Aging in the Primary Auditory Cortex

Abstract

Age-related auditory dysfunction affects half of all individuals 60 years and older, yet its causes are poorly understood. While commonly associated with cochlear dysfunction, a growing body of literature suggests that dysfunction originating in the auditory cortex itself is also a major contributor. Here, we review recent literature that describes the effects of aging on the primary auditory cortex in humans, non-human primates, rodents, and a variety of other species. During aging, individuals with auditory cortical dysfunction experience deficits in spectral and temporal processing of sounds, resulting not only from a loss of inhibition but also from an extensive restructuring of cortical circuits. Importantly, aging in the auditory cortex is sex-dependent, yet few studies account for this variable. A lack of comprehensive knowledge on aging in the auditory cortex hinders the path toward restoring cortical function through auditory training or broader cortical rehabilitation paradigms. Thus, we propose a cohesive mechanism of aging in the primary auditory cortex that involves a complex interaction between excitatory and inhibitory neurons, which several factors can modify. These factors include input from higher-order cortical areas, such as the orbitofrontal cortex, as well as the wide-ranging effects of neuromodulators and the external sensory environment, which must be accounted for in a sex-dependent manner.

Introduction

Approximately half of the people aged 60, and increasing in prevalence from age 60 onwards, will suffer from age-related auditory dysfunction, both when self-reported and measured audiologically [1]. Classically, patients present with two main symptoms of auditory dysfunction: either difficulty in detecting sounds or in perceiving sounds [2]. While sound detection often manifests as not being able to hear sounds of a low amplitude as a result of disrupted cochlear function, perceptually, patients frequently struggle with sound source localization, temporal perception, and gap detection even in older individuals with normal cochlear function, suggesting dysfunction in central components the auditory pathway, including the auditory cortex (ACtx), the auditory thalamus, and portions of the auditory brainstem and midbrain [3–10] (Fig. 1a). In day-to-day life, central auditory system dysfunction often affects older individuals, causing them to struggle to discriminate speech from a noisy background [7, 11].

Fig. 1.

The aged auditory cortex represents auditory stimuli with less specificity. a Schematic of the central auditory system necessary for proper auditory processing. Adapted from Fig. 2.2 of Goodrich & Kanold, [199]. The lemniscal auditory pathway is shown with the solid arrows, and the non-lemniscal pathway is shown with dotted arrows. b Adapted from Fig. 6 of Turner et al., [26]. Collapsed peristimulus time histogram from layer V primary ACtx neurons with complex response profiles from young and aging rats in response to a 200 ms current pulse ranging from 1 to 10 nA (represented by green bar). c Adapted from Fig. 2 of Ramamurthy & Recanzone, [20]. Exemplar ON and OFF intensity rate-level functions (top) and peristimulus time histograms (bottom) from both young and aging non-human primate neurons in response to a 200 ms unfrozen Gaussian noise stimulus (represented by green bar). d Schematic containing the main point made in this figure: as inhibition is lost in aging, the network becomes increasingly homogenous in its stimuli encoding. Circles represent neurons, and lines represent connections. Excitatory neurons are in blue and inhibitory neurons are in red

Cochlear dysfunction is exceedingly common in adult humans [12], likely due to exposure to harsh auditory environments, such as noise pollution, prolonged use of personal listening devices at high volumes, or loud environments. Additionally, independent dysfunction can also be observed throughout the remainder of the ascending auditory pathway in the majority of the aging population, contributing to this auditory dysfunction, although the prevalence of central auditory system dysfunction is difficult to measure in isolation from peripheral auditory system loss [4, 13–16]. To disambiguate central auditory system dysfunction from peripheral hearing loss, carefully controlled animal studies have been used. To this effect, multiple studies have searched for and found independent effects of aging on ACtx in non-human primates, rats, and mice [17–32]. Studies evaluating the effects of aging on single neurons in the ACtx of the macaque monkey suggest that aging increases both the spontaneous and sound-driven activity of these neurons [18, 19], increasing the effects of noise on auditory processing. These findings significantly advanced the field’s mechanistic understanding of the basis of age-related dysfunction in the ACtx and closely fit the human phenotype of decreased fidelity in auditory perception. To supplement this new knowledge, mouse models were used due to the wide availability of transgenic methods that enable the simultaneous recording of the activity of thousands of neurons in large cohorts of animals raised in controlled housing conditions designed to minimize variations in peripheral hearing experience not easily possible with non-human primates [21–23]. Here, we attempt to synthesize literature obtained from humans, non-human primates, rodents, and a variety of other species about aging in the ACtx.

The ACtx is thought to represent sound stimuli using sparse neural networks of excitatory and inhibitory neurons that respond preferentially to specific stimuli composed of numerous variables, such as the frequency, intensity, or location of the sound [33–35]. In aging, numerous changes occur in the ACtx, ranging from brain hemisphere-scale changes in the asymmetric processing of spatially distributed sounds [36] to smaller-scale, network-scale changes in the excitation-inhibition balance in local cortical circuits [25]. Over the past decade, a large body of literature has supported the idea that decreased inhibition is responsible for most dysfunctions observed in the aging ACtx [32, 37–42] although changes can also be seen throughout the entire auditory pathway [42]. However, recent findings suggest that changes in the ACtx involve both decreased inhibition and altered excitation rather than decreased inhibition alone [21, 22, 24, 25]. Therefore, this review describes the recent advances in our understanding of aging in the ACtx.

