A plethora of factors contributing to frailty are associated with the heterogeneity of tinnitus disorder. Genetic factors, early development, and early- and midlife adverse experiences, such as socioeconomic disadvantages, abuse, divorce, and job loss, contribute to age-related allostatic load increases, which are associated with frailty (43–45), pain (46), personality change (47), and chronic diseases, e.g., depression (48), cardiovascular disease (49), and hearing loss (50, 51). Allostatic load resulting in frailty and other health disorders might contribute to the auditory and non-auditory components of tinnitus disorder and the impairment of the physiological, cognitive–emotional, and behavioral response systems. Although there is no direct relationship between the allostatic load and tinnitus disorder, chronic stress, including genetic factors and genetic correlations with hearing loss and depression (52), socioeconomic status and long-term stress exposure (2, 9, 53–55), are directly related to tinnitus disorder.

Other demographic and social factors, such as advanced age, loneliness, maladaptive personality traits, and living alone, which increase the frailty risk, are also linked to age-related hearing loss and chronic tinnitus (41, 56, 57). Lifestyle factors, including diet, smoking, alcohol intake, sleep hygiene, and physical inactivity (58, 59), are associated with tinnitus disorder. Later-life risk factors for frailty, such as multimorbidity (4, 60), chronic diseases (61, 62), polypharmacy (63), malnutrition (64), micronutrient deficits (65, 66), depressive symptoms (5, 7, 10, 18, 19, 25), and cognitive impairment (13–15, 25, 26), are also linked to the onset and progression of tinnitus disorder. Among these risk factors, biological aging and psychosocial stress might be the main risk factors.

Risk factors for frailty, including individual differences in vulnerability, such as genetics, early development, life-course experiences, personality traits, lifestyles and behaviors, and other physical, psychosocial and environmental challenges, interact and result in accelerated aging at the subcellular and cellular levels (37, 67). These include cellular senescence secretory phenotypes and mitochondrial dysfunction-induced oxidative stress and chronic systemic inflammation. Deregulated nutrient-sensing systems, including the insulin and IGF-1 signaling pathways for glucose sensing, the mammalian target of rapamycin (mTOR) pathway for the sensing of high amino acid levels, and AMP-activated protein kinase (AMPK) and sirtuins 1 and 3 for the sensing of low-energy states, also participate in the onset and development of frailty (37). Impaired hypothalamic–pituitary–adrenal (HPA) axis and sympathetic adrenal–medullary (SAM) axis function induce a decline in the response of the neuroendocrine, immune, and autonomic nervous systems to stress exposure (24, 67). Age-related hormonal changes, including dysregulation of thyroid, estrogenic, androgenic, growth and stress hormones also contribute to a decline in the reserve/resilience of multiple physiological systems (24, 37, 67). Complex physiological systems adapt to challenges or stress by dynamic allostasis. A decline in resilience and long-term adaptation results in frailty (Figure 1). Minor stressors, such as minor infection or surgery, might cause decompensation and maladaptation due to dysregulation of physiological systems and disproportionate changes in health status caused by allostatic load (24, 68). Different frailty phenotypes can lead to disability in different domains, including the physical, cognitive, and psychosocial domains. The vulnerability to different domains of frailty could explain why older subjects are more likely to present with disability, cognitive impairment, anxiety and depression associated with tinnitus than younger tinnitus patients do. The long-term failure of a single organ or system to adapt due to the allostatic load might cause pathological changes and chronic diseases, including hearing loss, chronic tinnitus and mood disorders (69). The pathological changes include neuronal atrophy, white matter lesions, demineralization, and dysfunction of the immune and metabolic systems.

The biological mechanisms of frailty onset and adverse outcomes also contribute to tinnitus disorder, including chronic inflammation and oxidative stress, mitochondrial dysfunction, dysregulation of the neuroendocrine and immune systems, HAP and SAM axes, and neurotransmitter activity (2, 24, 32, 70–72), are also involved in tinnitus. Tinnitus patients exhibit blunted reactive cortisol and dysregulation of negative feedback in the HPA axis (2, 24, 32, 73). Tinnitus can act as a stressor and results in adverse outcomes of frailty by inducing dysregulation of the HAP and SAM axes, including anxiety, depression, and sleep disturbances. Neuromodulation by stimulation of the vagal nerve or the auricular branch of the vagal nerve can alleviate not only neuroinflammation (72) but also tinnitus-related stress through improvements in parasympathetic activity and the balance of the autonomous nervous system (74). Peripheral inflammatory markers, such as significantly lower C-reactive protein (CRP) (75) and IL-10 levels (76) and higher oxidative stress indices and total oxidant statuses (77), have been reported in tinnitus patients than in healthy controls. Animal experiments have shown that sirtuin 3 can enhance the mitochondrial glutathione antioxidant defense system and prevent age-related hearing loss by sensing low-energy states (78).

