Otoacoustic emissions in humans, birds, lizards, and frogs: evidence for multiple generation mechanisms

Abstract

Many non-mammalian ears lack physiological features considered integral to the generation of otoacoustic emissions in mammals, including basilar-membrane traveling waves and hair-cell somatic motility. To help elucidate the mechanisms of emission generation, this study systematically measured and compared evoked emissions in all four classes of tetrapod vertebrates using identical stimulus paradigms. Overall emission levels are largest in the lizard and frog species studied and smallest in the chicken. Emission levels in humans, the only examined species with somatic hair cell motility, were intermediate. Both geckos and frogs exhibit substantially higher levels of high-order intermodulation distortion. Stimulus frequency emission phase-gradient delays are longest in humans but are at least 1 ms in all species. Comparisons between stimulus-frequency emission and distortion-product emission phase gradients for low stimulus levels indicate that representatives from all classes except frog show evidence for two distinct generation mechanisms analogous to the reflection- and distortion-source (i.e., place- and wave-fixed) mechanisms evident in mammals. Despite morphological differences, the results suggest the role of a scaling-symmetric traveling wave in chicken emission generation, similar to that in mammals, and perhaps some analog in the gecko.

Appendix: Anatomical overview

This section provides background on the anatomy and physiology of peripheral auditory structures potentially relevant to OAE generation.

All the species examined here have a tympanic membrane enclosing the middle ear, although the presence of a tympanic membrane is unnecessary for the detection of DPOAEs (van Dijk et al. 2002). In contrast to mammals, the middle ear in the non-mammalian species examined here consists of a single bone, the columella, that couples the tympanic membrane directly to the stapedial footplate. The middle ears of both mammals and non-mammals play similar functional roles, providing both forward (stimulus going in) and reverse (OAE coming out) transmission to and from the inner ear.

The greatest amount of diversity across the examined species lies in the inner-ear anatomy (Fig. 9). Analogous to the organ of Corti in the mammalian cochlea, bird and lizard hair cells are situated in a structure called the basilar papilla (BP). The frog has two morphologically distinct auditory papilla in the inner ear. Specifics of the inner ear structures of each species are summarized below. The main qualitative differences in inner ear anatomy and physiology are summarized in Table 2. For brevity, we provide only brief descriptions of human cochlear anatomy and hearing perception (see Dallos et al. 1996). The term “cochlea” is reserved here for the auditory organ of the mammalian inner ear.

Fig. 9

Comparative schematic of inner-ear anatomy. Two perspectives are provided for each group: a cross-sectional view (left) and a top-down view of a section of the sensory epithelium (right). Except for the frog, the arrows in the top-down view represent an individual hair cell, the direction indicating the bundle’s polarization (pointing from shortest to tallest row). For the frog, the entire longitudinal length of only the amphibian papilla (AP) is shown and arrows indicate gross trends of the hair cells (the finely dashed bounding box corresponds to where the cross-section would lay and the coarsely dashed line represents where the sensing membrane extends down from the roof of the recess). Cells known to exhibit cell body somatic motility are indicated by a star on their tail. White regions are fluid-filled, grey regions correspond to overlying tectorium (with grey lines indicating fibrillar structure), grey striped area represents bone and stippled areas are non-stereociliary cellular regions (e.g., supporting cells). Distinction between scala vestibuli and scala media is omitted. Legend is as follows: AP amphibian papilla, AR amphibian recess, BM basilar membrane, BP basilar papilla, FN fundus, LL limbic lip, SA sallet, SC sallet chain, SE sensing membrane, SM scala media, ST scala tympani, TC tunnel of Corti, TM tectorial membrane

Humans

The human cochlea (i.e., BM length) is typically 33–35 mm in length (coiled over two and a half turns) and contains around 15,000 total hair cells (Ulehlova et al. 1987). Typical thresholds in a healthy human ear are relatively flat between 0.5 and 7 kHz, being in the range of −5 to 15 dB SPL. Peak sensitivity occurs in the frequency range of 3–4 kHz. Psychophysical human Q _ERB values typically range from 7 to 10 around 1 kHz to 9–17 around 4 kHz (e.g., Glasberg and Moore 2000; Shera et al. 2002) ¹³.

Chickens

Chickens have a short BM (∼3–4 mm) that curves gently over its length. The BM width and thickness change along its length (as well as hair-cell bundle properties such as height and number of stereocilia), correlating to the tonotopic gradient observed from ANF responses (Manley et al. 1987; Chen et al. 1994). Evidence from avian species suggests that a longitudinally traveling wave is present along the BM (von Bekesy 1960; Gummer et al. 1987). There are ≈5,000 hair cells situated in a hexagonal fashion (Tilney and Saunders 1983). In general, two distinct types of hair cells have been characterized: short hair cells sitting directly atop the BM (receiving the bulk of the efferent innervation) and tall hair cells (with the bulk of the afferent innervation) (Tanaka and Smith 1978). Chick hair cells do not exhibit somatic motility (He et al 2003; Köppl et al. 2004). The overlying TM is relatively quite thick, with dense radial and longitudinal fibers apparent under a light microscope. Cavities that are present in the TM over each hair cell extend back towards the homogene cells (at the neural edge), making the TM appear porous through a given cross-section (Cotanche 1987). For all hair cells in the papilla, the tallest row of the stereociliary bundle is tightly coupled to the TM, which is also attached to the papillar surface via fibrillar connections coupling to the microvilli of the supporting cells (Tanaka and Smith 1975).