Throughout the entire auditory system, auditory stimuli can be primarily characterized by their frequency spectrum, temporal envelope, and binaural differences [43]. Within the ACtx, neurons then integrate this information into complex readouts of multiple components of each auditory stimulus, including its spatial location, amplitude, spectral and temporal components, including as sound onset or offset leading to high stimulus selectivity [33–35, 43–47]. Furthermore, sound features are also encoded by neuronal ensembles in ACtx [21, 45, 46, 48, 49] and neuronal responses are modulated by arousal and behavioral state, such that neural and ensemble responses vary between passive tone presentation and during behavioral tasks [22, 47–53]. Sex has also been shown to affect the activity of neural populations and circuits in the ACtx, especially during aging [21, 22, 24]. Given the processing complexity in the brain, even slight circuit and functional changes associated with aging can have a dramatic impact on cognitive abilities, including those in the sensory systems. However, the detailed circuit and processing changes in the ACtx with aging were not well-studied until recently. The cerebral cortex demonstrates a tremendous capacity for plasticity, and an expanding body of research highlights promising strategies, such as auditory training and vagal nerve stimulation, to restore or preserve auditory cortical function [23, 54–58]. Thus, this review will focus on the changes in the primary ACtx that disrupt the spectral and temporal representations of sound before concentrating on the effects of sex and plasticity in the aging ACtx. While this work focuses on the primary ACtx, all ACtx references can be assumed to refer to the primary ACtx unless otherwise stated. Available research about the secondary ACtx is referenced when available, but differences in aging between primary and higher-order areas remain a central open question.

Decreased Functional Response Selectivity in the Aging Auditory Cortex

The aged ACtx fundamentally differs from the young ACtx in both structure and function. Thus, the aged ACtx represents stimuli differently and often with less fidelity than the young ACtx. Research conducted in humans has revealed distinct cortical activity patterns in young and older adults when listening to auditory stimuli. When an individual listens to auditory stimuli, the ACtx performs a type of pattern recognition, allowing for proper speech recognition [59, 60]. However, sustained neural activity resulting from this speech recognition pattern appears to be lost in the aged human ACtx when measured using magnetoencephalography (MEG) [61]. Accordingly, speech has been shown to recruit more neurons in the ACtx of older individuals as compared to young individuals [62, 63], suggesting that decreased auditory processing abilities, which limit speech comprehension, correlate with an overrepresentation in, and thus less specific, neural responses [64]. As such, older listeners often place excessive emphasis on the context of speech to aid in word recognition, which can lead to a misperception of the actual spoken word, a phenomenon termed false hearing [65]. This, too, has been hypothesized to be a result of decreased inhibition of high-probability responses, leading to the misperception of speech [66].

Moreover, multiple studies in humans have found that a decrease in inhibition, as measured by magnetic resonance (MR) spectroscopy, in the ACtx correlates with impaired speech-in-noise comprehension [38, 39, 67]. However, due to the inherent limitations of working with human subjects, these experiments lacked many finer observations made possible by invasive experiments in animal models. Thus, studies using non-human primates nicely complement human studies by allowing for direct, temporally precise measurements of neural activity in the ACtx. Such studies have revealed that auditory stimuli cause a higher spike rate of individual excitatory neurons in the aged ACtx, in addition to an increase in spontaneous activity present in aged neurons [17–19], perhaps partially due to a compensatory loss of inhibition in the ACtx to make up for decreased peripheral excitation in the case of accompanying peripheral hearing loss [26, 68]. Thus, the increase in cortical activity, concomitant with a decrease in functional response specificity to any given sound, suggests a fundamental change in the underlying circuitry of the ACtx throughout one’s lifespan.

With the advances in transgenic models, the relative ease of controlled housing, and shorter lifespan, single-neuron recordings and circuit-level in vivo calcium imaging in rodents have proven invaluable for studies on the individual neuron and circuit levels. Early studies that recorded single-neuron activity in response to a GABA_A receptor antagonist indicated that local inhibition plays a crucial role in maintaining sparse tuning in the ACtx [69, 70], and subsequent molecular and slice-physiology studies found that aging disrupts levels of inhibition, leading to broader tuning throughout the ACtx [40, 71]. In one such experiment, the intrinsic excitability of layer V rat auditory cortical neurons with complex receptive fields, namely those heavily shaped by inhibitory inputs, was recorded in young and aging animals in response to a 200-ms current pulse [26]. While neurons from young rats were difficult to excite, presumably due to the plentiful inhibitory influences (Fig. 1b, top), those from aging rats were easily excitable (Fig. 1b, bottom). Decreased inhibition is consistent with the observation that the aged ACtx employs less stable encoding, resulting in greater variability in single-neuron responses compared to its younger counterpart [28] and exhibits broader frequency tuning [27]. Despite this apparent lack of functional inhibition, in vitro studies in mice demonstrate that both the excitatory and inhibitory circuits in layer 2/3 of the ACtx change across age, likely by equivalent amounts [24, 25]. To complement this work, in vivo, 2-photon calcium imaging of excitatory neurons in the mouse ACtx allows imaging of many thousands of single neurons in awake animals. This technique, thus, enables the dissection of circuit dynamics at single-neuron resolution. Using mice of the CBA/CaJ strain, which have little early peripheral hearing loss and thus allowing for the study of central auditory system dysfunction in isolation [72], imaging studies showed that local inhibition might well play a role in the dysfunction of the ACtx seen in old age, but that widespread increases in correlated activity observed on the circuit level appear to also indicate a fundamental change in excitatory network structure [21, 22, 32]. Furthermore, these results demonstrate that while single-neuron tuning is largely unchanged with age, network-wide responses to auditory stimuli are more functionally homogeneous with age. Thus, aging seems to affect network responses more profoundly than responses of single neurons, likely due to the effects of aging on top-down or neuromodulatory inputs or due to remodeling of intrinsic ACtx circuits.

