Post-Metamorphic Development of the Frequency Selectivities and Sensitivities of the Peripheral Auditory System of the Bullfrog, Rana Catesbeiana

Anurans are capable of discriminating conspecific mating calls from those of other species (Martof & Thompson, 1958; Capranica, 1965; Littlejohn & Loftus-Hills, 1968). One of the underlying mechanisms of this discrimination is that the peripheral auditory system acts as a filter that is selectively responsive to the dominant spectral energies present in the call (Frishkopf, Capranica & Goldstein, 1968).

Comparative studies have shown that an inverse correlation exists between the body size of the anuran and the dominant spectral energies present in the calls, i.e. smaller species of anurans produce higher frequency sounds (Blair, 1963). A similar relationship has also been observed within a given species (Capranica, 1965 ; Oldham & Gerhardt, 1975; Ryan, 1980). Comparisons of the peripheral auditory selectivities from several anuran species (Sachs, 1964; Frishkopf et al. 1968; Capranica & Moffat, 1974, 1975; Moffat & Capranica, 1974; Feng, Narins & Capranica, 1975; Capranica, 1976 ; Narins & Capranica, 1976; Mudry, Constantine-Paton & Capranica, 1977) as well as central auditory responses (Loftus-Hills & Johnstone, 1970; Capranica, Frishkopf & Nevo, 1973 ; Loftus-Hills, 1973) also showan inverse relationship between body size and high-frequency selectivity (response to frequencies at a constant intensity) and sensitivity (response to intensities at a constant frequency) of the auditory system. Furthermore, mechanical measurements of the vibration of the middle ear in several anuran species reveal that the upper cut-off frequency is lower in large species than in small species (Saunders & Johnstone, 1972 ; Moffat & Capranica, 1978). Thus, the size of the anuran presumably influences the selectivity of the auditory periphery due to the differences in the sizes and masses of the peripheral auditory structures found among the various species.

A relationship between auditory selectivity and body size has not been demonstrated ontogenetically within a given species, in spite of the fact that there can be a dramatic increase in body size as well as in the size of the peripheral auditory structures during post-metamorphic development. In maturing bullfrogs, for example, there is a 10-fold increase in the diameter of the tympanum (Fig. 1). The sizes of the middle ear cavity, mouth cavity and columellar bones also increase with age and body size, and thus the acoustic transmission characteristics of these structures are presumably altered. These observations suggest that the frequency selectivity and sensitivity of the bullfrog auditory periphery may undergo some changes during the development following metamorphosis.

Previous single unit studies of the frequency selectivities of adult VIIIth nerve fibres have revealed that three populations of auditory fibres generally exist (see Capranica, 1976 for review): low-fiequency selective fibres which show two-tone inhibition, and mid-and high-frequency selective fibres which are non-inhibitable. With this in mind, the purposes of this study were: (i) to examine the acoustic response properties of single primary auditory fibres from early post-metamorphic bullfrogs and (ii) to compare these response properties with those of adult bullfrogs in order to gain an understanding of the functional development of the anuran auditory periphery. Evidence is presented that a number of the response properties change during post-metamorphic development.

Two groups of bullfrogs (Rana catesbeiana) were obtained from Charles Sullivan (Nashville, Tenn.): adults and early post-metamorphic frogs with snout-vent lengths ranging from 152 to 178 mm and 27 to 46 mm, respectively. During surgery the animals were anaesthetized by surrounding them in crushed ice (Kaplan, 1969), and the Vlllth nerve was exposed by a dorsal approach (for details, see Feng, 1980). Briefly, the skull overlying the nerve was removed, and the choroid plexus was carefully laid medially to expose the nerve. The dural membranes surrounding the nerve were removed with a sharpened tungsten needle. The dorsal approach has the advantage of recording Vlllth nerve activity with the mouth cavity closed, thus preserving its acoustic property.

The animals were allowed to recover from hypothermia for 2–3 h and were later immobilized with an intramuscular injection of d-tubocurarine chloride (3 mg/ml) during the recording session. Adults were injected initially with 4 rnl/kg body weight whereas froglets received 2 ml/kg. Periodic injections were administered to the animal during the recording session to maintain immobilization. Wet gauze was placed over the animal to facilitate cutaneous respiration and prevent evaporative water loss. Blood flow through the vessels of the choroid plexus served as a useful monitor of the physiological condition of the animal.

