The continued importance of comparative auditory research to modern scientific discovery

Abstract

A rich history of comparative research in the auditory field has afforded a synthetic view of sound information processing by ears and brains. Some organisms have proven to be powerful models for human hearing due to fundamental similarities (e.g., well-matched hearing ranges), while others feature intriguing differences (e.g., atympanic ears) that invite further study. Work across diverse “non-traditional” organisms, from small mammals to avians to amphibians and beyond, continues to propel auditory science forward, netting a variety of biomedical and technological advances along the way. In this brief review, limited primarily to tetrapod vertebrates, we discuss the continued importance of comparative studies in hearing research from the periphery to central nervous system with a focus on outstanding questions such as mechanisms for sound capture, peripheral and central processing of directional/spatial information, and non-canonical auditory processing, including efferent and hormonal effects.

1. Introduction

Comparative research has led to many important discoveries in the auditory field and neural systems in general, exemplified by the pioneering work of Ramon y Cajal who used a diversity of taxa for his early neuroscience studies. From small mammals (gerbils, ferrets, cats) to avians (barn owl, chicken) to amphibians (frogs, salamanders), diverse “non-traditional” model vertebrates have been leveraged to reveal common, convergent/divergent, and unique mechanisms for the reception and processing of sound information (Köppl and Manley, 2014). In some cases, mammalian models have been selected because they approximate key features of human hearing, such as audiometric range (e.g., gerbils and chinchillas), that traditional laboratory models (e.g., mice) do not (Heffner and Heffner, 2007). In other cases, comparative hearing models offer an opportunity to study processes not observed in mammals, such as the regeneration of auditory hair cells following damage (Brignull et al., 2009) or to evaluate functionally analogous mechanisms that provide insight into the complex processes underlying common auditory tasks, such as sound localization circuits in mammalian and non-mammalian tetrapods.

The comparative approach provides a framework for understanding highly varied mechanisms of hearing, including instances of structural and functional divergence and convergence across taxa. At its best, comparative work facilitates a synthetic understanding of hearing, offering deep insights on biology and evolution, prospects for biomedical advances, and avenues for technological innovation. Although contributions from invertebrate models for hearing have also yielded important insights (e.g., basal chordates and mollusks, Burighel et al., 2011; arthropods, Göpfert and Hennig, 2016), the current review will focus on contributions of non-traditional vertebrate model species to the auditory field. In this review we highlight several areas, working from the periphery to central nervous system, in which further work leveraging the comparative approach offers to move hearing science forward: diversity in mechanisms of sound capture, peripheral and central processing of directional/spatial information, and non-canonical auditory processing in which top-down control serves to modulate auditory function. While these are certainly not the only areas of continued importance for hearing research discovery, we have limited the scope of this review to these topics based on our collective expertise and based on the many outstanding questions associated with each.

2. Peripheral mechanisms for sound capture: Comparative studies reveal a panoply of pathways

Sound propagates as a longitudinal wave, giving rise to local fluctuations in pressure – a scalar quantity – but also to corresponding motion of the particles comprising the medium (i.e., acoustic particle motion) – a vector quantity – thus defining two signal components that ears might detect (reviewed in Popper and Hawkins, 2019). The particulars of sound propagation vary considerably across acoustic media, impacting transmission speed (approximately 340 m/s in air and 1450 m/s in water), and the form and extent of interaction with surfaces including ears and heads. To detect airborne sound, the ears of most terrestrial tetrapods leverage a tympanic receiver that moves in response to local fluctuations in sound pressure, efficiently transmitting the received signal to inner ear structures and ultimately the auditory hair cells. However, the earliest form of hearing relied on direct reception of particle motion generated by sound sources. This ancestral ‘inertial’ mode of hearing remains the dominant form of hearing in the largest extant group of vertebrates, the teleost fishes, as well as the elasmobranchs and their relatives. Underwater, particle motion is easily transmitted through the tissues of the fish’s head to the end organs of the ear because of the similar density of tissue and water (e.g., Taylor 1921). Comparative work has highlighted similarities across taxa that suggest atympanic ‘tissue-borne’ forms of hearing may be a relatively versatile means of sound reception, with analogous modes of auditory stimulation in evolutionarily ancient terrestrial tetrapods such as the tuatara (Gans and Wever 1976), but also more recently derived species (including many frogs) that exploit atympanic mechanisms (Capshaw et al., 2022; Christensen et al., 2012; Womack et al., 2017, 2018).

While the problem of sound reception is thus well-defined, additional comparative work is necessary to answer fundamental questions about sound capture and transduction through complex structures (e.g., pinnae, middle ear spaces and associated cranial anatomy). Increased accessibility of tools to non-invasively probe anatomy and physiology, including auditory brainstem response recordings, computed tomographic imaging, and finite element modeling, will allow future work to incorporate diverse species and more readily test explicit structure-function hypotheses by virtually isolating and manipulating hearing structure properties (as in Livens et al., 2020; Muyshondt et al., 2016, among others). Here we consider the progress and potential of comparative work in expanding our understanding of acoustic reception by vertebrate ears. In this section, we outline insights gained from previous studies of external and middle ear variation and identify outstanding questions about sound transduction from the external environment to the inner ears.

2.1. On the diversity of tympanic ears

The vertebrate ear is highly diversified across and within taxa, with significant structural variation influencing every stage from reception to transduction. Most terrestrial species detect airborne sound using a specialized peripheral receiver, the tympanic ear, consisting of a thin membrane that is coupled to the inner ear via one or more middle ear bones spanning an air-filled cavity. Comparative work has made clear the efficacy of the tympanic ear in terrestrial hearing, with evidence that tympanic ears evolved independently in all major tetrapod clades (reviewed in Manley, 2017; Figure 1). However, divergent evolutionary development of tympanic ears has generated a great deal of variation, including in the presence, absence or elaboration of external structures (e.g. pinnae), differing structures and mechanics of communication between the tympana and the auditory end organ(s) of the inner ear, and variation in the nature and extent of inter-tympanic coupling (see also section 3.1). Parallel evolution of tympanic middle ears among terrestrial vertebrate lineages has generated many structural and functional analogues in the auditory periphery that reveal fundamental characteristics of aerial sound reception, revealed by cross-taxa study. The rich evolutionary history of the vertebrate ear provides natural case studies that can be leveraged to understand different mechanisms and solutions for resolving auditory ambiguities (e.g., front-back confusions of sound source location) and enhancing responsiveness to relevant stimuli (e.g., contributions of the external auditory canals and middle ear bones to resonance and gain at the peripheral level of the auditory system). Comparative work has improved our understanding of the effects of structural variation of the auditory periphery on hearing and highlights fundamental unanswered questions about mechanisms for sound reception and transmission through complex structures.

Figure 1: Tympanic middle ears vary among tetrapods and are hypothesized to have arisen evolutionarily at least three times.

