitus is a constant phantom sound (either a pure tone, hissing or roaring) in the ears in the absence of the corresponding external auditory stimulus ^[1][2][51,52]. This phantom sound is more noticeable in silent environments and hence an array of studies has compared the brain activity in tinnitus patients and healthy controls in silent environments to investigate the abnormal brain activity linked to tinnitus using electrophysiological (e.g., EEG) or magnetoencephalographic (MEG) recordings. Abnormal spontaneous activity, such as elevated gamma oscillations ^[3][4][5][6][53,54,55,56] and reduced alpha ^[1][7][8][51,57,58] activity, has been reported in the auditory brain regions of tinnitus patients. Interestingly, a spectral examination of the brain activity in a salicylate-induced animal model of tinnitus also showed a significant decrease in alpha and an increase in gamma band activity in the auditory cortex ^[9][59], which is consistent with neuromagnetic recordings in humans with acute tinnitus, reported by Lorenz et al., 2009 ^[10][60]. Therefore, it was generally thought that synchronised gamma-band activity links sensory events into a single cohesive conscious perception ^[11][61], and the ongoing gamma-band activity in the auditory cortex is necessary for tinnitus to occur ^[4][54]. The theoretical foundation of these observations forms the basis of the thalamocortical dysrhythmia (TCD) model ^[12][62] and the synchronization by loss-of-inhibition model (SLIM) of tinnitus ^[13][63]. The TCD model identifies the establishment of increased spontaneous firing of thalamic fibres as a crucial component of the development of tinnitus ^[12][62]. In particular, due to the reduced excitatory sensory stimuli from the damaged inner ear to thalamic relay cells, the hyperpolarised cell membrane induces the relay cells to fire low-threshold calcium spike bursts in a slow-wave form ^[12][62]. The formation of this slow-wave rhythm in the cortical neurons is subsequently brought on by thalamocortical feedback loops, which are detectable as continuous delta activity on the scalp. Tinnitus patients exhibit a decrease in delta activity in the cortical area ^[14][64]. The SLIM model hypothesises that the rise in the gamma frequency range may potentially be caused by a reduction in lateral inhibition processes in the auditory cortex as a result of the reduced activation of inhibitory neurons such as those which are GABAergic ^[13][63]. Increased gamma activity could also be due to hearing loss-related reduction in sensory input, which in turn reduces alpha-mediated inhibition. This leads to heightened gamma activity, as alpha band activity ordinarily suppresses gamma activity and hence results in increased neuronal synchrony, which is thought to play a role in the cause of tinnitus ^[2][52]. Thus, this discrepancy between cortical suppression and excitation offers a theoretical explanation for the alpha-up, delta-down pattern commonly observed in the resting-state non-invasive magneto- and electroencephalography (M/EEG) data of tinnitus patients ^[15][65]. In summary, based on the TCD and SLIM models, hearing loss may cause tinnitus by interfering with the coherent oscillatory activity between the thalamus and cortex. Although the mechanism which connects hearing loss with tinnitus remains unclear, both are well-known and mutually related medical conditions ^[16][17][66,67].

In addition to the above, reduced alpha activity is correlated with a desynchronised neuronal network that is often linked with auditory attention ^[18][19][68,69]. However, the source of alpha activity and whether it is induced by inhibitory activity or some other network-generated factors, remains elusive. Given that only 10% of synchronously activated neurons can generate an amplitude that is around 10-fold greater than that of unsynchronized neurons ^[20][70], and that approximately 10–15% of cortical neurons are GABAergic ^[21][71], EEG alpha activity, in theory, could be due to periodic fluctuations in the activity of inhibitory neurons. Therefore, one can speculate that tinnitus may be associated with a problem in areas of the brain where irrelevant incoming information from sensory regions of the human brain is actively suppressed under normal conditions ^[22][23][72,73], i.e., the “noise cancellation” model.