Decreased Temporal Selectivity and Response Diversity in the Aging Auditory Cortex: The Danger of Losing the Beat

In the ACtx, neural timing is crucial for ensuring that each discrete sound stimulus can be processed independently [73]. In the primary ACtx, neurons show a variety of temporal response shapes and primarily encode large changes in sound intensity, namely sound onset and offset [44, 74–76]. These neural representations of stimulus timing correlate with behavioral readouts of auditory perception, resulting in the robust detection of stimulus onset and offset during a tone detection task [22]. Similarly, the ability of the human ACtx to encode the temporal envelope of speech is thought to underlie speech intelligibility [77, 78].

Aging degrades temporal processing in the ACtx, leading human listeners to considerable difficulty in using temporal cues to isolate components of the auditory scene, such as speech [61, 79, 80]. More specifically, aging causes a decrease in the ability of the ACtx to encode the envelope of speech and temporal fine structure [64, 81]. This loss in fidelity is likely due to prolonged neural responses in the ACtx causing discrete stimuli to bleed into one another [82] and a distinct rise in prestimulus energy, thought to represent background noise, that occurs with aging, which in turn distorts auditory perception [83]. As such, older listeners must rely on a reactive strategy, where the ACtx only transiently uses cues from the external auditory environment because of difficulty in maintaining a cognitive state focused solely on the auditory task at hand [84]. Younger individuals, in contrast, typically employ a proactive listening strategy. Work done in the aged ACtx of non-human primates has found that although an equal number of neurons responded to amplitude-modulated noise, a lesser proportion were able to synchronize their firing rate to the amplitude modulation of the stimulus [19], although the literature paints a more conflicting picture in rats [28, 68], perhaps due to their higher-shifted range of frequencies where phase-locked responses become far less feasible. Thus, at least in non-human primates, of the overall smaller proportion of neurons that phase-lock with the temporal envelope of a stimulus as compared to young, those neurons can only do so weakly instead, leaving the network with primarily a firing rate code where neurons simply spike more often during the stimulus to encode the timing of stimuli, rather than a more precise temporal code where neurons phase-lock with the temporal envelope of a stimulus. Phase-locking is essential to temporal encoding throughout the ascending auditory system [85–87], likely contributing to the changes seen in the ACtx.

Aging disrupts even relatively basic temporal responses of neurons to sounds. Throughout aging, sound onset is overrepresented in the human ACtx by way of its response magnitude [82]. In non-human primates, the rate-level function differs between onset and offset responding neurons. This difference is influenced heavily by age [20] (Fig. 1c). Similar results are found in mice [21]. These disruptions that lead to an overrepresentation of sound onset are accompanied by a loss of sound offset representation [21], likely as a result of lessened inhibition in the ascending auditory pathway thought to give rise to sound offset encoding.

Sound offset representation is crucial for temporally precise stimulus representation [88]. Sound offsets are robustly encoded in the ACtx in many species [20, 44] and their encoding has been shown to decrease with age [20–22]. Consequently, the disappearance of a tone offset response may underlie some of the auditory difficulties experienced by older listeners. In aged mice, the sound offset is no longer a viable cue for sound detection [22], effectively eliminating half of the temporal information of a sound. Given that sound offsets are thought to arise from inhibition in the ascending auditory pathway [44], a decrease in sound offset representation is consistent with a reduction in inhibitory influence in aging or selective degradation of ascending pathways mediating these signals (Fig. 1d).

Not only does the temporal response pattern of individual neurons change with aging, but the variety of unique temporal responses of neurons in response to a sound stimulus decreases with age [21], likely decreasing the coding diversity and perceptual accuracy of the sound stimulus as fewer neurons are encoding unique temporal aspects of the stimulus. One prime example of this change is a distinct rise in prestimulus activity, which robustly predicts a false perception of a sound stimulus found only in older animals [22]. In young animals, this likely representation of noise in the external environment is easily suppressed by complex inhibitory networks that work to create noise-invariant representations of the attended-to auditory stimulus [53, 89]. Instead, this neural distraction is present in the ACtx of older animals, leading to increased difficulties in auditory perception.