A simplified cladogram representing the evolutionary relationships among the four main tetrapod groups (Amphibia, Lepidosauria, Archosauria, and Mammalia). Species diversity and the estimated percent of acoustically communicating species are displayed above each group (data derived from Chen and Wiens (2020)). Black ovals indicate the independent evolutionary origins of tympanic middle ears. Top diagrams show a glimpse of middle ear diversity seen across tetrapods. Abbreviations: AB: auditory bulla, EAC: external auditory canal, ES: extrastapes, ET: Eustachian tube, IE: inner ear, In: incus, LJ: lower jaw, Ma: malleus, MEC: middle ear cavity, Op: operculum, Qu: quadrate, St: stapes, TM: tympanum,

2.1.1. The external ear: pinnae and external auditory canals

In therian mammals, i.e. marsupials and placental mammals including humans, sound reaches the tympanum via an elaborate external ear, the pinna, that collects and funnels airborne sound into an external auditory canal. The mammalian pinna is highly diversified across species with a wide variety of shapes, sizes, and degree of folding, as well as a range of motility from nearly static to highly mobile. To what extent is this variation useful for extracting information from acoustic stimuli? Examination of ear structure and auditory performance in diverse species indicates that, although much of the inter- and intraspecific diversity follows ecogeographic rules for thermal physiology (e.g. Allen’s Rule - Allen 1877; Alhajeri et al., 2020), pinna variation has important consequences for hearing. For example, the shape and folding of the pinnae generate location cues for high frequencies that enable mammals to determine sound source elevation (Calford and Pettigrew, 1984; Musicant and Butler, 1984; Heffner et al., 1996; Rice et al., 1992). Shared reliance on pinna-based sound localization cues among diverse mammalian species supports the importance of spectral filtering by these specialized external structures to auditory processing – a feature that is most apparent in the size and elaboration of the pinnae of echolocating bats (Jen and Chen, 1988). The pinnae also create binaural cues to location that are larger in magnitude and lower in frequency than the head produces alone, a feature of particular importance to small mammals. For example, in chinchilla the head alone produces meaningful interaural level difference (ILD) cues (> ~5 dB) only for frequencies greater than 10 kHz, but the head with the pinnae produces these ILDs for frequencies of 4 kHz and above (Jones et al., 2011). In the rat, the pinnae create interaural time difference cues that are ~36% larger than those produced by the head alone (Koka et al., 2008). Given the large variability in pinna size and shape, augmentation of binaural cues across mammalian taxa is likely to be complex and varied; additional work in this area could point to common themes and translational insights (e.g., to inform the design of hearing devices and acoustic sensors that generally lack ‘pinnae’).

Aside from humans, most mammals, particularly highly mobile predators, can and do move their pinnae individually and quite extensively. In the cat, a nocturnal predator with exceptional sound localization abilities (Tollin et al., 2005), 22 muscles are dedicated to moving the pinnae (Populin and Yin 1995). In contrast, eye position is controlled by only six muscles. Such highly mobile pinnae enable species to actively alter their acoustic axis in response to sound source location, a phenomenon also studied in diverse echolocating bat species (e.g., Lawrence and Simmons, 1982; Wotton et al., 1995; Müller et al., 2008; Chiu and Moss, 2007) and inspiring numerous insights into the technological development of sonar sensing devices (e.g., Guarato et al., 2015; Schillebeeckx et al., 2011; Yin and Müller, 2021). Mobile pinnae also create additional complexities regarding our understanding of sound source localization cues, physiology and behavior. Moving the pinna on the head dramatically changes all of the cues to sound location (Young et al., Re 1996), yet sound localization ability remains unaffected by these altered cues (Populin et al., 2004). Remarkably, this implies that there is not a single set of cues that specify sound location relative to the head and that, somewhere in the auditory pathway, the impact of pinnae position on the head must be accounted for. The beginnings of such compensation may be found in the dorsal cochlear nucleus (e.g. Kanold and Young 2001), but additional comparative work examining differences in neural circuits in mammals that can and cannot move their pinnae could yield valuable insight as to where and how such transformations occur.

Additionally, mobile pinnae can actively track sound sources despite movements of the head via a mechanism called the vestibular auricular reflex (VAR) (Tollin et al., 2009). Tollin et al. (2009) proposed that compensatory movements like the VAR allow the auditory spatial scene to be stabilized during head movements in the same way that the vestibular ocular reflex (VOR) stabilizes the visual world. Here, the VAR functions to keep the attentional direction fixated at a particular sound source regardless of movements of the head. Like the classical VOR, the VAR would play an important function to stabilize the auditory world of predators like the cat, despite the rapid movements of the head that occur during the active pursuit of prey. Given the clear advantage conferred by an ability to dynamically alter the shape and orientation of the pinna in response to sound location, it remains unclear why some species, notably humans but also other anthropoid primates, have reduced pinna mobility and therefore must rely on location cues generated by the static spectral shape of the pinna (and head). Still other species, such as horses, have highly mobile pinna yet show reduced localization abilities (Heffner and Heffner 1984). Continued comparative study of spatial hearing in species that vary in these parameters will further reveal the relative contributions of dynamic cues contributed by the movement of the pinna and head versus static cues intrinsic to pinna shape.

In contrast to mammals, most non-mammalian tetrapods have a comparatively simple external ear consisting of a shallow external auditory canal leading to the tympanum as in birds, or completely absent external ears as in amphibians and most squamates. Even among mammals, reduced pinnae are evident in subterranean species (e.g., naked mole-rats and blind mole-rats), co-occurring with reduced sensitivity to high frequencies (Pyott et al., 2020), reduced pathways for sound processing in the auditory brainstem (Gessele et al., 2016; McCullagh et al., 2022), and reduced localization abilities relative to other mammals (Heffner and Heffner 1992, 1993). Among non-mammalians that are sensitive to high-frequency sound and that may rely on high-frequency information for important tasks including source localization, some external adaptations are evident that may represent functional analogues to the therian external ear. For example, among avians, barn owls show external modifications to the auditory periphery including facial ruffs and bilaterally asymmetric ear openings that generate location cues analogous to pinna cues, particularly in the vertical plane (Norberg, 1977; Knudsen and Konishi, 1979; Coles and Guppy 1988). Within amphibians, some frog species have recessed tympana that are analogous to ear canals, a putative adaptation for high frequency reception of their ultrasonic mating calls (Arch et al., 2008; Cobo-Cuan et al., 2020; Feng et al., 2006; Feng and Narins 2008).

But high-frequency hearing is generally uncommon among non-mammalian tetrapods, and low-frequency (long-wavelength) sound is not effectively altered by comparatively small external ear anatomy. In this context, other tetrapod ‘external ear’ structures may serve primarily to regulate the amplitude of sound reaching the tympanum, or to offer a level of protection for the tympanum and other delicate auditory structures. For example, crocodilians have muscular ‘earlids’ that cover the tympanic membrane and may be opened or closed to modify interactions of the tympana with the surrounding media (Shute and Bellairs 1955). Medial to the earlids is an enclosed meatal chamber that terminates at the tympana, analogous to the external auditory canal of mammals. The association of these external ear structures with terrestriality among basal crocodilians could represent anatomical specialization for sound reception (Montefeltro et al., 2016); however, the contributions of these structures to hearing in crocodilians is unknown and in need of further study. Many open questions remain about the significance of structural variability in external ears among vertebrates and how this may correlate with functional variation. More thorough study incorporating diverse species will expand our understanding of how different size and shape parameters of the external ear structures, as well as their mobility (or lack thereof), contribute to auditory specialization and how this may influence auditory processing at higher levels.

2.1.2. The tympanic membrane

Despite morphological and functional variation of external ears among vertebrates, all tympanic ears share a common functional output: the detection of sound pressure via a specialized receiver, the tympanic membrane (tympanum, pl. tympana). Although structurally similar across taxa, the tympana of terrestrial vertebrates represent evolutionary convergence following independent evolutionary derivations of the tympanic middle ear. The effects of tympanic membrane variation across taxa (e.g., differences in size, shape, and stiffness of the membrane) on its functional output as the peripheral pressure receiver represents an important question in terrestrial sound capture mechanisms. In amniotes, the evolution of the tympanic membrane is strongly linked to the evolution of the jaw joint, illustrated by variation in its relative position to the upper and lower elements of the jaw. The tympanum is attached to the tympanic ring, a homolog of the angular bone of the lower jaw, in synapsids, and to the quadrate bone of the upper jaw in diapsids. Comparative developmental work using mouse and chick models showed that, although the tympana derive from the first pharyngeal pouch in both groups, topographical shifts in the relationship of the jaw joint and first pharyngeal pouch led to differential development of the tympana in these lineages (Kitazawa et al., 2015). Differential interaction of the tympana with the middle ear bone(s) in synapsids and diapsids influences the overall shape of the membrane, resulting in a concave or convex curvature, respectively. These curvature differences are predicted to influence middle-ear sound transmission and amplification (Fay et al., 2006; Funnell and Lazlo 1978; Muyshondt and Dirckx 2020), but more comparative work is needed to test these predictions in natural systems.