As proposed by Rauschecker, et al. ^[24][74], in a non-tinnitus case, unwanted repetitive auditory information is cancelled out at the level of the thalamus (medial geniculate nucleus; MGN), whereas in tinnitus, the cancellation mechanism i.e., the auditory gating mechanism, is compromised, which causes the MGN to become uninhibited, and eventually results in the perception of tinnitus sound ^[24][25][74,75]. Anatomically, the auditory gating system is thought to involve the auditory cortex, thalamus, prefrontal cortex, and limbic and paralimbic areas. The thalamus is innervated by the serotonergic fibres originating from limbic and paralimbic structures ^[26][27][76,77] and plays a role in auditory-limbic interactions. It has been speculated that at least certain limbic areas work as a component of a feedback circuit to the auditory system that plays an “inhibitory gating role” for auditory perception ^[28][78]. To predictably adjust and enhance auditory performance, auditory gating may also act as an adaptive process to filter out unnecessary event-based or temporal information ^[29][79].

Indeed, mounting pieces of evidence over the years have suggested that tinnitus is associated with functional and structural alterations in brain regions involved in emotional regulation, presumably as a result of consciously or subconsciously evaluating the continuous noise negatively, and failing to acclimate to it ^[30][80]. Functional imaging studies in tinnitus-affected humans and animals demonstrated activation of regions, both inside and outside the auditory systems. To explain this, one of the proposed hypotheses is that the abnormal filtering of auditory information by the limbic regions may play an exclusive role in tinnitus generation/perception. In the late 1980s, a study proposed a cross-talk between tinnitus and the emotional state of patients by showing that treating the comorbid major depression may lessen tinnitus disabilities ^[31][81]. It was later recognised by other groups that the limbic system may be actively involved in modulating or perpetuating tinnitus ^[32][33][34][36,82,83]. For example, the limbic and primary auditory areas showed dynamic linkages between tinnitus-related abnormalities, highlighting the significance of auditory-limbic interactions in tinnitus ^[35][84]. Functionally, the limbic system is an area of the brain that controls learning, memory development and storage, and emotions ^[36][85]. The hippocampus, amygdala, basal ganglia, cingulate cortex, and subcallosal area, among other limbic structures, are thought to contribute to a generalised “distress circuit” that can be triggered by real or phantom stimuli related to auditory, nociceptive, or other sensory stimuli ^[37][38][86,87].

The hippocampus is situated medio-temporally in each hemisphere of the brain, and its role in mediating spatial memory and memory consolidation is substantially investigated and widely acknowledged ^{[39][40][41][42]}[88,89,90,91]. The hippocampus receives auditory information primarily through the entorhinal cortex and is responsible for the temporal processing of the information ^[43][44][92,93]. The hippocampus also projects directly from area CA1 to the auditory association cortex and even to the primary auditory cortex ^[45][94] and plays a role in language, music processing and the development of long-term auditory memories ^[46][47][95,96]. For example, auditory recognition tests show very poor results in patients with extensive bilateral hippocampal loss ^[48][97], and research on monkeys suggests that auditory cues may be involved in spatial memory formation mediated by the hippocampus ^[49][98]. Using resting-state functional Magnetic Resonance Imaging (MRI), several studies have reported structural alterations such as significant loss of grey matter in the hippocampus of noise-induced tinnitus patients ^[50][99]. In addition, hippocampal activity and cognitive changes in human patients were found to be linked with the unpleasantness of tinnitus ^[51][52][100,101]. These findings suggest anatomical and functional connections between the hippocampus and tinnitus perception and/or tinnitus-related distress. Genetic factors have recently been associated with severe tinnitus, such as an increased occurrence of rare variants in ANK2 and TSC2 synaptic genes, which exhibit high expression in the hippocampus and cortex ^[53][12]. Furthermore, one study reported an increased occurrence of specific genetic changes such as missense and large structural variants in regions of the genome that are highly conserved. They identified CACNA1E, NAV2 and TMEM132D as potential genes that may play a role in contributing to severe tinnitus ^[54][13]. However, additional research is necessary to establish a definitive genetic linkage between tinnitus and the hippocampus.

Interestingly, noise exposure and listening to music have also been shown to cause structural and molecular changes in the hippocampus. For example, noise-exposed animals exhibited elevated mitochondrial areas inside the hippocampal neurons ^[55][102], possibly due to the variations in synaptic transmission that trigger presynaptic mitochondria’s ultrastructural plasticity to respond to the metabolic demand. Noise-exposed animals also exhibited a significant reduction in the optical density of Nissl bodies ^[56][103], a reduced postsynaptic density and outspread synaptic clefts ^[57][104] in the hippocampus, indicative of impaired synaptic transmission and neural activity.