Decreased Behavioral Modulation of Responses in the Aging Auditory Cortex

The ACtx encodes stimuli in an environment- and task-dependent manner. For example, a listener can distinguish the speech of a single person from that of individuals not relevant to a conversation. As such, the ACtx preferentially encodes stimuli when presented in a behaviorally relevant context compared to passive exposure [22, 47, 50–53]. This adaptive capability of the ACtx is thought to enable humans to understand speech in noisy environments, a process often disrupted in aging [7, 11]. In mice, a greater evoked neural response to an auditory stimulus is also observed in 87% of excitatory layer 2/3 neurons when presented as part of a behavior compared to passive exposure [48, 49]. Moreover, task engagement leads to a refinement of the excitatory layer 2/3 active functional circuits, reducing their size to nearly a quarter of that found during passive exposure, thereby enhancing the representation of the behaviorally relevant stimulus [48, 49]. In its most basic measurement, this enhancement can be observed in the modulation of single-neuron responses via task engagement, such as during a tone detection task (Fig. 2a). In a previous study using 2-photon imaging of excitatory layer 2/3 during behavior in aging animals [22], we found a gain enhancement in both young and aging animals (Fig. 2b). When we grouped neurons with an increase in calcium activity during tone presentation and those with a decrease (~ 20%) (Fig. 2c), we further found that the only difference between young and aging animals exists in the gain effect in those suppressed neurons (Fig. 2d). Circuit-scale observations showed that many effects observed in non-behaving animals, such as increased neural correlation, were similar to those in behaving animals [21]. In the ACtx, a population of neurons increases their correlated activity when a stimulus has behavioral relevance (36%), and another population of neurons becomes decorrelated by this behavioral relevance (64%) [22, 48]. However, only in the aging ACtx do the neurons that become more correlated in response to a behaviorally relevant stimulus have a larger increase in correlation from the passive condition than those in the young ACtx. Additionally, those neurons that became uncorrelated in response to a behavioral stimulus already had a higher baseline correlation level than young neurons [22]. This result suggests that, during behavior, neurons in the aging ACtx show a dramatic increase in attentional-driven correlated responses and a higher baseline of correlated activity found within the network. However, this difference in circuit dynamics was not largely driven by the outcome of the mouse’s behavior, as observed in the younger ACtx, but rather by a highly correlated network of neurons responding to the vast majority of stimuli, whether behaviorally relevant or not. Together, these results could explain why older listeners experience significant difficulty focusing on only salient aspects of the auditory environment (Fig. 2e).

Fig. 2.

Aging in the auditory cortex leads to a loss of differential suppressive responses during a behavioral task. Adapted from Figs. 1 and 3 of Shilling-Scrivo et al., [22] and Fig. 1 from Mittelstadt et al., [23]. a Neurons in the awake mouse primary ACtx were recorded during a tone detection task in both young and old mice. b Fluorescence traces of each neuron illustrating the basic gain calculation using fluorescence intensity in both the passive and active conditions. c Average suppressive and facilitative gain traces. d Average suppressive and facilitative gain as calculated by behavioral stimulus. e Schematic containing the main point made in this figure: in aging, the influence of the remaining inhibitory neurons in the auditory cortex decreases, leading to less differential responses to behaviorally relevant sounds. Circles represent neurons, and lines represent connections. Excitatory neurons are in blue and inhibitory neurons are in red. Dashed lines indicate loss of synaptic modulation

Sex Differences in the Aging Auditory Cortex

In the human brain, the expression of numerous genes, and thus their effects, differ between females and males [90]. These gene expression differences persist into old age [91], and it has been shown that both male and female brains age differently, dependent on a second X-chromosome [92]. Accordingly, both males and females have different baseline levels of auditory processing [93, 94], with females displaying better harmonic, speech onset, and even better hearing sensitivity, illustrating differences in structures throughout the auditory system from outer hair cells in the cochlea to centers of the brainstem. Across both humans and mice, females possess a larger ACtx than their male counterparts as measured either anatomically in post-mortem tissue or using neuroimaging [95, 96]. Functionally, some processes, such as representations of auditory stimuli derived from only silent lipreading, can only be observed in the left primary ACtx of females [97], but more often, there are slight changes in processing between the sexes as seen in a study comparing bilateral activation of the ACtx to noise or tones [98]. In earlier stages of the ascending auditory system, it has been demonstrated that there are differences between males and females, such as susceptibility to age and noise-induced hearing loss, a cochlear phenomenon primarily due to sex differences in noise exposure, preferentially affecting males and volume of the visual cortex correlating with hearing difficulty only in females [99–102]. Furthermore, in both mice and humans, females show increased synchrony in spiral ganglion neuron activation, as evident in wave I of the auditory brainstem response [103–105], at least partially due to the likely expression of glutamate subunits with faster kinetics [103], reflecting a fundamental difference in the auditory system between males and females that can be observed even further upstream in centers such as the cochlear nucleus and inferior colliculus [103, 106, 107].

However, until recent work in animal models [21, 22, 24], little research investigated any potential differences in the aging ACtx between males and females on a neural circuit level. This work found large changes in neuronal, temporal, and spectral encoding between the aging male and female ACtx. More specifically, only males exhibited increased amplitudes, lower bandwidths, greater spectral homogeneity, and the elimination of a tone offset response [21] (Fig. 3a), all previously considered universal hallmarks of the aging ACtx. Similar sex-specific effects were observed during behavioral task performance, where males exhibited a distinct rise in prestimulus neural activity that could reliably predict a wrong early response from the mouse, which was nearly absent in females [22]. These sex-dependent differences in auditory processing are likely due to sex-dependent changes in aging ACtx circuits as these mice had little peripheral hearing loss and thus likely equivalent peripheral inputs between males and females, a scenario not easily found in isolation in humans but likely playing a large, and previously undiscovered role, in sex-based differences in auditory processing abilities. Should there have been additional differences in peripheral input in addition to differences in cortical circuitry in aging, these findings would likely have been exacerbated.