In general, the size of the tympanic membrane scales allometrically in vertebrates. A broad range of body sizes among vertebrates generates diversity in tympanum size across species that may influence its acoustic response properties and therefore its functional output. Examination of diverse taxa, particularly those like frogs that rely on acoustic communication and encompass a broad range of body sizes, provides the opportunity to explore the effects of scaling on the acoustic response of the tympanum. Although there is some evidence that tympanic membrane size is positively correlated with low- and high-frequency hearing sensitivity in frogs (Fox 1995), other studies suggest this correlation may be due to variation in body size versus more specific auditory variation (James et al., 2022). Studies in dogs have shown that sensitivity does not vary with tympanic membrane size, casting additional doubt on the effect membrane size has on hearing thresholds (Heffner 1983). Despite this, diminutive frog species provide a clear demonstration of a lower-size limit on tympanic function in which miniaturized tympana are less responsive to sound pressure than undifferentiated skin (Hetherington 1992), likely indicating that terrestrial hearing in these small species is more reliant on non-tympanic pathways (discussed below). Size-effects on hearing may therefore reflect changes to tympanic membrane flexibility or stiffness, which can influence the overall impedance of the system in a frequency-specific manner. Although studies have attempted to address the role of tympanic membrane properties on hearing sensitivity, many include confounding factors, like body size, that make interpretation difficult. More studies that explicitly sample with confounding factors in mind or that model tympanic membrane properties in isolation are needed.

The detailed internal structure of the tympanic membrane also varies across taxa. The mammalian tympanum is the most well-studied, and is divided into two parts based on morphological and biomechanical differences: the stiff, collagen rich pars tensa which interacts with the middle ear ossicles, and the triangular, elastic pars flaccida that represents a small portion of the tympanic membrane in humans (Lim, 1995). The relative size of these structures varies among mammals ranging from complete absence of the pars flaccida in chinchillas, hamsters, and red squirrels, to equal ratios of each structure in Mus musculus, the traditional laboratory mouse model, to the oversized pars flaccida of goats and pigs that takes up more than half of the tympanum area (Kohllöffel, 1984). Any functional consequences of this variation are largely unknown, and comparative studies could reveal the contributions of these mechanically distinct components of the tympanum to sound conduction and pressure regulation within the mammalian middle ear. Further, the pars tensa and flaccida differ in their susceptibility to various congenital and acquired pathologies in humans (reviewed by Mozaffari et al., 2020), highlighting the utility of comparative study that takes advantage of the natural diversity of tympanic membrane composition across mammalian species to better understand the influence of tympanum structure on disease etiology in humans.

2.1.3. The middle ear bone(s)

The vertebrate middle ear bones represent another fascinating case in diverse mechanisms for sound reception, and their study has informed bio-inspired strategies for hearing restoration in humans. The ancestral tetrapod ear had a single middle-ear ossicle, the stapes (or columella), that derived from the hyomandibular bone of the jaw suspensorium. This ossicle was non-functional in the early tetrapods, and a functional, movable middle-ear apparatus originated independently in mammals, archosaurs, anurans and lepidosaurs (Clack 1997). Many extant tetrapod vertebrates retain the single-ossicle middle ear configuration, including amphibians, archosaurs, and squamates; however, evolution of tympanic middle ears among these lineages has generated diversity in the size and shape of the stapes as well as in its interaction with adjacent cranial bones (Fig. 1). Morphological variation of the middle-ear ossicle is most notable across non-mammalian species in which the stapes ranges from a slender, rod-like bone in species with tympanic ears, like birds, crocodilians, frogs, and most lizards, to a bulky structure that is often reduced to a footplate in the oval window of atympanate species, such as snakes, amphisbaenians, salamanders and caecilians. This variation reflects differences in the sound transmission pathways among tympanate and atympanate species, in which the absence of a tympanic ear precludes high-frequency sound transmission and restricts hearing to low-frequency sound and vibration detection.

Despite its relative simplicity compared to the three-ossicle mammalian middle ear, the single-ossicle middle ear configuration of non-mammalian tetrapods demonstrates flexibility (e.g., at the stapes-extrastapes articulation of frogs, some lizards and birds) that enables a lever action about a rotatory axis to increase force applied by the stapedial input to the cochlea (Mason and Farr, 2012). In birds, joint flexion between the stapes and extrastapes generates a rocking motion of the footplate in the oval window rather than a piston-like motion, increasing middle-ear pressure gain at low frequencies while providing protection against potentially damaging high-amplitude displacements (Mills and Zhang, 2006; Muyshondt et al., 2019; Muyshondt and Dirckx, 2020). The mechanics of the single-ossicle middle ear are simple to reconstruct and manipulate, making it an appealing design for total ossicular replacement prosthetics (Manley, 2021). Indeed, bio-inspired “bird-type” prostheses that mimic the avian single-ossicle configuration show great promise for restoring sound transmission to the ear in human cases of conductive hearing loss (Arechvo et al., 2013; Beleites et al., 2007; Rönnblom et al., 2020). Recent comparative studies across a large number of taxonomically diverse species of bird of the morphology of the columella (Peacock et al., 2020a) and the functional consequences of these differences (Peacock et al., 2020b) could yield the kinds of insights needed to design the next generation of ossicular prostheses.

In Mesozoic mammaliforms, a novel jaw joint emerged that enabled the articular and quadrate bones of the jaw to become specialized for sound conduction and join the stapes to form the three-ossicle configuration of the mammalian middle ear (reviewed by Luo, 2011). Developmental and fossil evidence surveying diverse taxa support two independent derivations of the middle ear bones in therians and monotremes (reviewed by Anthwal et al., 2013). Structural variation of the middle ear ossicles across different mammalian species affects their mobility in response to acoustic input and therefore influences the frequency response of the middle ear. For example, the presence of a bulky orbital apophysis on the malleus in small mammals such as mice and bats increases its rotational inertia and co-occurs with high frequency specialization (Fleischer, 1978; Mason 2013). Conversely, radiation into subterranean habitats is correlated with convergent enlargement of the middle ear ossicles that are inferred to improve low-frequency sound transmission and bone conduction hearing (Burda et al., 1992; Mason 2001, 2003; Koyabu et al., 2017). Continuing comparative study of middle ears in diverse mammalian species can potentially reveal different strategies for enhancing transmission of acoustic energy to the inner ear, particularly among high- and low-frequency specialists. Additionally, comparative study of the effects of variable size and configuration of the middle ear ossicles among mammals on the middle-ear acoustic transfer function could contribute to our understanding of the perceptual effects of congenital and acquired pathologies in humans such as middle ear bone deformities, fixation, and/or fusion that influence the mass and flexibility of the ossicular chain.

Recent advances in imaging technology have facilitated in-depth study of structural diversity in vertebrate middle ears. In particular, non-destructive micro-computed tomographic imaging, often paired with novel contrast methodologies to visualize soft tissue structures in addition to bone (e.g., diceCT methods reviewed by Gignac and Kley, 2018) have provided the opportunity to sample diverse species that have historically been difficult to access. Comparison of high resolution reconstructions of the middle ear in echolocating bats and cetaceans, for example, has revealed a novel placement of the stapes and oval window that may influence stapedial input to the cochlea in certain species (Ketten et al., 2021). The stapes generally inserts near the hook region at the base of the cochlea in most non-echolocating mammalian species. However, the stapes-oval window placement in two echolocating bat species and the harbor porpoise is shifted relative to the basal hook region – a convergent morphological change that could influence traveling wave dynamics within the cochlea (Ketten et al., 2021). The functional implications of species variation in oval window location have not yet been investigated, although Ketten et al. (2021) have proposed the intriguing idea that the unique placement of the stapedial input to the cochlea could generate standing waves that enhance frequency selectivity in echolocators. Future work could leverage finite element modeling or optical coherence tomography, another non-destructive technique, to investigate cochlear traveling wave dynamics in these species and provide further insights into the physiological and biomechanical effects of spatial variation in location of the middle ear input to the cochlea.