Tinnitus disorder can interact bidirectionally with frailty. Tinnitus disorder might be aggravated by frailty, and tinnitus disorder induced by the aforementioned risk factors might lead to an allostatic load and maladaptation in older frail individuals. Tinnitus symptoms might be a stressor, accelerate the presence of adverse outcomes from frailty in different domains, increase the transition from tinnitus to tinnitus disorder, and increase the heterogeneity of tinnitus disorder (Figure 1). Therefore, frailty might cause heterogeneity in tinnitus disorder, and tinnitus or tinnitus disorder could become a stressor and lead to adverse outcomes in frail older individuals, including the worsening of tinnitus disorder.

Tinnitus disorder involves phantom percepts and accompanying psychopathological reactions due to the abnormal function of auditory and non-auditory networks (28, 29, 33, 34, 97). Tinnitus is processed by three anatomically separable but interacting pathways, including lateral and medial ascending pathways and descending noise-canceling pathways (98). The bottom-up hyperactivity of auditory pathways is the critical trigger factor of tinnitus, and the medial tinnitus pathway, overlapping with the salience network, contributes to the cognitive, psychological, and behavioral responses of patients with tinnitus disorder. The medial and lateral pathways are separable and are commonly balanced by the top-down noise-canceling pathway, and induce auditory sensation Tinnitus perception and the transition from tinnitus percept to tinnitus disorder depend on further processing of auditory stimulation by the abnormal synchronization of the central executive control network, the salience network, and the DMN. Therefore, some individuals may have tinnitus without suffering, and other individuals may suffer without tinnitus. Hearing loss in some individuals cannot induce tinnitus percept but leads to adverse structural changes in the medial ascending pathways, such as atrophy of the auditory cortex, frontotemporal regions, cingulate cortex, insula, and amygdala (56). The vulnerability of auditory and cognitive control and emotion processing circuits causes dysfunctional activation, including decreased cognitive reserve for other executive functions, due to increased support of the cognitive control network to effortful listening, abnormal emotion regulation and reactivity (56). Hearing loss also contributes to behavioral responses, including social isolation and loneliness, which aggravate dysfunction in cognitive and psychological domains. Thus, individuals with hearing loss without tinnitus may experience similar disorders as individuals with tinnitus disorder, such as cognitive, emotional, and behavioral reactions.

The ascending lateral sound pathway encodes the auditory component of tinnitus. Auditory nerve fiber deafferentation (reduced FC within auditory network) results in maladaptive structural and functional plasticity, including homeostatic downregulation of tonic inhibition and reorganization of the cortical tonotopic map in the central auditory pathway, beginning in the dorsal cochlear nucleus (DCN), then in the inferior colliculus (IC), and finally in the auditory thalamocortical system (28, 29). Somatosensory and auditory afferent projections could be integrated into the fusiform cells of the DCN, which is correlated with the development of somatosensory tinnitus (28). Frequency-specific increases in spontaneous firing rates, abnormal neural synchrony, and burst firing in the DCN, IC and medial geniculate body (MGB) lead to abnormal theta-range resonant interactions between the thalamus and cortex, referred to as thalamocortical dysrhythmia, and tinnitus generation (28, 29, 33). Thalamocortical dysrhythmia is characterized by maladaptive changes in the auditory cortex resulting from deafferentation. The auditory cortex can obtain missing information from neighboring cortical cells through a selective increase in cortical excitability due to imbalanced neuronal excitation and inhibition or by dendritic and axonal rewiring. If the bandwidth of deafferentation is large, alternatively, the auditory cortex could pull the missing auditory information from parahippocampal memory (33, 35).

The ascending non-specific medial “suffering” pathway encodes affective components of pain, tinnitus, and other pathologies (56). Amygdala–thalamic reticular nucleus (TRN) circuit is the projection of the basolateral amygdala of the limbic system to the TRN (99, 100). The amygdala regulates TRN gating auditory information by excitatory projections from the basolateral amygdala to the TRN (101, 102). The cortex and thalamus simultaneously send excitatory collaterals to the TRN (103). Thus, the amygdala–TRN functions as an ascending gatekeeper, regulating the affective value from the medial pathway. The mediodorsal and ventromedial posterior nuclei of the thalamus relay the ascending auditory information to the rostrodorsal anterior cingulate cortex, anterior insula, and auditory cortex, which are correlated with tinnitus suffering (33, 98, 104).