Fig. 3.

Neural circuits in the male, aging auditory cortex are preferentially disrupted. a Adapted from Fig. 5 of Shilling-Scrivo et al., [21]. Stimulus-dependent correlations from young and aging animals split by sex. b Adapted from figure 8 of Xu et al., [24]. Correlation of excitatory and inhibitory circuits across all layers. c Schematic containing the main point made in this figure: with aging, the male auditory cortex is more disrupted as compared to the female and thus will have less unique stimulus encoding. Circles represent neurons, and lines represent connections. Excitatory neurons are in blue and inhibitory neurons are in red

In vitro studies have shown that compared to adult mice, the aging ACtx of male mice has altered interlaminar excitatory and inhibitory connectivity. In contrast, aged female mice do not appear to exhibit changes [24] (Fig. 3b), a finding that was once thought to be a universal hallmark of the aging ACtx. These latest data from animal models underscore the need for further research on the aging ACtx to be stratified by sex (Fig. 3c).

Preventing Age Changes in Neural Cortical Processing

The cortex, including the ACtx, has been shown to possess a remarkable ability to change in response to long-term alterations in the external environment [55–57]. Both decreased and increased richness of auditory experience can change the ACtx. For example, in individuals with hearing impairment, the responses of ACtx neurons become less distinct, while parts of the visual cortex are recruited for auditory processing of the degraded signals [108, 109]. However, in individuals with hearing aids, and thus an increase in external signal availability, there is enhanced speech processing, and thus likely cortical plasticity [110–112]. In contrast, musicians generally have more extensive and interconnected cortical areas devoted to auditory processing than average listeners [113–115]. Importantly, these cortical changes persist into old age [116], and can have long-lasting functional consequences such as improvements in speech-in-noise perception [117]. There is also evidence suggesting that auditory training in older adults can improve auditory processing abilities [118], but there has been little work in humans looking at the effects of auditory training on cortical plasticity.

In rodent models, short-term auditory training (~ 2 months), even when attempted at old age, has been shown to forestall neural changes or even revert aged auditory cortical neurons to a more youthful state while also increasing the population of inhibitory neurons [29–31]. This learning-induced rejuvenation was most strikingly shown in the increase in ACtx spectral organization of the rat, regardless of age [31] (Fig. 4). On a network scale, after a longer-term training experience (~ 6 months), a similar phenomenon is observed, but it is not as simple as the entire network being frozen in time and remaining similar to that of a younger one [23]. The mice trained on a tone detection task with a water reward exhibited a suppressed tone-onset response more in balance with the limited tone-offset response found in aging mice, and neuronal responses were less correlated during tone presentation as compared to responses in mice only passively exposed to the same set of tones. Additionally, the network seems to rebalance itself by introducing more heterogeneity in neurons tuned to tone onsets and more homogeneity in neurons likely influenced by inhibitory inputs, thus deemphasizing tone onset and amplifying the effect of inhibitory neurons in the network. However, for these effects, mice had to have nearly continuous access to auditory training for nearly six months, beginning around middle age, completing tens of thousands of correct tone-detection trials in exchange for their daily water. Thus, this strategy, although effective, is likely too intensive to apply proactively to humans, and the water-reward scenario cannot be directly replicated. However, one could envision that a purposefully designed auditory game or increased, focused auditory exposure may preserve ACtx function in old age, albeit likely with lower efficacy.

Fig. 4.

Auditory training in adulthood can improve auditory cortex spectral organization. Adapted from Fig. 1 of de Villers-Sidani et al., [31]. Exemplar characteristic frequency (top), tuning curve width (middle), and receptive field overlap relative to the respective recording sites identified on the maps by “*”

Long-term auditory training may not be efficient for people other than musicians who have lifelong high-engagement experience. In rodents, cross-modal plasticity induced by short-term visual deprivation has been shown to alter the ACtx, potentially providing a novel mechanism for recovering central auditory processing circuits lost during aging. The sensory cortices show a high degree of interconnectivity, and this is well demonstrated during periods of sensory deprivation when one or more of the non-deprived sensory cortices see heightened function [119, 120]. In mice with short-term (1–2 weeks) periods of visual deprivation, there are durable changes to the intracortical circuits of the ACtx, increasing the frequency tuning within the neurons of the ACtx [121–123]. Functionally, a visual deprivation paradigm has been shown to most likely increase central auditory system function in mice using a tone-detection task [58]. Thus, temporary sensory deprivation may be a promising strategy for restoring some of the auditory function lost with aging, although this remains unexplored in the context of aging.

Another possible way to reorganize the ACtx and potentially revert age-dependent changes is the selective engagement of neuromodulatory systems, potentially paired with selective auditory stimulation. As such, stimulation of the vagus nerve, when paired with auditory stimuli, has been shown to cause reorganization of the ACtx, which is highly dependent on the nature of the auditory stimulus presented [54, 124], extends to the temporal domain [125], and impacts behavioral output in an auditory discrimination task in mice [124]. These phenomena are thought to occur from a release of numerous neurotransmitters such as acetylcholine and norepinephrine, which likely causes increased expression of NMDA and GABA_A receptors, which dramatically change neural dynamics [124, 126, 127]. Although most vagus nerve stimulation studies currently focus on its application in tinnitus [128–130], this could likely be applied to a variety of central auditory processing disorders [54, 131], including aging.

