DPOAEs and tympanal membrane vibrations reveal adaptations of the sexually dimorphic ear of the concave-eared torrent frog, Odorrana tormota

Changes in tympanum sensitivity in Odorrana females

In females of O. tormota, we found that the auditory sensitivity of the middle ear changes with the ET closure as in males. The morphology that allows closing the ET is also similar to that described in males (Gridi-Papp et al. 2008). The distal end of the hyoid horn forms a hinge at its attachment to the skull that allows it to pivot. This pivoting mechanism allows the cartilaginous horn of the hyoid to arch over the lumen of the ET and close it. The capability of females to close the ET was discovered using anatomical observations and manipulations, much like was done for males of this species (Gridi-Papp et al. 2008). In subsequent behavioral observations of captive female frogs, we did not notice closure of the ET during vocalization, and only once only during swallowing. Further investigation is needed to determine whether female Odorrana frogs use the same neuromuscular mechanism as males to close the ET.

The tympanum velocity in response to acoustic stimulation with pure tones was evaluated in four females with the ET open and closed. With the ET open, the tympanum responded with vibration velocities above 1 mm/s/Pa over a broad range of frequencies (2–10 kHz); maximum velocities occurred at 3 kHz (Fig. 2b, red curve). With the ET closed, the tympanum response exhibited a shift in sensitivity towards a higher spectral range (10–21 kHz) and peak velocities occurred at 14 kHz (Fig. 2b, brown curve). This shift in middle ear tuning was described previously in males: a high sensitivity at frequencies < 10 kHz with the ET open changed to 10–32 kHz after the closure of the ET (Gridi-Papp et al. 2008). In both sexes, the ET closure tends to attenuate low frequencies and amplify higher frequencies, but the effects this mechanism could have in sound perception could differ widely considering the sexual differences in the anatomy and physiology of the auditory system of this species (Shen et al. 2008).

Middle ear sensitivity to natural stimuli

The tympanum’s sensitivity to conspecific calls was investigated in four males with the ET open and closed (Fig. 3). To assess the frequency-dependent relative differences in sensitivity between these conditions, we selected three spectral bands (B1–B3). For the multi-harmonic calls (natural and equalized), the analysis focused on the spectral bands centered around the dominant frequency (5.5–8 kHz) and around the two first harmonics (11–16 kHz, 17–23.5 kHz). For calls with NLP, we defined analysis bands around the three main spectral peaks which appeared at 6–9, 12–15 and 16–18 kHz in the female call, and at 3.5–6, 7–10 and 11.2–16 kHz in the male call.

Fig. 3figure 3

Tympanum response of males of Odorrana tormota to species-specific calls. Spectrograms of the vocal signals used as stimuli and velocity profiles of the tympanic membrane in response to them: a multi-note short-call (male), b multi-note short-call with equalized harmonics (male), c long-duration call (female), d short-duration call (male). Dark green: open-ET condition, orange: closed-ET condition. Normalized power spectra of the calls as recorded by the reference microphone are shown as gray areas. Scatter plots represent mean coherence between the stimulus (natural or synthetic call) and the tympanum vibration for the selected spectral bands (B1–B3, gray bands on top of spectra). Bold horizontal lines denote the median, and the bottom and top edges of the box indicate minimum and maximum, respectively

In general, the frequency of the maximal vibrational response corresponded to the dominant frequency of the calls indicating that males’ tympana are mechanically well-tuned to the conspecific calls. The mean coherence between the vibration velocity of the tympanum and the sound pressure recorded by the reference microphone adjacent to the tympanum was also calculated at the specified spectral bands (Fig. 3). The highest coherence corresponded to the low-frequency band (B1) both in the ET-open state and in the ET-closed state. However, B1 coherence values corresponding to the recordings with the ET open (~ 0.8) were always higher relative to the recordings with the ET closed, where greater variability was also found (0.4–0.6). In contrast, median coherence values for the other spectral bands (B2 and B3) tend to be higher in the recordings with the ET closed. It is remarkable, however, that the changes in sensitivity caused by the closure of the ET, loss of sensitivity at low frequencies and gain of sensitivity at high frequencies, are not equivalent, and the relationship between them varies with the type of calls.