The top-down noise-canceling pathway separately regulates abnormal auditory activity in the auditory thalamus from the lateral and medial pathways. The inhibition deficiency of the descending noise cancellation pathway plays a critical role in the generation of tinnitus in individuals without initial tinnitus triggers or deafferentation and without map reorganization (105). The noise-canceling pathway contains a frontostriatal top-down gating system circuits, which is related to the affective value of internal and external percepts (34, 99). The substrates of the frontostriatal gating system involve the ventromedial prefrontal cortex (vmPFC) and the nucleus accumbens (NAcc) of the basal ganglion and the subgenual cingulate, which are closely related to the generation and maintenance of tinnitus (106, 107). The projections of the vmPFC and limbic structure of the NAcc extend to TRN, a region that consists of a layer of inhibitory GABAergic neurons present between the thalamus and neocortex and produces a direct inhibitory input on the neurons of the sensory thalamocortical relay neurons (108). The significance of the salience of the acoustic signals due to abnormal TRN gating sensory information is evaluated by the circuits of the vmPFC, limbic structure of the NAcc and auditory cortex. The vmPFC does not properly suppress tinnitus-related hyperactivity in the thalamus, and aberrant neuronal excitability in the NAcc results in tinnitus-related distress (106).

However, the abnormal auditory signals induced by the interaction between the ascending and descending pathways do not yield tinnitus percepts, including tinnitus loudness and lateralization, tinnitus duration, and tinnitus suffering. The signal must be processed in multiple parallel, dynamically changing, and partially overlapping non-auditory networks with specific spontaneous oscillatory frequencies (35). Tinnitus loudness percepts require synchronized activation of the salience network, including the anterior cingulate cortex (ACC) and anterior insular cortex, via the lateral pathway and transmit signals into the awareness network or perception network, which includes the subgenual ACC, dACC, pACC, precuneus, frontal cortex, and parietal cortex (34, 35, 98, 109). Two subpathways of the lateral pathway, the tonotopic lemniscal and the less tonotopic extralemniscal ascending auditory pathways, can induce the tone and noise types, respectively (33). The signal from the medial and noise canceling pathways must be processed by the salience network and distress network (including the subgenual and dorsal ACC, anterior insula, and amygdala), which results in tinnitus suffering (106, 109). The dysfunctional plasticity in these circuits may be critical for the tinnitus percept and percept-induced distress response.

Synchronized activation of the executive control network, including the dorsal ACC, dorsolateral prefrontal cortex, and inferior parietal lobule, is involved in cognitive and behavioral impairments, such as negative automatic thoughts and selective attention to tinnitus (22). Thus, communication between the executive control network and distress network leads to worsening tinnitus suffering. The default mode network, including the medial prefrontal cortex, posterior cingulate cortex, and precuneus, which controls self-representational processing, could result in tinnitus chronification by becoming tinnitus and tinnitus suffering as an integral part of the self, the new normal default state when pathologically communicating with tinnitus-provoking networks (98, 110–112). The imbalance of the positive reward system, in which the main hub is the NAcc, and the negative reward system, in which the main hub is the lateral habenula, receives dopaminergic projections from the ventral tegmental area and influences reward functions, including valuation, decision-making, and learning, and behavioral changes, such as pain, anhedonia, and motivational disturbances (104, 113). The functional connectivity between the pregenual anterior cingulate cortex and auditory cortex with the NAcc induced by tinnitus might also be involved in tinnitus chronification (104, 114). Changes in connectivity between the lateral pathway and motor network are associated with physical disability. Nonetheless, the difference between underlying tinnitus and tinnitus disorder needs further investigation. The dysfunctional plasticity in each network could explain the heterogeneity of tinnitus disorder and the response to interventions.

Frailty is a major modifiable factor of biological age and an extreme aging status. Individuals with frailty show specific changes in the brain microstructure and FC. Gray atrophy and white matter lesions are associated with physical frailty (115, 116). Microstructural neuroimaging of those with physical frailty has revealed gray atrophy in the medial frontal cortex; the basal ganglia (BG) region, including the putamen, caudate, and thalamus; the anterior cingulate cortex; and white matter lesions in the body of the corpus callosum (117). Brain structural markers of cognitive frailty, such as frontotemporal and subcortical atrophy, increased white matter hyperintensities, and decreased white matter microstructure integrity, have been indicated in previous studies (118, 119), which is different from the characteristic medial temporal lobe atrophy in early Alzheimer's disease (120). These brain structural degenerations result in decreased segregation within subnetworks and increased integration between subnetworks, including the goal-oriented executive control netework, DMN, salience network, motor network, and reward network.