2.2. Atympanic hearing: Bone conduction and related pathways

Tympanic ears enable auditory systems to detect minute fluctuations in air pressure associated with even very faint sounds (whether emanating from distant sources or low-intensity sources nearby). The selective advantage of detecting such sounds is clear, and a loss of key structures of the tympanic ear is associated with higher detection thresholds for airborne sound. However, provided the inner ear is intact, hearing remains possible via the bone conduction pathway (or related forms of tissue-borne conduction) in which sound is transmitted directly to the inner ear via vibrations traveling through the bones and tissues of the head. Much of our knowledge of bone conduction hearing comes from work in human listeners (e.g., von Bèkèsy 1948, 1949; reviewed in Stenfelt, 2011) but all vertebrate ears are likely sensitive to bone conducted stimulation. Indeed, recent comparative work has highlighted similarities across taxa that suggest atympanic sound reception may be evolutionarily persistent, and an important target for future research to advance a synthetic understanding of auditory function.

Tympanic middle ears have been evolutionarily lost in many species, with multiple loss events evident in some clades (e.g., amphibians and squamates - Pereyra et al., 2016; Baird 1970; Wever 1973; Capshaw et al., 2022). In some “earless” atympanate species, relaxed selection on airborne hearing may be correlated with environmentally imposed challenges to sound reception, such as subterranean habitats in which acoustic transmission of high frequencies is limited (e.g., Wever 1973) or environments with high levels of masking noise (Arch et al., 2011). In other atympanate species, tympanic middle ear loss underscores the ability of organisms to transfer sound using non-tympanic pathways, such as via bone conduction or through secondary inputs such as adjacent air-filled cavities that augment sound transmission to the inner ear – e.g., the lung-ear pathway in tympanate frogs that serves as a band-pass filter to modulate the frequency response of the tympana (Lee et al., 2021) and could function similarly in some atympanate frog species (Lindquist et al., 1998; Hetherington and Lindquist 1999; but see also Womack et al., 2018). For example, as suggested in Sec. 2.1., small-bodied frogs demonstrate scaling limitations to the efficacy of sound reception by the tympanic structures of the ear. Impinging pressure waves generate equivalent or greater vibration velocities in the undifferentiated skin surfaces of miniaturized frog species relative to their tympana (Hetherington 1992). In these and other species such as salamanders (Capshaw et al., 2020) and snakes (Christensen et al., 2012), the small acoustic size of the head (i.e., its size relative to the wavelength of the impinging sound) enables hearing via a general mode of bone conduction in which vibrations of the animal’s head are detected by the auditory system (Capshaw et al., 2022). This mechanism for sound reception requires no specialized receptive surfaces and can convey directional information (Capshaw et al., 2021), therefore, similar bone conduction pathways could represent an evolutionarily ancient pathway for hearing in ancestral atympanate tetrapods. Interestingly, there is evidence that species lacking tympanic middle ears behaviorally orient to sound (Lindquist and Hetherington 1996; Diego-Rasilla and Luengo 2004, 2007; Madden and Jehle 2017). Although this remains understudied in atympanate terrestrial vertebrates, especially among the many species of earless frogs that retain acoustic signaling behaviors, a rich literature on atympanate aquatic vertebrates, particularly the teleost fishes which hear via inertial sensing of acoustic particle motion (subject to modification by secondary structures; reviewed in Hawkins and Popper 2018), suggest that tympanic ears are not necessary for directional hearing. Further research on atympanate species could reveal mechanisms of directional hearing via bone conduction and provide a clearer picture of how this sound transmission pathway could have influenced the early evolution of the terrestrial vertebrate auditory system. Additionally, understanding how this non-tympanic directionality is processed could inform the design of improved sensors and bone conduction hearing devices (Håkansson et al., 1985; Caspers et al., 2022).

2.3. Amphibious hearing: Sensing across acoustic media

The tympanic middle ear is often seen as an adaptation for hearing airborne sounds, but several recent studies of amphibious or aquatic tetrapods (Christensen-Dalsgaard, 2011; Christensen-Dalsgaard and Elepfandt, 1995; Larsen et al., 2020) have shown that minor changes to the tympanic ear can turn it into an efficient receiver for underwater sound. Earlier, it was supposed that hearing underwater could function efficiently by bone conduction because the similar impedance of water and animal tissues would ensure an efficient transmission of energy to vibrate the inner ear fluids without any accessory middle ear structures. However, the particle motion cues that stimulate the ear are very small because of the higher impedance of water. Rather, a common factor in the investigated species is that they have a stiffer tympanum (a cartilaginous, disk-like structure in the red-eared slider turtle and the clawed frog) and an air-filled middle ear cavity. Underwater, the tympana of these animals move with higher amplitudes than the water particle motion, and the tympanic motion is probably driven by resonance of the air in the middle ear cavity, analogous to the amplification of motion cues by fish swim bladders (Christensen-Dalsgaard et al., 2012). A recent study comparing human in-air and underwater hearing thresholds (Sørensen et al., 2022) is interesting in this respect, since it shows that, in a purely terrestrial tetrapod, sensitivity to underwater sound is still higher than expected from bone-conduction alone, potentially by virtue of the air-filled middle ear cavity (though so-called ‘soft tissue’ pathways may also contribute, e.g. Chordekar et al., 2015; Sohmer 2017).

In most amphibious animals there is a trade-off between airborne and underwater hearing due to the adaptations for enhancing underwater sensitivity, as shown in cormorants (Larsen et al., 2020), red-eared slider turtles (Christensen-Dalsgaard et al., 2012), and clawed frogs (Christensen-Dalsgaard et al., 1990, Christensen-Dalsgaard and Elepfandt 1995), where thresholds in air are higher than non-amphibious species. Pinnipeds represent an interesting contrast in this context, because they are very sensitive to both airborne and underwater sounds (e.g., Kastak and Schusterman, 1998; Reichmuth et al., 2013) suggesting that they may be able to change the properties of the middle ear when moving between land and water. Further research is necessary to understand the underlying structural adaptations and control mechanisms that may facilitate sensitive hearing across media. Further study of amphibious hearing could also provide key insights for the improvement of underwater and amphibious hearing devices and acoustic sensors for divers (e.g. Smith 1969; Shupak et al., 2005).

3. Divergence and convergence in mechanisms for directional hearing

The localization of sounds generated by predators and prey is an important task of the auditory system, and even the earliest auditory systems of vertebrates were probably able to process sound direction. Here we consider the diversity of directional ears revealed by comparative work, but also evidence for convergence in central mechanisms for directional processing.

3.1. On the diversity of directional ears

In most terrestrial vertebrates, sound direction is processed in the central nervous system by comparison of the sound spectra reaching the two ears and by analysis of the binaural differences determined by sound propagation velocity and diffraction caused by the head and body of the animal. However, the origins of these spectral and binaural differences depend on the configuration of the ears. There are three major ear configurations in vertebrates. One is the non-tympanic ear found in chondrosteans and most osteichthyans, including a large number of tetrapods (Capshaw et al., 2022, see section 2.2), that respond to particle motion and are inherently directional (Fay, 1984). The other two major types are different configurations of tetrapod tympanic ears (Christensen-Dalsgaard and Carr, 2008). Most tetrapods have ears coupled across the mouth cavity. This may be the ancestral state in non-mammals: early tympana formed at the spiracular openings, leading to a middle ear cavity that was broadly confluent with the mouth cavity (Clack, 1997; reviewed by Christensen-Dalsgaard, 2011; van Hemmen et al., 2016). Depending on the efficiency of sound transmission from one eardrum to the other, the interaction of sound at the tympana produces robust frequency-dependent binaural cues, resulting in large time and amplitude differences between the motion of the two tympana. Directionality by way of binaural coupling, similar to the configuration observed in frogs and lizards, has inspired the engineering of several directional sensing devices (reviewed in Christensen-Dalsgaard and Manley, 2019; Rahaman and Kim, 2022), including a Braitenberg-type robot that integrates amplitude differences across two parallel nanowires to identify sound source location (Shaikh et al., 2009), and a photosensing device that exploits an analogous coupling of optical resonators to determine incident angles of light (Yi et al., 2018).