The Molecular Basis of Aging

Within the ACtx, and indeed the wider brain, circuit-level changes are preceded by those on a molecular level. Often, this is achieved via changes in neurotransmitter receptor expression levels, including NMDA and AMPA receptors for the excitatory neurotransmitter glutamate and GABA receptors for the inhibitory neurotransmitter GABA. In the rodent ACtx, all of these receptors are age and experience-dependent [132–135], highlighting their role in ACtx dynamics. In more advanced age, the expression of both NMDA receptors and an enzyme responsible for the production of GABA have been shown to increase compared to young adult rats [136], and changes in NMDA receptor subunit expression have been shown to affect a type of neuroplasticity called long-term potentiation in the ACtx [137]. Although the work on NMDA receptor changes has thus far been limited, a much more extensive body of literature exists on the role of GABA in the aging auditory system. From post-mortem human tissue studies, it has long been established that aging is associated with a decrease in glutamic acid decarboxylase (GAD), an enzyme that produces GABA, in the ACtx [138], and that GABA loss in the ACtx with aging has functional consequences [39, 67]. Further studies in rats demonstrated that GABA receptor subunit expression changes with age [71] and that GABA binds these receptors less efficiently [40, 139]. Together, these results indicate that the ACtx is more prone to hyperactivity with age due to a decrease in inhibitory weight on the networks of the ACtx. Thus, these networks would provide a less specific representation of sounds in the ACtx.

Although neurotransmitter receptor expression may be the most direct molecular mechanism capable of modifying neural circuits in aging throughout the brain, the actions of neuromodulators need to be considered as well. Indeed, neuromodulators, including acetylcholine, serotonin, norepinephrine, and dopamine, are thought to modulate the neural activity of the human ACtx, contributing to neuroplasticity [140], and their receptors can be found at synaptic terminals throughout the cortical layers [141]. At the level of the synapse, acetylcholine has been shown to modify a neuron’s response to the excitatory neurotransmitter glutamate, and the effects of acetylcholine can, in turn, be modulated by dopamine [142–144]. Similarly, serotonin and norepinephrine have also been shown to affect auditory cortical neurons [145, 146]. In aging, levels of these neurotransmitters differ. For serotonin, the expression has been shown to decrease with age [136], but the functional effects are unclear. A bit more is known about changes in the cholinergic modulation in aging. There is an increase in the precursor enzyme, choline acetyltransferase, with age [147]. Functionally, acetylcholine yielded decreased post-synaptic responses in auditory cortical neurons in an in vitro patch clamp experiment from aged rats [148], likely as a result of altered acetylcholine receptor subunit expression. These results highlight the complexities of altered neuromodulation and neurotransmission in aging, underscoring the need for further research into the molecular underpinnings of the circuit changes described throughout this review.

Altered Inputs in Aging: Complex Interactions Throughout the Auditory Pathway

Although aging does independently affect the ACtx, as illustrated by the multitude of studies referenced above, the ACtx is far from the only component of the auditory system to be affected by aging. In the peripheral auditory system, specifically the cochlea and its output spiral ganglion neurons, the effects of aging can be observed in age-related hearing loss. Work dating back several decades used post-mortem human temporal bones to observe clear degradation in the cochlea as a result of aging, namely loss of the signal-transducing stereocilia that sit atop the hair cells, loss of the signal-amplifying outer hair cells, and even loss of cochlear supporting cells that are crucial to hair cell function [12, 149–151]. Importantly, outer hair cell loss is often most prevalent in the basal turn of the cochlea, the region responsible for encoding high-frequency sounds, and leads to increased audiometric thresholds beginning in the higher frequencies before progressing down the frequency range with old age. There is also a loss of cochlear innervation as a result of the loss of spiral ganglion neurons [12, 149–152]. These two forms of peripheral degeneration combine to decrease the encoding fidelity of auditory stimuli before they ever reach the central auditory system. With a weaker and altered signal, the central auditory system often struggles to distinguish similar sounds or perceive them at all.

Upstream from the peripheral auditory system, the lower-order auditory processing centers of the brainstem and thalamus also experience the effects of aging that can be demonstrated either independently of or in tandem with aging in the peripheral auditory system. In the first auditory relay, the cochlear nucleus, the sound is parsed primarily by its temporal aspects, leading to precise, phase-locked representations of sound in addition to frequency representation. There is conflicting literature about the effects of aging on measures such as cochlear nucleus volume and cell counts in humans [13, 153–155], likely due to the morphological complexity found in the human cochlear nucleus, making its precise delineation challenging. However, the effects of cochlear synaptopathy are widely observed in animal models, resulting in deficits in tone-in-noise encoding and specific loss to certain types of cochlear nucleus neuron subtypes [156, 157]. In addition, there is a widespread loss of inhibition, primarily glycinergic, in the cochlear nucleus with age [158, 159]. This loss of inhibition likely contributes to a loss of temporal processing accuracy in the cochlear nucleus that is seen with age [160, 161]. Similar changes can also be seen in the next auditory processing center, the inferior colliculus in the midbrain [162–166]. Less is known about aging in the auditory thalamus, the medial geniculate body, likely due to its small physical size and deep location within the brain. To date, rodent studies have not found any significant age-related changes [167], yet this may be caused by the relatively small proportion of inhibitory neurons found in the medial geniculate body in rodents as compared to other species [168]. Many of the above-referenced changes occur independently of peripheral hearing loss. When combined with peripheral hearing loss, the effects compound.