Physically frail individuals present reduced FC (synchronization) in fronto-partietal areas (the executive control netework, for maintaining and processing information in working memory, problem-solving and decision making) (121) and supplementary motor areas associated with motor function and reduced intranetwork FC within the fronto-partietal, ventral attentional, and posterior DMN (122). In an Irish longitudinal study, frailty, assessed by the frailty index (FI), a self-reported multidimensional deficit, could capture physical but not cognitive impairment. Connectome-based predictive modeling results indicated a positive correlation between the FC of the visual network and FI and a negative correlation between FC in the BG and FI (123). The highest node of both networks was the caudate, but with different FC patterns: from the caudate to the visual network, and from the caudate to the DMN-related areas, respectively. The different connectivity patterns along with FI could reflect different recruitments in the brain network to maintain daily physical performance (123). Patients with mild cognitive impairment and frailty (cognitive frailty), which were assessed by the multidimensional frailty index, indicated that increased FC between the right hippocampus and clusters in the temporal gyrus was positively associated with higher frailty index scores (124). Late-life depression (LLD) is usually combined with physical frailty, referred to as psychological frailty (41). Individuals with LLD exhibit aberrant FC within and between the salience network, the DMN, the executive control network, and the frontal striatal reward network (125, 126). Network-level disruptions in connectivity result in cognitive deficits with diminished top-down control salience of negative stimuli, negative thoughts and emotions, and motivational disturbance, and in turn, maladaptive behavioral manifestations.

Tinnitus or tinnitus disorder likely interacts with frailty to produce ‘vicious cycling', which contributes to the heterogeneity of tinnitus disorder and the response to interventions. The hypothesized frailty model of tinnitus disorder revealed a bidirectional relationship between frailty and tinnitus disorder (Figure 2). Frailty accelerates biological aging and increases the vulnerability to morbidity in multiple physical, cognitive, and psychological dimensions. Frailty is associated with tinnitus severity (27, 127), cognitive impairment (128), depression (129), and other physical diseases. The risk factors for frailty, such as chronic inflammation, cerebrovascular pathology, and pathological protein burden in the brain, can cause white lesions in the brain (115–117), cortical and subcortical atrophy (118–120), and homeostatic imbalances in intrinsic functional networks (121–126). The intrinsic neural networks, such as the salience network, DMN, executive control network, motor network, and reward system, are disrupted to varying extents in patients with different frailty phenotypes or dimensions. The decreased FC intranetwork (segregation) and increased FC internetwork (integration) among the salience network, DMN, executive control network had been proposed to underline somatic symptom and neuropsychiatric disoders, including tinnitus, evidently contribute to the heterogeneity of disorders and responses to tinnitus interventions (Figure 2). Frailty results in dysfunction in the salience network, DMN, executive control network, and the interaction of these networks, including increase salience of negative stimuli and negative self-referential thinking, decreased cognitive control and selective attention, and the salience network driven switch between DMN and executive control network. Dynamic executive functioning hypothesis, referred to as a balance between the salience and the executive control networks had been proposed for individual to dynamically and quickly adjust the ongoing behavior when faced with a stressful event (130). Frailty decreases cognitive reserve and more cognitive effort demand for auditory function in older adult with tinnitus. The reallocation of cognitive resource might not meet the salience network to identify tinnitus and initiate the switch between DMN and executive control network, which lead to failure in tinnitus inhibitory control, and dynamic transition of tinnitus percept to tinnitus disorder, tinnitus percept and suffering chronification.

Tinnitus or tinnitus disorder may also accelerate frailty development and progression and contribute to the adverse outcomes of frailty (Figure 2). Typically, tinnitus percept or suffering, as a stressor, can result in the adaptation of neural and physiological responses by dynamic allostasis. Segregation and integration in the brain network could maintain homeostasis. However, repeated or continuous tinnitus, together with other risk factors, may cause frailty or the vulnerability of multiple physiological systems, including neural networks for cognitive and emotion processing. The sensation information of tinnitus percept from the ascending lateral auditory and suffering information from medial suffering regulated by top-down-noise canceling pathway activates salience, percept network, DMN, and executive control network by dynamic allostasis compensatory mechanism to trigger goal-oriented behaviors, including decreasing the salience of tinnitus percept and suffering, negative automatic thoughts of tinnitus. Moreover, the adaptive behaviors will be added to the repertoire of successful strategies for the response to the following same tinnitus percept or suffering and saving cognitive resources. In combination with aging and other factors, such as psychological stress, the vulnerability of dynamic executive functioning to maintain homeostasis will increase, resulting in tinnitus percept and/or suffering, the transition from tinnitus percept to tinnitus disorder, and frailty. furthermore, the interaction of tinnitus and frailty might cause allostatic load and maladaptation and adverse physical, cognitive, and psychological outcomes and chronic morbidity.