In the third tetrapod ear configuration, sound transmission is highly attenuated between the two middle ears, which functionally isolates them. This configuration is chiefly found in mammals and in specialized birds like the barn owl. Here, the subsequent processing by the auditory pathway focuses on time delays caused by the time-of-arrival differences at the two tympana (interaural time difference - ITD) and on the difference in eardrum vibration amplitude caused by sound shadowing of the head and body leading to interaural level differences (ILD) between the two ears. These binaural cues are also frequency-dependent: the neural representation of ITD depends on neural phase locking that degrades with frequency, and conversely, the sound shadowing generating ILD depends on the acoustic size of the animal and is often found at relatively high frequencies, although large ILDs are also observed at low frequencies for nearby sources (e.g. Jones et al., 2015). While ITD and ILD are understood to be separate cues and were traditionally thought to be encoded (at least initially) in separate pathways, accumulating evidence suggests a more complex account (Grothe et al., 2010; Joris and Trussell 2018; van der Heijden et al., 2019; see Sec. 3.2.1.).

3.2. Convergence on opponent-channel coding schemes

Despite the different types of ears, comparative studies support the idea that many auditory systems may converge upon similar opponent-channel coding schemes. In barn owls, responses consistent with the opponent-channel theory of sound localization emerge in premotor neurons of the owl’s midbrain (Cazettes et al., 2018), while in mammals, a transformation from a place code to an opponent-channel code has been proposed for midbrain control of eye movements (Groh et al., 2001). In barn owls, midbrain spatial maps (Knudsen and Konishi, 1978) are transformed into non-topographic representations in forebrain (Beckert et al., 2017; Reches and Gutfreund, 2009; Vonderschen and Wagner, 2012). In geckos, which have directional coupled ears, passive binaural processing occurs already at the eardrum, and gecko auditory nerve responses closely resemble a binaural opponent-channel coding scheme (Fig. 2, Christensen-Dalsgaard et al., 2021). The overlap between lizard, bird and mammalian coding strategies suggests unifying principles in binaural processing (Peña et al., 2019). A recent study also suggests that the basis for auditory spatial adaptation may be comparable in ferret and owls (Keating et al., 2015). These findings account for growing evidence connecting neural coding of auditory spatial cues across species, highlighting the importance of comparative studies to reveal similarities and differences in coding schemes based on peripheral and central nervous system features. Important questions also remain regarding similarities and differences in the central representations of the two canonical binaural cues, ITD and ILD.

Figure 2: Directional hearing.

A, Averaged ITD tuning curves in midbrain tegmentum reveal neural responses like the opponent-channel code hypothesized in mammals in left (red) and right (blue) hemispheres. The curves cross at the midline and the steepest slopes are positioned within the physiological range (black dashed lines). Modified from Peña et al., 2019 and Cazettes et al., 2018. B, Directional responses of auditory nerve fibers in the gecko. Combination line and histogram plot of the distribution of spatial receptive field sizes for all fibers. Recordings shown from the right auditory nerve (red, data were reflected to simulate left side responses in blue). From Christensen-Dalsgaard et al., 2021.

3.2.1. Simultaneous sensitivity to ITD and ILD in the lateral superior olive

In the mammalian auditory system, the encoding ITD and ILD is classically thought to be handled by two separate neural circuits: ITDs by neurons in the medial superior olive (MSO), and ILDs by neurons in the lateral superior olive (LSO). There have been numerous reviews in recent years (e.g., Joris and Yin 2007; Grothe et al., 2010) on the encoding of ITDs in the MSO, so this topic is covered only briefly. Here, we instead focus on the encoding of ILDs and ITDs by LSO neurons. This narrow focus is in response to relatively recent but accumulating evidence suggesting that intricate specializations in the neural circuits responsible for the initial encoding of ILDs by the LSO allows these neurons to also be exquisitely sensitive to all types of ITDs, including ITDs conveyed by temporal fine structure, amplitude envelopes, and transient stimuli (Franken et al., 2018; Joris and Trussel, 2018; Franken et al., 2021; Owrutsky et al., 2021).

The results of the Franken et al. (2018; 2021) studies and others (Tollin and Yin, 2005) suggest that LSO neurons encode the timing of sounds at the two ears very similar to those of the MSO: they both act as coincidence detectors. Yet, in contrast to MSO, the LSO detects coincidences of excitation from one ear and inhibition from the other. Rather than conveying information about ITDs through changes in firing rates, LSO neurons do so via shifts in the delay of the interaction between excitation and inhibition, which are dependent on both the ITD and the relative intensity of the sound between the ears (i.e., the ILD) (e.g., Joris and Yin, 1995; Tollin and Yin, 2005). This has been supported by computational models of LSO that detect the coincidence of excitation and inhibition from the two ears and was recently shown to effectively create sensitivity to ITDs for high-frequency sounds while also preserving sensitivity to level differences (Ashida et al., 2016). An interpretation of these results is that the circuit provides ILD sensitivity as a byproduct of exquisite ITD coding via an anti-coincidence mechanism. Consistent with this, there is also evidence that additional (including de novo) neural extraction of ILDs occurs at more central nuclei in the auditory system than the LSO (Li and Pollak, 2013; Tsai et al., 2010). In particular, the neurons in the auditory midbrain integrate inputs over longer time windows, long enough to compute the ILDs in a way consistent with empirical behavioral sensitivity to ILD – something that the LSO is too ‘fast’ to do (e.g., Brown and Tollin, 2016).

So what does the LSO do? One possibility is that the LSO allows the detection of sound sources away from the midline, while MSO neurons are modulated maximally by sound sources in front of an animal, where the gaze is directed. For sources with small or near zero ITDs, LSO neurons are completely or nearly completely inhibited. As a sound moves laterally to the side of an animal, the firing rates of MSO neurons are modulated less, but those of the LSO become more optimally modulated. A differential comparison of MSO and LSO firing rates could potentially give an index of the laterality of a sound source that is much less influenced by changes in firing rates that occur simply due to changes in sound source level (see Tsai et al., 2010). This could yield a level-independent code for horizontal sound location. For small-headed mammals such as bats, mice, rats, etc., where there is uncertainty whether the MSO is binaurally functional, the LSO could act as a mechanism to direct gaze to sound sources. For a laterally located source, the head could be moved until the LSO responses from the two sides are effectively nulled.

A fuller understanding of LSO function will likely require across-species comparisons. It is well known that the relative size of the MSO and LSO, and the numbers of neurons comprising them, varies significantly across mammalian species (Glendenning and Masterton, 1998). Masterton et al. (1975) exploited this knowledge in an approach dubbed “natural ablation”. They reasoned that behaviors and physiological responses that rely on a particular nucleus might be expected to scale with its size across species. For example, mice and rats have large and well developed LSOs but small or nonexistent MSOs (Fischl et al., 2016). Consequently, both species likely can use high-frequency ILDs and transient ITDs (Li et al., 2019, in rat), but neither can use low-frequency ITDs for sound localization (Heffner et al., 2001). Other low-frequency hearing mammals (including cats, chinchillas, guinea pigs, gerbils, and humans) have both LSO and MSO nuclei and use both high-frequency ILDs as well as low-frequency, high-frequency, and transient ITDs for sound localization. Recent studies have leveraged this “natural ablation” approach to demonstrate that the origin of a binaural auditory evoked potential likely arises from the LSO because the morphology of the potential waveform and the way that binaural cues to location modulated the amplitude of the potential were conserved across species ranging from mice and rats to primates and humans (Benichoux et al., 2018; Peacock et al., 2021).