For example, when we analyzed the tympanum response to a multi-harmonic call for which the harmonic amplitudes roll off as the frequency increases, we found that the decline in the response to the fundamental frequency is not counteracted by an increase in the detection of the third harmonic, for which there was almost no vibration with either the open or closed ET (Fig. 3a). The use of equal-amplitude tones in experimental approaches helps to assess the spectral sensitivity of the auditory structures, but still falls short of illustrating their performance in a natural context. As with tones of equal amplitude (Gridi-Papp et al. 2008), the tympanum response to equalized harmonics suggests that the ET closure facilitates the detection of higher harmonics (Fig. 3b). However, the energy content of these high-frequency harmonics in the natural calls could be so low that they do not cause the eardrum to vibrate significantly (Fig. 3a). The capacity of detecting the higher harmonics potentially benefits from the ET closure but it would depend on the energetic roll-off characteristic of the call.

Our assessment of the tympanum’s vibration also used representative calls with NLP from both sexes. In the ET-closed condition, the response of the tympanum to the female call increased for frequencies greater than 10 kHz relative to the response corresponding to ET-open (Fig. 3c). In contrast, the tympanum response to the male call, which concentrates most of the energy below 10 kHz, mostly diminished and did not benefit from the ET closure (Fig. 3d). Generally, dominant frequencies of female calls are in a higher spectral range relative to male calls (~ 7.5–9.2 kHz, Shen et al. 2008; Zhang et al. 2017); consequently, harmonics and NLP appear at very high frequencies (> 15 kHz). These findings suggest that the closure of the ET in males of O. tormota could be relevant in a behavioral context beyond helping to extract the desired signals from background noise. It might play a role in intraspecific acoustic communication by helping to better detect the high-frequency components of female calls while decreasing sensitivity to low-frequency calls mainly produced by other males.

DPOAEs in O. tormota

Distortion product otoacoustic emissions were recorded in all 20 ears investigated (10 females, 10 males). A median DP-gram for each sex was obtained by determining the median DP2f1–f2 amplitude for each f1 frequency across same-sex subjects (Fig. 4). In females, the DP2f1–f2 was distinguishable from system noise (≥ noise + 2SD) when f1 was between 2.27 and 6.73 kHz, except for one individual in which the DPOAE response extended to 8.36 kHz. In males, DP2f1–f2 was observed in a more extended frequency range, from 3.09 to 14.73 kHz. In two males, DPOAE response was detected as high as 18.19 kHz, the highest f1 frequency used.

Fig. 4figure 4

DPOAE response in Odorrana tormota. Individual DP-grams recorded in 10 females (a) and 10 males (b) reveal the sexual differences in the inner ear sensitivity. Data recorded with the ET open. Bold lines represent median DP-grams. System noise is indicated in gray. L1 = L2 = 80 dB SPL; f2/f1 = 1.1

In most of our DPOAE analyses, we refer to the primary frequency f1, which is commonly used in the literature, thus allowing direct comparison with other species. Nevertheless, also revealing the frequencies at which the cubic distortions 2f1–f2 and 2f2–f1 were detected helps to grasp the remarkable extent of the hearing range in O. tormota. The DPOAE response in females appeared from 2 to 7.4 kHz, while in males this spectral range extended between 2.5 and 20 kHz. The inner ear of this species exhibits a reciprocal matched filtering (Cobo-Cuan and Narins 2020) that could facilitate the detection of the opposite sex's vocalizations. Sexual differences found in inner ear tuning are in agreement with the data from behavioral, vibrometric and electrophysiological evaluations of O. tormota (Gridi-Papp et al. 2008; Shen et al. 2011).

The bimodal DP-grams typically recorded in frogs (Van Dijk and Manley 2001; Van Dijk et al. 2003; Meenderink and Van Dijk 2004) are also confirmed in the DP-grams of O. tormota females. They exhibit two frequency regions with elevated DPOAE amplitudes which are separated by a notch in DPOAE amplitude. The two relative amplitude maxima reflect two spectral filters as derived from the amphibian papilla (AP) and the basilar papilla (BP). The best sensitivity in the lower frequency range, presumably corresponding to the AP, was found at 3.36 kHz (2.91–3.81 kHz). The second sensitivity peak, presumably corresponding to the BP frequency range, showed the best sensitivity at 5 kHz (4.73–6.19 kHz).