Based upon ANF recordings from previous studies, P21 (number of days post-hatching) chicks have a flat mean threshold of ≈20 dB SPL from 0.2 to 3 kHz, increasing sharply at higher frequencies (Manley 1990). Psychoacoustic studies in adult chickens correlate well to these measurements (Saunders and Salvi 1993). Q ₁₀ values from the single units are typically around 2–5 (though some units exhibit significantly higher Q values), and increase with characteristic frequency (Salvi et al. 1992). While DPOAEs have been measured in the chicken (Kettembeil et al. 1995; Ipakchi et al. 2005; Lichtenhan et al. 2005), the authors know of no published reports of spontaneous emissions (SOAEs) or SFOAEs in Gallus gallus domesticus.

Geckos

Two species of gecko were examined in this study: Leopard geckos (E. macularius) and Tokay geckos (G. gecko). Both have similar peripheral auditory anatomy (Wever 1978), including a short (1.2–1.8 mm) and straight BM. Both the width and thickness of the BM and BP vary considerably over the longitudinal length. The BP contains ≈1,000–2,000 hair cells (Wever 1978). The BM in G. gecko is slightly longer and supports ≈40% more hair cells. Hair-cell bundles are oriented both uni-directionally (all in the same direction) and bi-directionally (180° relative to one another). There is a unique tectorial topology along one region of the papilla that consists of sallets, discretized sections of TM loosely coupled to each other via a fine strand overlying their top surface called the sallet chain. The sallets couple a single row of bi-directionally oriented hair cells together as shown in Fig. 9 (Wever 1978). Evidence suggests the absence of both somatic hair-cell motility (Köppl et al. 2004) and traveling waves (Peake and Ling 1980; Manley et al. 1988, 1999) along the gecko BM. A thickened tissue called the fundus runs along the length of the BM underneath the BP. Both afferent and efferent innervations are present, though the latter appears exclusive to the uni-directional segment of the papilla (Manley 1990).

Previous studies have looked at microphonic responses in both species (Wever 1978) and ANF responses in G. gecko (Eatock et al. 1981; Sams-Dodd and Capranica 1994; Manley et al. 1999), giving an indication of the thresholds and sharpness of tuning ¹⁴. Based upon microphonic and ANF data, the G. gecko ear has a threshold of ≈10–15 dB SPL in its most sensitive region of 0.5–0.8 kHz. E. macularius appears to be a further 10–15 dB more sensitive than G. gecko based upon microphonic comparisons, suggesting thresholds at or below 0 dB SPL. Derived Q ₁₀ values from the ANF studies for G. gecko were ∼2–4, increasing with characteristic frequency. Spontaneous emissions have been reported in both E. macularius and G. gecko (Manley et al. 1996; Stewart and Hudspeth 2000), but the present authors are unaware of any previous reports of evoked emissions in either gecko species.

Frogs

Frogs have two papillae that are sensitive to sound, the amphibian papilla (AP) and the basilar papilla (BP) (Wever 1985). In contrast to chicks and geckos, both papillae in frogs lack a flexible BM altogether, and the hair cells sit atop relatively rigid tissue (Wever 1973). Unlike those of the human and chicken, the hair cells in the papillae do not exhibit any obvious morphological distinctions (such as short vs. tall hair cells), but do exhibit a degree of bi-directionality (similar to that seen in the gecko). Shaped roughly like a horseshoe and ∼0.5–0.6 mm long (Wever 1973), the AP is tonotopically organized (Lewis et al. 1982) and is sensitive to frequencies below ∼1.2 kHz. Containing ≈800 hair cells, the AP has a thick TM punctuated by many small holes (Wever 1973). The hair cells couple tightly to the TM. A tectorial curtain (or sensing membrane extends from the bony roof of the AP recess down to the central portion of the TM. There does not appear to be a smooth gradation in either bundle or TM properties along the length of the AP (Shofner and Feng 1983; Lewis and Leverenz 1983). The BP, sensitive to higher frequencies (above ≈1.3 kHz) is smaller, containing only about 70 hair cells. The BP is thought to act as a singly tuned resonator (Ronken 1991). Unlike the AP, the BP does not appear to receive any efferent innervation in ranid frogs (Ronken 1991), though efferent innervation to the BP has been found in other amphibian species (Hellmann and Fritzsch 1996). Similar to lizards, a great deal of diversity is seen in the inner ear anatomy across different species of frogs.

Microphonic measurements in other frog species of the same family examined here indicate airborne thresholds near ∼20–40 dB SPL, being smallest in the range 0.2–0.6 kHz (Wever 1985). ANF responses in L. pipiens revealed higher mean thresholds, typically 50 dB SPL around 0.5–1 kHz and increasing at both lower and higher frequencies (Ronken 1991). Q ₁₀ values range between 1 and 2, increasing at frequencies below 0.5 kHz and above 2 kHz (Ronken 1991). The existence of SOAEs has been reported for L. pipiens (van Dijk et al. 1996) while both DPOAEs (Meenderink et al. 2005) and SFOAEs (Meenderink and Narins 2006) have also been reported.