3.3. Central adaptations for encoding natural sound statistics in directional hearing

Across species, sound localization is generally better for sounds that come from sources in front of the animal than from sources by their sides. This may be due in part to the number of neurons available to detect sounds from each location (Colburn 1977; Stern and Colburn 1978), differences in sound-evoked neural firing rates (McAlpine et al., 2001; Harper and McAlpine 2004), or how filtering of incoming sounds by the head and external ear determines patterns of changing binaural cues across locations (Grothe et al., 2010; Mills 1972; Harper and McAlpine 2004; Yost 1974; Feddersen et al., 1957; Brown et al., 2018; Kuhn 1977; Stern and Colburn 1978). Recent studies using barn owls to test whether enhanced localization of sounds in the front represents optimality in spatial perception have shown that the owl’s brain has evolved the ability to anticipate natural statistical patterns of ITD that exist in sounds as a result of their frequency and location (Fischer and Pena 2011; Cazettes et al., 2014, 2016, 2018). Further studies showed that humans also use these natural statistics to optimize the spatial perception of sounds (Pavão et al., 2020).

ITDs for sound sources located in front of a listener are generally more informative (sharper modulation across azimuth, less variance across other stimulus dimensions) than those for lateral sources, although these patterns may vary across frequency (Kuhn, 1977; Benichoux et al., 2016; Pavão et al., 2020). The same holds true (with more complex frequency dependence) for ILD cues (Brown et al., 2018). Human spatial discrimination across frequency investigated using free field stimulation (Mills 1958) is consistent with anticipated natural statistics of ITD and ILD cues, although magnitude-dependent variation in the resolution of sensitivity to such cues may also be important (Brown et al., 2018; Pavão et al., 2020). Playing tonal sounds to healthy humans while removing natural statistical differences using inserted earphones also showed that a person’s ability to tell where a sound was coming from was correlated with combined anticipated statistics across frequency and location (Pavão et al., 2020). Pavão et al. (2020) analyzed both the rate of change and the variability of ITD cues and found that natural stimulus ITD statistics were good predictors of auditory spatial perception. Thus, the auditory system anticipates natural features of the auditory scene to facilitate sound-source localization. Improved understanding of these features could in turn facilitate the development of hearing devices that leverage natural binaural cue statistics to efficiently transmit the most salient information. These observations also suggest that spatial hearing, like other spatial senses including vision (e.g., Ma et al., 2011), may have evolved to respond near-optimally to the environment (Harper and McAlpine 2004; Pavão et al., 2020).

4. Non-canonical auditory processing

In addition to the need for continued comparative work on the ascending auditory pathway extending to the most peripheral elements highlighted in the foregoing sections, the descending auditory system, and other neglected pathways for auditory modulation, provide fertile ground for further study. Here in particular, comparative work is essential to parse the origins and elucidate the functional significance of these factors across taxa and to push the field of auditory neuroscience toward a fuller understanding of the dynamic nature of the central nervous system.

4.1. Efferent feedback and diverse mechanisms for auditory tuning

Although the majority of auditory research focuses on ascending pathways, an important parallel pathway of descending efferent projections allows for central control and modification to sensory inputs at multiple levels of auditory processing. In mammals, the auditory efferent system consists of corticofugal projections from the cerebral cortex to lower brain centers and olivocochlear projections from auditory brainstem nuclei to the hair cell receptors of the inner ear. These descending pathways are proposed to improve signal detection within noise, permit selective attention to relevant stimuli, and protect the auditory periphery from damaging sounds (reviewed by Guinan 1996; Terreros and Delano 2015).

The efferent auditory system is most well-studied among mammals; however, comparative data from diverse species have revealed that efferent control of auditory processing, particularly via descending projections from the brainstem to the hair cells of the inner ear, is a key ancestral feature of the vertebrate ear. In this section, we focus on contributions of comparative work to our understanding of the evolution, variation, and function of the brainstem olivocochlear efferent system. However, studies using non-traditional species (e.g., bats, cats, chinchillas, and non-human primates), in conjunction with work using traditional laboratory models (mice and rats), have similarly elucidated the crucial role of the fore- and midbrain corticofugal pathway, including descending cortico-thalamic and cortico-collicular projections, in controlling downstream efferent centers. Through the combined modulatory effects of the corticofugal and olivocochlear pathways, the vertebrate auditory efferent system dynamically influences sensory processing at all levels, including frequency tuning (e.g., in bats: reviewed by Suga 2018, and cats: Villa et al., 1991), state-dependent changes such as those influenced by seasonality (e.g., in fish: Perelmuter et al., 2019, Forlano et al., 2015), and attentional gating such as that observed in the presence of non-auditory distractor stimuli (e.g., in cats: Oatman 1971 and chinchillas: Delano et al., 2007) and during auditory stimulus-specific adaptation (reviewed by Duque et al., 2015). Comparative work in efferent processing of sound information may yet reveal important information on whole brain top-down neural processing as well as provide an evolutionary perspective on modulatory pathways in the auditory system.

4.1.1. Organization of the auditory efferent system across species

All vertebrates, gnathostomes and agnathans alike, have inner ears with hair cells that receive efferent feedback from the brain via cholinergic fibers originating in the brainstem, suggesting that the efferent system co-evolved with the inner ear in craniates (Fritzsch, 1999). Comparative developmental study has shown that inner ear efferent fibers are evolutionarily derived from facial branchial motor neurons (reviewed by Chandrasekhar 2004; Gilland and Baker 2005), which project unilaterally in lamprey (Fritzsch et al., 1989) and bilaterally in gnathostome vertebrates including bony and cartilaginous fishes (Baker et al., 2008; Higashijima et al., 2000, Tomchik and Lu 2006), amphibians (Fritzsch 1981; Fritzsch and De Camprona 1984; Straka et al., 2005), birds (Whitehead and Morest 1981; Raabe and Köppl 2003), and mammals (Bruce et al., 1997) - Fig. 3. Vestibular and auditory efferents originate from a single efferent nucleus in the brainstem of fish and amphibians, the octavolateralis efferent nucleus, and branch extensively to innervate many hair cells within and across different inner ear end organs (Bell 1981; Bleckmann et al., 1991; Claas and Münz 1980; Edds-Walton et al., 1999; Fritzsch 1981; Gonzalez et al., 1993; Hellmann and Fritzsch 1996; Meredith and Roberts 1986). The efferent system of birds and mammals, in contrast, shows greater specialization in which vestibular and auditory efferents are topographically segregated into dorsal and ventral projections respectively in archosaurs, and are totally separate in mammals (Strutz 1982; Roberts and Meredith 1992). Separation of the auditory and vestibular efferent streams in birds and mammals is hypothesized to indicate a functional shift in the auditory system to support sensitive hearing on land; however, it remains unknown whether similar functional segregation exists in the efferent system of frogs and lizards (reviewed in Köppl 2011).

Figure 3: Generalized brainstem organization of the vertebrate octavolateral efferent system.

Auditory (filled circles) and vestibular (open circles) efferent neurons send projections (red) to peripheral sensory hair cells to modulate afferent inputs (blue). Abbrev: HC: hair cell (IHC: inner hair cell, OHC: outer hair cell, SHC: short hair cell, THC: tall hair cell), mlf: medial longitudinal fascicle, OC: olivocochlear (LOC: lateral olivocochlear, MOC: medial olivocochlear), V: trigeminal nucleus, VI abducens motor nucleus, VII: facial branchial motor nucleus.

The mammalian olivocochlear efferent system is further subdivided into lateral and medial pathways depending on their locations of origin in the brainstem and termination in the inner ear (Warr and Guinan 1979). Medial olivocochlear (MOC) efferents originate near the medial superior olive and send large, myelinated projections to ipsilateral and contralateral outer hair cells, whereas the lateral olivocochlear (LOC) neurons originate near the lateral superior olive and project small, unmyelinated fibers that form primarily ipsilateral axodendritic terminals on type one cochlear afferents (reviewed in Guinan 1996). Much of what we know about efferent control of cochlear responses derives from research on MOCs that, by virtue of their size and myelination, are easier to record and manipulate. Generally, activation of the MOC efferents will reduce the gain of the cochlear amplifier via hyperpolarization of the outer hair cells, and this system therefore supports reflexive modulation of the auditory response to both ipsilateral and contralateral sound stimulation. Mammalian species vary in proportion of crossed and uncrossed MOC fibers, and therefore in the lateralization of the MOC reflex (Warr, 1992); however, the functional significance of this variation is not well understood. Further, species differences in organization and distribution of MOC cells in the periolivary region (Aschoff and Ostwald 1987) indicates that much remains to be discovered with regards to the evolution of efferent control mechanisms for hearing across mammals.