In males, however, it is noteworthy that most DP-grams exhibited three sensitivity peaks instead of two. The best sensitivity was found at 5.05 kHz (3.64–7.55 kHz) and at 7.64 kHz (6.36–10.45 kHz) in the low- and mid-frequency regions, respectively. The sensitivity peak in the high-frequency region appeared at 11.27 kHz (10.55–16.09 kHz). The presence of three relative amplitude maxima at low-, mid- and high-frequency regions was also described in DPOAE recordings from males of Huia cavitympanum (Cobo-Cuan et al. 2020). It is tempting to assume that the three spectral ranges of sensitivity may be related to three separate functional epithelial patches in the inner ear. To this end, it is important to consider that the AP itself appears to be a combination of two distinctly functioning auditory epithelia that fuse during development to form a single organ (Li and Lewis 1974). These AP regions have oppositely polarized hair-cell populations that differ in their electrical properties and ionic currents (Smotherman and Narins 1999). Interestingly, the AP caudal patch exhibits an elongation whose shape, relative length, and proportion of hair cells apparently correspond to the extension of auditory sensitivity to higher frequencies (Lewis et al. 1992). It is possible that in O. tormota, like in H. cavitympanum, DPOAEs corresponding to the low- and mid-frequency ranges are generated in the rostral and caudal regions of the AP, respectively. Compared to most anurans, these two species are exceptional in their ability to communicate over an extensive spectral range that reaches high frequencies. It could be possible that frogs’ species with auditory sensitivity over a wider spectrum have less overlapping between the frequency ranges detected by the two regions of the amphibian papilla. Further studies are necessary to test this hypothesis. Both the AP and the BP were tuned to significantly higher frequencies in males than in females (AP: U = 3, p < 0.001; BP: U = 0, p < 0.001).

Examples of matrices of the DP2f1–f2 amplitude in response to a set of primary-tone frequency-level combinations are shown as color maps in Fig. 5a, b. From these response matrices, DPOAE threshold tuning curves were obtained by interpolating the stimulus levels necessary to elicit a DP2f1–f2 amplitude ≥ noise + 2SD across the frequency range tested. A Wilcoxon signed-rank test indicated that the minimum thresholds corresponding to each auditory papillae were not statistically different in females. The median minimum threshold of both the AP and the BP was 62 dB SPL. To examine differences in sensitivity between the auditory end-organs of males, only the minimum thresholds in the low- and high-frequency regions were compared. A Wilcoxon signed-rank test indicated that the AP minimum thresholds in males were significantly lower than the BP minimum thresholds (Z = 2.203, p = 0.027). The median of the minimum threshold was 63 dB SPL for the AP (60–64 dB SPL) and 66 dB SPL for the BP (62–72 dB SPL). These differences in sensitivity between papillae have been found in other anurans such as Rana pipiens and H. cavitympanum) (Meenderink and Van Dijk 2004; Cobo-Cuan et al. 2020). The involvement of active processes in the DPOAE generation could explain the higher sensitivity of the AP (Meenderink and Van Dijk 2006).

Fig. 5figure 5

Frequency tuning and sensitivity of the inner ear. DPOAE responses recorded in a female (a) and a male (b) of Odorrana tormota. Normalized colormaps show the DP2f1-f2 amplitude for each frequency-level combination. The white arrows delimit the frequencies selected to evaluate DPOAE growth. Regions with diagonal-lines pattern indicate stimuli combinations that generated system distortions c, d DPOAE growth functions extracted from each sensitivity peak found in the corresponding DPOAE response matrices shown above. Solid lines correspond to DP2f1–f2 amplitude ≥ noise + 2SD. Background noise is indicated in gray. In both sexes, the growth functions were analyzed for low- and high-frequency regions, but mid-frequency regions were additionally evaluated in males. e, f Histograms including all the slopes of DPOAE growth registered at the different frequency ranges of high sensitivity in all individuals. Slopes’ distributions corresponding to low- and mid-frequency regions appear to the left, and slopes’ distributions corresponding to high-frequency regions appear to the right. DPOAE amplitude grows steeply (> 2 dB/dB) with high stimulus levels (> 80 dB SPL, lighter colors) while it decreases or grows with shallow slopes (≤ 1 dB/dB) with low stimulus levels (< 75 dB SPL, darker colors)

DPOAE growth functions (DPOAE amplitude as a function of the stimulus level) were extracted from the individual DPOAE response matrices at ± 0.5 kHz around the frequencies of highest sensitivity (Fig. 5c, d). The slope of the growth functions at each data point was calculated by fitting a straight line through its two neighbors (i.e., the data points evoked with stimulus levels 2 dB above and below the data point in question). Frequency distributions of the slopes corresponding to each sensitivity peak are depicted as histograms in Fig. 5e, f. In females, the growth functions were evaluated for low- and high-frequency regions, and a two-sample Kolmogorov–Smirnov test indicated that their slopes’ distributions were significantly different (p < 0.01). Both histograms show a prevalence of 1 dB/dB slopes for low stimulus levels (< 70 dB SPL) and much steeper slopes (> 2 dB/dB) for high-stimulus levels (> 80 dB SPL). However, a slow DPOAE growth (slopes < 1 dB/dB) was found exclusively in the histograms corresponding to the low-frequency region, which is consistent with the compressive nonlinearity typical of DPOAE growth functions from the frog’s AP (Meenderink and Van Dijk 2004; Vassilakis et al. 2004). In contrast, linear growth at a rate > 1 dB/dB prevails in the high-frequency region consistent with the DPOAE growth functions from the BP (Meenderink and Van Dijk 2005, 2006).