Within the auditory system, there are significant descending pathways that go through many of the same auditory processing centers as the ascending auditory pathway. While the communication between the bottom-up and top-down auditory systems is poorly understood, the descending auditory pathway is thought to play a crucial role in processing more complex auditory environments, enabling speech perception when presented in a high level of background noise [169–171]. However, the descending system seems to be particularly responsive to peripheral hearing loss [172] and thus a good example of the close interplay between the peripheral and central auditory systems. When older individuals lose part of their peripheral auditory system function, they often show an overreliance on contextual clues to understand speech [65, 173, 174]. In animal models, deafness leads to anatomical and functional differences in the descending auditory system, both at the level of the central and the peripheral auditory systems [175–177], and even auditory training has been shown to rely on cortical projections to the inferior colliculus [178]. Thus, although this review focuses on the ACtx, the entirety of the auditory system is closely interconnected, both in a top-down and bottom-up fashion, and future research should focus on changes in this interplay with aging, both with and without peripheral hearing loss.

Discussion

We are beginning to understand that auditory dysfunction in old age is not solely due to peripheral hearing loss but also largely due to changes within the ACtx. Moreover, the ACtx shows sex-specific circuit changes with aging, leading to sex-specific differences in hearing abilities. These circuit changes in aging lead to a loss of diversity in spectral and temporal selectivity, especially in males. While a lot of progress has been made, this review highlights the critical gaps in our knowledge, especially surrounding the sex differences in the aging ACtx, as these sex differences likely require different treatment approaches.

In aging, it seems that the proper excitatory/inhibitory balance is disrupted in the ACtx, which leads to a rise in background activity and decreased stimulus encoding fidelity for simple stimuli due to reduced temporal precision. For example, if there were a similar number of neurons remaining in the aging ACtx, a simple loss of inhibitory drive could be enough to account for the observed phenotype. Imagine three different networks of auditory cortical neurons: one with weak inhibition, one with moderate, and one with strong inhibition. In response to a simple, tonal stimulus, each network would encode the sound differently. In low background noise, neurons in the weakly inhibited network would display the largest response to tone-onset and the largest amplitude response overall. Such neurons would be critical for hearing sounds of low amplitude but would suffer from poor temporal precision, as the lack of inhibition would lead to a prolonged period of excitation following the stimulus. Instead, the moderately inhibited network has high temporal precision, as there is a balance of excitation and inhibition. Neurons in this network would have a tone-offset response as inhibition and excitation are in tight enough coordination to give the precise temporal offset cue. However, in the strongly inhibited network, neurons would likely be non-responsive, as the tone stimulus by itself is not strong enough to overcome the inhibition. Therefore, in an environment with low background noise, weak and moderately inhibited neurons would optimally encode the stimulus. In a noisy environment, however, these networks are no longer optimal. With increased background noise, there will be increased excitatory drive onto each network. The strongly inhibited network will now be the network that contains the most information about the tone. The strong inhibition of the network makes the neural responses largely invariant to noise and increases the threshold of the neurons themselves. In the weakly inhibited network, neurons will strongly respond to noise and will, therefore, have no more dynamic range to respond to the tone. The moderately inhibited network will have some tonal information but will not fully be able to suppress its response to noise.

During normal aging in the ACtx, inhibitory function decreases. As such, this creates an imbalance in which most neurons become weakly inhibited if inhibited at all, and the previously strongly inhibited neurons are now weakly inhibited (Fig. 5a). This would explain the increase in neural correlation seen in old animals, in addition to the reduction in response types found in old animals. This shift is inconsequential without background noise, as a wider range of neurons can detect the tone. However, in background noise, there are no longer strongly inhibited neurons, so fewer neurons optimally encode the tone in a noise-invariant manner. Therefore, old animals would likely have to rely on moderately inhibited neurons to detect tones in noise. As these neurons cannot fully suppress their responses to noise, they will often be active when no tone stimulus is present. As the animal has no way of knowing if the moderately inhibited neuron is active due to the tone or the noise, the animal will commonly respond to the background noise instead of the tone [22] (Fig. 5b). As such, a decrease in inhibition can explain many of the functional consequences of aging, but more complex explanations focused on a change in inhibition corresponding to a restructuring of the excitatory networks could also be true.

Fig. 5.

Aging disrupts the excitation-inhibition balance to negatively affect processing of complex auditory stimuli in the auditory cortex. Adapted from Fig. 5.1 of Shilling-Scrivo, [200]. a Model of inhibition in young and aging animals. b In a noisy environment, only the young neurons have sufficient inhibitory drive to properly encode the timing of a sound stimulus

Although the recent network-scale work supports much of the existing literature on the decline of inhibitory influence in aging, research directly linking it to network-scale loss of inhibition is lacking. Instead, most of the literature focuses on a single-neuron or whole-brain scale. Still, some research suggests that both the number and strength of inhibitory neurons decline with age [37, 41, 179]. However, emerging evidence indicates that altered inhibition may be one of many sex effects observed in aging [21, 24]. The lack of accounting for sex effects leaves only suppressed responses of excitatory neurons as indirect confirmation of network-wide loss of inhibition in male mice [21, 22] consistent with the existing aging literature, which shows an increase in response amplitude [180].