The physiology and function of the LOC efferent system represents a key gap in our knowledge, and much remains unknown about the perceptual effects of its activation. In general, the effects of LOC activation likely occur on slow time scales due to their lack of myelination; however, a great diversity of neurotransmitters and neuromodulators have been immunolocalized to LOC terminals, indicating that LOC effects on hearing could be highly complex (Guinan 1996; Groff and Liberman 2003; Lauer et al., 2022). LOC efferents notably release dopamine onto auditory afferents in an inhibitory reflex circuit activated by noise exposure and hypothesized to preserve peripheral sensitivity by preventing excitotoxic damage (Gil-Loyzaga, 1995; Lendvai et al., 2011; Ruel et al., 2001). A similar dopaminergic efferent circuit observed in vocal midshipman fish (Perelmuter and Forlano 2017; Perelmuter et al., 2019) may functionally parallel the mammalian LOC, and represents an informative comparison for probing efferent modulation of peripheral inputs to the auditory pathway.

There remain many open questions regarding the functional significance of the auditory efferent system and its effects on hearing and sound localization in the presence of noise, attentional modulation of auditory sensitivity, and protection from noise-induced hearing loss. Efferent projections, their functional role, and the neurotransmitters that regulate them are largely understudied across taxa and represent an interesting area where further comparative work may provide insights into the function and consequences of modulation of auditory processing from the top-down perspective. As has been employed in previous studies of the contributions of the LSO and MSO to sound localization (described above in section 3.2.1), a “natural ablation” approach incorporating diverse species that vary in number of cells and/or their projection patterns could contribute key insights into the functional role and physiological parameters of the mammalian auditory efferent system. A greater understanding of natural cross-taxa variation in efferent control of hearing will undoubtedly yield insights into different processing strategies that may serve to enhance or protect hearing.

4.1.2. Ancestral and innovative mechanisms for cochlear amplification

Another major open question in sound capture concerns the evolution of active and passive amplification provided by specialized hair cells that greatly enhances the sensitivity and dynamic range of the ear, allowing it to detect and process a wide range of sounds. Evolutionary processes following the divergence of the major vertebrate lineages shaped the diversity of not only the auditory end organs, but also their innervation patterns and cellular specializations. For example, parallel evolution of the papillar end organs in mammals and archosaurs led to functionally distinct hair cell subpopulations that differ in their afferent and efferent supply. Mammals have two hair cell types: inner hair cells (IHCs) that are innervated by afferent fibers only (although these afferent fibers themselves receive efferent innervation) and outer hair cells (OHCs) that are innervated by both afferent and efferent terminals. Mammalian OHCs exhibit somatic electromotility (Brownell et al., 1985) driven by the motor capabilities of the membrane protein prestin (Zheng et al., 2000). Voltage-dependent conformational changes to prestin mediates OHC length changes that, through contact of OHC hair bundles with the tectorial membrane, act to amplify the mechanical response of the cochlea and enhance sensitivity to low-intensity sounds (Avan et al., 2019; Fettiplace 2020; Zhang et al., 2000).

The harnessing of prestin as a motor appears to have occurred more than once in amniote evolution (Beurg et al., 2013), but may have reached the epitome of its development as a motor molecule capable of operating at acoustic and ultrasonic frequencies in the OHCs of the mammalian cochlea (reviewed by Fettiplace 2020). The evolution of prestin-derived cochlear amplification facilitated the extension of the mammalian hearing range to high and even ultrasonic frequencies; however, it is not well-understood whether or how prestin is able to overcome the kinetic limitations of the OHC membrane resistance-capacitance time constant to confer cycle-by-cycle feedback at frequencies greater than 100 kHz (Frank et al., 1999; Li et al., 2022; Reuter et al., 1994; Santos-Sacchi and Tan 2018, 2019; Vavakou et al., 2019). Additionally, although molecular evolutionary studies highlight the diverse selective processes shaping the evolution of prestin in mammals (Franchini and Elgoyhen 2006; Liu et al., 2012), particularly among auditory specialists such as echolocating bats and toothed whales (Li et al., 2010), little is known about how molecular divergence and subsequent structural changes to prestin among mammals influences functional differences in cochlear physiology. Key prestin sequence differences observed among monotreme, marsupial, and placental mammals influence its voltage sensitivity and nonlinear capacitance behaviors, and therefore may result in variable OHC electromotility capabilities among diverse mammalian species (Liu et al., 2012).

An analogous amplification system in archosaurs relies on tall and short hair cells that function similarly to mammalian IHCs and OHCs, respectively. Although short hair cells lack the uniquely mammalian mechanism for prestin-based electromotility (He et al., 2003; Köppl et al., 2004), they are hypothesized to act as amplifiers via motility of their large stereocilliary bundles and are innervated by large-diameter efferent fibers that form calyx-like boutons (Fischer 1994; Keppler et al., 1994; Whitehead and Morest 1981; Köppl et al., 2009; Sul and Iwasa 2009). In fact, the force-displacement relationship of the hair bundles of both auditory and vestibular hair cells is nonlinear, indicating an active process capable of amplifying responses to low amplitude stimulation in a mechanism first demonstrated in the auditory hair cells of turtles (Crawford and Fettiplace 1985) and the saccular hair cells of frogs (Howard and Hudspeth 1987). Recent work has shown that the gecko papilla also shows fast electromechanical responses from hair bundles that augment the passive mechanical tuning of high-frequency hair cells through an as-yet unknown molecular mechanism (Beurg et al., 2022).

Active hair bundle motility has now been demonstrated in a diverse array of tetrapods, including frogs, lizards, turtles, birds, and mammals and likely represents an ancestral mechanism for response amplification in the inner ear (Crawford and Fettiplace 1981; 1985; Manley et al., 2001; Nin et al., 2012; Tinevez et al., 2007). This is supported by observations of spontaneous and evoked otoacoustic emissions traced to hair bundle motility in mammalian and non-mammalian species alike (reviewed in Köppl 1995). The effects of contralateral acoustic stimulation on otoacoustic emissions measured in barn owls suggests a potential role for efferent modulation of response amplification in archosaurs (Manley et al., 1999), although the physiological influence of the efferent system on the auditory response remains understudied in nonmammalian vertebrates. Additionally, the manner in which active mechanisms may combine to modulate auditory sensitivity - for example, in the mammalian cochlea, where hair cells show both active bundle mechanics and somatic electromotility - remains an important open question that comparative work is well-suited to address. Studies incorporating diverse species in which different molecular configurations of prestin may influence its functional parameters will be especially valuable.

4.2. Hormonal regulation of the auditory system

It is well known that hormones can influence sensory processing. A classic example is the hormonal role in chemical signaling that is critical for reproduction, kin recognition, and social behaviors. However, the importance of hormonal modulation of auditory circuitry has been underexplored, in part because the ascending auditory pathway and specifically the early processing of sound information, relies on extremely fast synaptic transmission compared to the longer, slower changes typically associated with hormones. Comparative studies have made important progress in defining what role hormones play in auditory processing and may hold the key towards future advancement in this area.

4.2.1. Steroid hormones

The steroid hormones are critical for reproductive function in many animals, but also play a role in seasonality of sensory responses, neuroprotection, and local synaptic impacts. The structure and function of steroid hormones is highly conserved across species and most research has focused on testosterone, progesterone, and 17β-estradiol due to their importance in production of gametes, maintenance of pregnancy in some animals, and sexual maturity (Guerriero, 2009). These hormones also have been shown to have important consequences for auditory information processing and, although there has been some focus on the auditory effects of estrogens in some species, overall the influence of steroid hormones represents an understudied area in auditory neuroscience.