In males, slope analysis included the three sensitivity peaks that appeared in the low-, mid- and high-frequency regions of the colormaps. In general, results were similar to those found in females. Slope distributions from the low and mid-frequency regions, presumably corresponding to the AP, are skewed toward shallow slopes (≤ 1 dB/dB), while the slope distribution corresponding to the high-frequency region, presumably corresponding to the BP, appeared skewed toward steeper slopes (~ 3 dB/dB) which is consistent with a passive DPOAE generation (Meenderink and Van Dijk 2006).

Effects of ET closure on inner ear tuning

For a small subset of frogs (2 females, 4 males), we investigated the effect of ET closure on inner ear sensitivity. DP-grams were recorded (as described above) with the ET open and closed. No DPOAEs were detected in females with the ET closed (Fig. 6a). By reopening the ET, the response recovered completely indicating that the ET closure procedure did not affect hearing permanently. The disappearance of the DPOAE response is understandable if we analyze the electrophysiological, vibrometric and DPOAE data available for females of this species. Shen et al. (2011) evaluated single-unit responses from the torus semicircularis of Odorrana females and only found units with best frequencies < 10 kHz. DP-grams recorded with the ET open showed that the spectral sensitivity of the inner ear of these female frogs is limited to frequencies below 10 kHz. By closing the ET, the middle ear tuning changes to a frequency range (> 10 kHz) where the inner ear and even the central auditory system have limited or no sensitivity (this study, Shen et al. 2011). Taken together these results suggest that ET closure temporarily impairs hearing in Odorrana females.

Fig. 6figure 6

Changes in inner ear sensitivity with the ET closure. DP-grams registered in a female (red) and a male (green) with the ET open and closed. In females, the DPOAE response disappeared after closing the ET. In males, the ET closure caused a sensitivity shift towards the high frequencies; DPOAEs attenuated at low frequencies and increased at high frequencies relative to the amplitudes recorded with the ET open. Shaded rectangles (top) represent the ranges of the fundamental frequency of Odorrana tormota vocalizations (male: gray, female: pink): I meow and staccato calls, II shallow-fm calls, III single and multi-note calls, IV female calls (data from Feng et al. 2009; Zhang et al. 2017)

In contrast, DPOAEs in males do not disappear in the noise with ET closure but undergo a shift in inner ear sensitivity (Fig. 6b). DP-grams recorded with the ET closed showed amplitude changes in the relative maxima: DPOAEs attenuated (median − 14 dB) at low frequencies and increased (median + 8 dB) at high frequencies relative to the amplitudes recorded with the ET open. This shift in the DP-grams towards the high frequencies suggests that the ET closure not only changes the middle ear tuning and inward sound transmission through the ossicles, but also affects outward transmission of the otoacoustic emissions generated by the inner ear amplification. Considering that maximum DPOAE amplitudes generally correlate to the frequencies of high auditory sensitivity (Kössl 1992; Van Dijk and Manley 2001; Bergevin et al. 2008), our results indicate that the ET closure also impacts hearing in Odorrana males but enhances sound perception at the higher frequencies.

The present study revealed the inner ear sensitivity of both sexes of O. tormota, an anuran species in which both sexes vocalize. Sexual differences in tuning, previously described at other levels of the auditory system (Shen et al. 2011), are now shown in the inner ear indicating that each sex is more sensitive to the frequencies of the other sex's vocalizations. Moreover, we confirmed that the capability of closing the ET is not only present in males but also in females of O. tormota. The ET closure seems to play a functional role in hearing; it impacts the sensitivity of middle and inner ear at frequencies used for social and sexual communication with conspecifics. In males, an increased sensitivity at higher frequencies would presumably confer advantages in locating calling females in the acoustically complex environment where calls of conspecific males compete with high levels of low-frequency noise (i.e., the sound of flowing water). We speculate that in females, the ET closure might play a role in hearing protection since it causes the disappearance of the inner ear response, but further investigation to verify this is required. This study broadens our understanding of anuran hearing and the diversity of adaptations to acoustically communicate in noisy environments.

留言 (0)

沒有登入
gif