Overall, these studies suggest that hearing loss produces a negative impact on hippocampal anatomy and function. However, the link between tinnitus and hippocampal function, especially the relationship between tinnitus and hippocampal synaptic plasticity, remains to be elucidated.

The basal ganglia consist of a group of subcortical nuclei, such as the striatum, subthalamic nuclei, substantia nigra and globus pallidus. They are situated at the top of the midbrain and the base of the forebrain and are primarily involved in controlling voluntary movements such as eye movements, aiding in balance, and supporting posture ^[58][105]. In the last two decades, the striatum (a nucleus present in the basal ganglia, critical for motor and reward systems) was found to not only mediate sensory information transmission to the cerebral cortex but may also play a role in tinnitus ^[59][60][106,107]. Studies in monkey ^[61][62][108,109], cat ^[63][110], rat ^[64][111] and human ^[65][112] demonstrate a functional and anatomical link between the caudate nucleus and auditory cortex. A case report of tinnitus cessation following a cerebrovascular accident that lesioned both the caudate nucleus and putamen (dorsal striatum) revealed the clinical viability of striatal neuromodulation to reduce tinnitus ^[66][113]. Over the years, a few studies have reported that deep brain stimulation (DBS) of the striatum can significantly suppress symptoms in patients suffering from tinnitus ^[67][68][69][114,115,116]. DBS of the striatum has also been tested in an animal model of tinnitus, where electrical stimulation of the caudate nucleus attenuated cluster neuronal firing in the auditory cortex and suppressed tinnitus-like behaviour ^[70][117]. Overall, the caudate nucleus may be involved in tinnitus; however, the mechanisms through which it regulates the disorder remains poorly understood.

The basal ganglia also have a limbic sector which consists of the ventral pallidum, nucleus accumbens (NAcc) and ventral tegmental area (VTA) ^[71][72][118,119]. The NAcc is strongly associated with addiction and depression ^[73][74][75][120,121,122]. The projection from the anterior cingulate cortex to the NAcc-VTA region has been found to mediate the effects of unpleasant sound ^[76][123], music ^[77][124], as well as tinnitus ^[78][125]. According to high-resolution magnetic resonance imaging studies in the brain, tinnitus patients exhibit structural and functional abnormalities such as a decrease of gray matter in the NAcc ^[38][78][87,125]. Recent research has also demonstrated that the severity of tinnitus and hyperacusis is linked to aberrant neuronal excitability in the NAcc, which results in emotional alterations to a sound stimulus ^[79][80][81][126,127,128]. Therefore, the NAcc has been suggested to be implicated in tinnitus and may contribute to changes in the limbic-auditory connections ^[82][129]. However, it remains unclear how the NAcc might be involved in the perception of tinnitus.

One critical structure of the basal ganglia, the subthalamic nucleus (STN), which is not directly linked to the auditory system, but is connected to the NAcc, has also been found to play a role in tinnitus ^[83][130]. For example, DBS in the STN of a tinnitus patient significantly improved tinnitus handicap inventory (THI) scores compared to the situation prior to DBS ^[68][115], suggesting that DBS of the STN may have a beneficial effect in the treatment of tinnitus. All of this evidence indicates the possibility that the STN contributes to tinnitus mechanisms, however, the exact nature of this connection still requires further investigation.