Although much debate surrounds whether the reduction of inhibition is due to a loss in inhibitory connections or simply a decrease in the strength of those connections, it appears that the inhibition loss is differentially affected by sex in that males lose inhibitory connections while females lose strength in their inhibitory connections [24]. What initiates this change remains unknown, but it appears that it can be altered via auditory training, suggesting that it is not a simple, irreversible rule of aging [23, 29–31]. However, more finely sampled time points should be used to elucidate this process throughout the aging process.

The ACtx is the primary site of complex auditory processing required for processes such as speech comprehension, but it can only process the information it receives from the periphery. Likewise, auditory training has been shown to alter cortical activity [23, 29–31], but this can only occur if the training paradigm itself is received by the peripheral auditory system. A primary method to ensure this is through the use of hearing aids, as they amplify all auditory stimuli in the environment to ensure they are received by the peripheral auditory system, even if they are degraded. The use of a hearing aid alone has been shown to induce cortical plasticity [110–112], demonstrating the crucial link between both the peripheral and central auditory systems. This further compounds the problem of developing auditory training paradigms when males and females have been repeatedly shown to have differing levels of peripheral degeneration as a result of increased noise exposure in males [99, 100, 102]. Thus, combining auditory training paradigms with increased auditory input via hearing aids may be the most effective tool in increasing cortical plasticity for both males and females.

While so far discussion has focused on changes occurring on the local scale within one ACtx, the mammalian brain has two hemispheres that work together to represent stimuli, albeit in a lateralized fashion for speech stimuli in humans [181, 182], and this lateralization is disrupted, and typically decreased, in aging, likely due to a decline in the inhibitory force placed on one hemisphere by the other [36, 183, 184]. Lateralization of auditory processing is also found in animal models [185–187] where the right hemisphere is thought to be more important for spectral processing, while the left is important for temporal processing, at least in rats [188, 189], or shows increased responses to vocalizations in mice [187]. In aging, the right hemisphere is preferentially affected via decreased neural firing synchronization with the amplitude modulation of a wide-band stimulus. However, neural synchronization increases in the left hemisphere [28], and leads to increases across both hemispheres in response to more narrow-band, tonal stimuli. This disparity between hemispheric responses to wide and narrow-band stimuli illustrates that aging involves a complex change in auditory processing and is not just a simple degradation of the respective networks. Given that the two hemispheres are coupled via inhibitory circuits [190] and show asymmetrical functional interactions [187], altered inhibition with aging could result in altered hemispheric interactions and lateralization. This small but illuminating body of work highlights the need for more cross-hemisphere studies on the effects of aging in the auditory cortices.

Although the scope of this review is the effect of aging on the ACtx, the ACtx receives numerous inputs from various regions in the frontal cortex, especially the orbitofrontal cortex [191–195]. The orbitofrontal cortex is generally thought of as being the brain region responsible for encoding reward and error [196], but it has been shown to respond to sound and cause plasticity in the ACtx [191, 192, 194, 195]. Additionally, the projections from the orbitofrontal cortex to the ACtx respond specifically to sounds and encode task-relevant information such as outcomes [194]. However, the role of top-down neural inputs to the ACtx is largely unknown, yet they may play a large role in shaping the dynamics of the excitatory network in the ACtx during aging. Indeed, the orbitofrontal cortex mediates many attentional effects that are largely altered in the aging ACtx [84, 194, 197], and the link between the ACtx and the prefrontal cortex is diminished with age [198].

In summary, the ACtx is a crucial component of the aging auditory system. While some of the difficulties associated with hearing are caused by damage to the peripheral auditory system, many hearing difficulties in aging are due to central changes. The aging central auditory system has a decreased ability to remediate the degraded peripheral signal and is prone to counterproductive changes that make it decreasingly specific when it comes to encoding all aspects of auditory stimuli, but especially the spectral and temporal attributes of a sound. However, while aging in the ACtx was previously primarily viewed as an issue that affects males and females similarly, recent research has shown this to be an oversimplification of what appear to be somewhat distinct aging processes. This sex difference makes the path of restorative neuroplasticity difficult to overcome during the aging process in the ACtx. Still, recent approaches, such as training on auditory tasks, visual deprivation, and vagus nerve stimulation, show early signs of promise. Together, the recent work on aging sheds light on a complex degradation of network dynamics; however, as summarized in this review, there are still fundamental gaps in the literature surrounding the basic mechanisms.

Author Contributions

All authors contributed to the writing of this manuscript (JKM, KVS-S, POK).

Funding

Funding was provided courtesy of National Institutes of Health grants T32NS091018 (JKM), T32DC000046 (KVS-S), RF1AG078378 (POK), P01AG055365 (POK), and R01DC017785 (POK).

Code Availability

No code was created for this manuscript.

Data Availability

No novel data were obtained for this manuscript.

Declarations

Artificial Intelligence

Artificial intelligence was not used in the preparation of this manuscript.

Competing Interests

The authors declare no competing interests.