Estrogens have important physiological consequences, including follicular development, wound healing, and bone development. Estrogens also function in the brain and auditory system throughout development in many species (Vahaba and Remage-Healey, 2018). For example, estrogens play key roles in protecting the auditory system from damage (e.g., noise-induced: Hu et al., 2016; Meltser et al., 2008; Shuster et al., 2021; Simonoska et al., 2009, ototoxic compounds: Hu et al., 2017; Nakamagoe et al., 2010), enhancing sensitivity during breeding season (Forlano et al., 2005; Maruska et al., 2012; Sisneros et al., 2004), providing local refinement of auditory processing and modulating auditory responses (Remage-Healey et al., 2011; Vahaba and Remage-Healey, 2018; Lynch and Wilczynski, 2006; Yovanof and Feng, 1983), and mediating acoustic learning (reviewed in Tremere et al., 2009). Estrogens may also partially underlie sex differences in auditory processing (Chakraborty and Burmeister, 2010; Zeyl et al., 2013 reviewed in Caras, 2013), although its diverse physiological effects throughout the body and differential activation and diversity of estrogen receptors leads to variability in estrogen impacts across species (Asnake et al., 2019; Fergus and Bass, 2013). Some studies have shown no sex-differences in estrogen receptor expression across the auditory system (Charitidi et al., 2010; Charitidi and Canlon, 2010; Motohashi et al., 2010); however, there may be variability in estrogen receptors across the reproductive cycle, indicating that estrus stage has an important influence in auditory processing (Charitidi et al., 2012).

Other studies focused on hormonal changes in response to acoustic information reveal additional steroid hormones important for auditory processing. For example, Gonadotropin releasing hormone (GnRH) has been observed to increase in immunoreactivity in green treefrogs after exposure to male advertisement calls (Burmeister and Wilczynski, 2005) and in damselfish during peak reproductive season (Maruska and Tricas, 2011), indicating that GnRH may play a seasonal and context-dependent role in hormonal regulation of auditory processing. Testosterone also has sex-specific effects in which neural thresholds increase in a frequency-specific manner for female green treefrogs, but not males (Miranda and Wilczynski, 2009). Many studies that show sex differences in acoustic sound perception may be due to several related factors including differences in efferent processing (see above) and hormonal mechanisms. The myriad influences of steroid hormones acting across the auditory system through specific receptor subtypes that vary across seasons, throughout development, and with advancing age, further the complexity of the auditory system and leave a diversity of open questions (Gall et al., 2013).

4.2.2. Neuropeptides

Acoustic communication is crucial for social relationships in many species. To establish and maintain social relationships, not only is the ability to vocalize important, but also the ability to process received acoustic information. Oxytocin (OT) and arginine vasopressin (AVP), the so-called “social hormones”, are two conserved neuropeptides that have important functions in the control of social behaviors, including communication and auditory plasticity (Borowiak and von Kriegstein, 2020; Marlin et al., 2015). These hormones serve distinct functions between the sexes: where AVP has been linked to aggression in males (Terranova et al., 2017), OT has a clear role in lactation in females (Matsunaga et al., 2020) and in pair bonding in both sexes (Walum and Young, 2018). OT and AVP may exert effects in some of the earliest sound processing regions, directly strengthening or changing circuits that help parse sound information into discrete pathways. Previous work has shown that OT and its receptor are expressed in the medial nucleus of the trapezoid body, an inhibitory brainstem nucleus important for the ability to localize sounds in mammals (Freeman et al., 2014; Kanwal and Rao, 2002). Interestingly, AVP expression is found in the female, but not male, guinea pig brainstem/midbrain, including in the ventral nucleus of the trapezoid body and inferior colliculus (Dubois-Dauphin et al., 1987), suggesting that AVP expression may be sex dependent and have different roles in auditory brainstem processing. Other work suggested that cochlear expression of OT and AVP could originate from the MOC system (Reuss et al., 2009). However, this study did not find that MOC neurons are the source of cochlear OT/AVP expression but rather express another similar neuropeptide PACAP (Reuss et al., 2009). Hormones that are important for social behaviors are likely important regulators of sound processing across species, but this area of research is vastly understudied and generally has focused on sound production for social communication rather than sound reception. Comparative research using closely related species that differ in their social behaviors, for example monogamous and polygynous Peromyscus deer mice or Microtus voles, could reveal how variation in hormonal influence on vocalization and auditory processing can subserve different acoustic behaviors.

4.2.3. Other hormonal influences on auditory processing

Hormonal regulation and modulation of the nervous system is complex, with hormones being released both within the brain and systemically through vasculature. There is potential for almost any hormone to have impacts on sensory processing, depending on receptor presence in tissues and receptor subtypes. For example, thyroid hormone receptor β1 and β2 are both expressed in the mouse cochlea, but only β1 has been implicated to play a role in hearing loss (Abel et al., 1999). Sex differences in auditory processing in humans are well known, however many outstanding questions remain, including when during development these differences arise, if they are hormonally dependent, if there are critical periods during development that underlie these differences, what mechanisms regulate differences, whether hormones provide any neural protective effects, and how aging and menopause may impact differences across the lifespan. These questions may be well suited for comparative studies with animals that are auditory specialists or have unique characteristics that lead to diversity in hearing ability. Hormonal correlates between human language acquisition and vocal learning in songbirds has led to a wealth of sensory learning studies in diverse archosaurian species that have enriched our knowledge of analogous neural control mechanisms shaping language acquisition. Recent work in mammals, including highly social bats (Lattenkamp et al., 2021; Prat et al., 2015; Vernes et al., 2022) and amphibious pinnipeds (Stansbury and Janik 2021; Verga et al., 2022), highlights additional taxa for which comparative assessment of shared and disparate neurochemical mechanisms underlying auditory processing and vocal learning may serve to boost our understanding of homologous mechanisms in humans.

Finally, mounting evidence suggests that environmental disruption through endocrine contamination may have detrimental impacts on hearing abilities in humans and other species by disturbing pathways discussed above. For example, environmental exposure to ototoxic and neurotoxic compounds, such as polychlorinated biphenyl and methyl-mercury, during development is correlated with a cascade of auditory deficits at both the peripheral and central level in songbirds (Wolf et al., 2017), rodents (Crofton et al., 2000; Kenet et al., 2007; Lee et al., 2021; Powers et al., 2009) and humans (Trnovec et al., 2008; Min et al., 2014; Poon et al., 2015). Unfortunately, a broad range of environmental contaminants are known to cause hearing loss in humans and laboratory animals (reviewed in Fábelová, 2019); without further comparative studies, our picture of environmental endocrine impacts on auditory processing will remain incomplete. The diversity of hormonal regulation already apparent - particularly in the context of additional variation in efferent signaling and fundamental variation in ascending inputs - defines a challenging variable space that careful comparisons across and within taxa and species will be essential to address.

5. Conclusion

Comparative research has provided many foundational insights on hearing. In this review, we have summarized contributions of vertebrate comparative work across only a small selection of topics including mechanisms of sound capture, directional hearing, and top-down processing, with an emphasis on opportunities for future work and technological advances. Comparative work will continue to be an essential driver of progress in the auditory field, particularly as new techniques, including new genetic tools in non-mouse models, non-invasive recording, and high-resolution imaging methods, continue to generate new opportunities for the study of fundamental auditory questions.

Highlights.

• Comparative research has driven many important discoveries in the auditory field

• There are many questions that a comparative approach is uniquely suited to address

• New tools allow for a diversity of measurements across species

Abbreviations:

AVP

arginine vasopressin

ILD

interaural level difference

ITD

interaural time difference

LOC

lateral olivocochlear

LSO

lateral superior olive

MOC

medial olivocochlear

MSO

medial superior olive

OT

oxytocin