The amygdala is a key mediator of the emotional and other behavioural responses to sensory stimuli, across all senses, and is connected to the limbic, executive, and other sensory areas of the forebrain ^[84][131]. Anatomical investigations in different animal species have shown relatively consistent networks of projections to the nuclear complex of the amygdala from auditory and auditory-associated regions ^{[63][85][86][87][88]}[110,132,133,134,135]. Importantly, there is a second pathway from the basolateral amygdala to the principal inhibitory nucleus in the thalamus, the thalamic reticular nucleus (TRN) ^[89][136]. The TRN consists of a layer of inhibitory GABAergic neurons present between the thalamus and neocortex and produces inhibition of the sensory thalamocortical relay neurons ^[90][137]. The cortex and thalamus simultaneously send excitatory collaterals to the TRN ^[91][138]. These connections put the TRN in the role of a gatekeeper, regulating the flow of sensory information from the thalamus to the cortex by evaluating sensory stimuli based on their behavioural relevance ^[92][93][94][139,140,141]. Using a computational model of thalamocortical circuitry, it was found that the stimulation of the basolateral amygdala inputs to the TRN results in decreased spontaneous thalamic activity ^[95][142]. Thus, the amygdala may play an important role in regulating auditory gating functions. Furthermore, resting-state fMRI investigations in tinnitus patients demonstrated an abnormal functional connection between the auditory cortex and the amygdala ^[96][97][143,144], which may be related to tinnitus-related distress given that the auditory cortex projection to the amygdala play a potential role in mediating auditory fear conditioning, and hence connects emotion with tinnitus perception ^[24][98][74,145]. Interestingly, a study reported that salicylate administration into the amygdala can significantly enhance sound-evoked local field potentials in the auditory cortex, changes indicating heightened perception and emotional salience of the tinnitus ^[99][146]. Thus, one can speculate that the functional connections between the “amygdala-TRN-auditory cortex” may play a potential role in auditory gating mechanisms, that may further contribute to tinnitus perception and tinnitus-related distress.

The significance of emotional memories is a major aspect of chronic tinnitus. Emotional memories contribute to chronic tinnitus by causing a sustained level of hypervigilance, thereby promoting a continuous level of awareness ^[100][101][147,148]. Sound-induced amygdala responses are found to be sensitive to their emotional strength ^[102][149] and align with the significance of sound in an individual’s sensory environment ^[103][150]. Overall, one possible theory could be that the amygdala may deliver a significant negative emotional signal to the auditory cortex, influencing the perception of acoustic information ^[104][151]. However, there is a lack of understanding of the mechanisms involved and investigating the role of the amygdala in mediating tinnitus induction or consolidation following noise exposure may offer a better understanding of the condition.

The cingulate cortex is another part of the limbic system and is primarily involved in emotional responsivity ^[105][106][152,153], emotional processing and inhibitory control ^[107][108][154,155]. Interestingly, tinnitus discomfort has been associated with increased activity in the cingulate cortex ^[109][110][156,157]. Reportedly, manipulations that induce tinnitus resulted in increased fos-like immunoreactivity and Arc protein expression in the cingulate cortex of gerbils ^[111][112][158,159]. In contrast, one study reported no changes in the neuronal excitability or frequency response in the cingulate cortex after tinnitus induction with salicylate ^[113][160]. Furthermore, the subcallosal area (medial prefrontal cortex, orbitofrontal cortex, and anterior cingulate areas) is an important hub linking limbic-affective systems with thalamocortical perceptual systems. Using anatomical MRI ^[78][125], significant grey matter volume reductions in the subcallosal region were found in tinnitus patients when compared to the controls. The subcallosal areas, such as the ventromedial prefrontal cortex (vmPFC) and NAcc, demonstrated a crucial function in the long-term habituation to persistently unpleasant noises by sending feedback projections to the TRN, which in turn selectively inhibits the MGN regions corresponding to the unpleasant sound frequencies ^[114][161]. Interestingly, a significant volume loss in the subcallosal area of tinnitus patients has been reported ^[50][78][99,125]. The engagement of the subcallosal region, for example, is also modified by pain anticipation and perception and responds to the unpleasant effects of discordant music to variable degrees ^[115][162].

Overall, the limbic system occupies a critical role in elucidating the underlying molecular causes of tinnitus, as tinnitus-related volume loss in the limbic regions such as the hippocampus, amygdala, and subcallosal area is often due to atrophy of neurons and glial cells, leading to impairment of synaptic plasticity mechanisms ^[116][163]. Thus, the limbic system is well placed to play a crucial role in mediating tinnitus sounds from being perceived and targeting and suppressing tinnitus signals at these subcortical levels before they reach the primary auditory cortex may open new horizons for tinnitus treatments.