Animals were placed in a sound-proof room (Tracoustics) which was maintained at 20–22 °C. Single unit responses were recorded using 3 M-NaCl-filled glass micropipettes (10–20 MΩ). The electrodes were advanced by a hydraulic microdrive (Kopf 1207) from outside of the sound-proof room. Extracellular action potentials were amplified, filtered from background noise, and displayed on a storage oscilloscope (Tektonix 5115) as well as audiomonitored. Firing rates were determined by a gated electronic counter (Coulboum Rn-25). Neural responses were recorded on magnetic tape (Akai GX-630D-SS tape recorder) for off-line computer analysis.

Acoustic stimuli were presented through an earphone (Beyer DT48) enclosed in a brass housing, which also held a condenser microphone (Bruel and Kjaer 4134) with a J in probe tube attachment. The earphone housing was sealed around the tympanum with non-toxic silicone rubber cement (General Electric RTV-162) to provide a closed acoustical system. The absolute sound pressure level at the tympanum was monitored on a sound level meter (Bruel and Kjaer 2209). The measured sound pressure level was corrected for the frequency response of the probe tube to give the actual sound pressure level at the tympanum in dB SPL with reference to 2 × 10−5 N/m2. The frequency response of the acoustic system was flat within ±5 dB over the range of 100–4500 Hz.

Acoustic stimuli consisted of white noise and pure tones. These stimuli had a duration of too ms and a symmetrical rise-fall time of 5 ms. Stimuli were presented at 1–1·2 s intervals. The intensity of the acoustic stimuli was controlled with a Hewlett-Packard 350D attenuator.

White noise at a sound pressure level of 110 dB SPL was used as a search stimulus for exciting auditory fibres of the VIIIth nerve. When an isolated unit responded to the search stimulus, single pure tones of varying frequencies and intensities were presented to determine the tuning curve of the unit and its best excitatory frequency (BEF), i.e. the frequency at which the unit had its lowest threshold of excitation.

Response characteristics of a total of 242 primary auditory fibres from 11 adults and 346 auditory fibres from 22 froglets were studied. All auditory fibres responded tonically to bursts of pure tones for all intensities above threshold levels as shown by the post-stimulus time histograms in Fig. 2. The response of the low frequency auditory fibres was phase-locked to the stimulus in both groups of frogs (Fig. 2). Spontaneous activity in the absence of any acoustic stimulus was noted in most auditory fibres from adults and froglets.

The BEFs of adult auditory fibres fell into three populations (Fig. 3 a): a low-frequency population with a peak around 100–300 Hz, a mid-frequency population with a peak around 500–600 Hz and a high-frequency population with a peak around 1200–1400 Hz. The high-frequency selective fibres in adults all had BEFs of less than 1700 Hz. As in the adults, three populations of auditory fibres were distinguishable from early post-metamorphic frogs (Fig. 3 b). However, the distributions of these populations covered broader frequency ranges than those in adults. For instance, the BEFs of the low-frequency population in post-metamorphic frogs ranged from 100 to 800 Hz. The BEFs of the mid-frequency selective fibres ranged from 900 to 1700 Hz,and those of the high-frequency population ranged from 1800 to 2500 Hz. Note that the distribution of the froglet high-frequency population was outside that of the adults.

Each adult auditory fibre possessed a V-shaped tuning curve with a distinct BEF. Typical tuning curves from adults are shown in Fig. 4a. It is interesting to note that the high-frequency selective fibre had an upper cut-off frequency at 96 dB SPL of 3125 Hz. In general, the upper cut-off frequencies at about 100 dB SPL were below 3500 Hz and no adult high-frequency selective fibre could be stimulated beyond 4000 Hz at any intensity.

The majority of froglet auditory fibres possessed V-shaped tuning curves, although some high threshold units having broader tuning curves mimicking a U-shape were also found. Typical tuning curves of the three populations from froglets are shown in Fig. 4,b. Inspection of the tuning curve for the high-frequency selective fibre (Fig.46) showed that the upper cut-off frequency was 5500 Hz at 86 dB SPL. The upper frequency limits of all high-frequency selective fibres at about 100 dB SPL was 6000 Hz, which was considerably higher than in adults. The tuning curve of the midfrequency selective fibre in the froglet (unit F2, Fig. 4,b) had a high cut-off frequency of 3500 Hz at 88 dB SPL and resembled that of the high-frequency selective fibre of the adult (unit A3, Fig. 4,a). Such a comparison is more clearly shown in Fig. 5. The tuning curve from an adult high-frequency selective fibre (dashed line) and that of a froglet mid-frequency selective fibre (solid line) with similar BEFs and thresholds of excitation were almost identical.

Thresholds of excitation at the unit’s BEF ranged widely for both adults (22–103 SPL, Fig. 7,a) and froglets (22–132 dB SPL, Fig. 7,b), with some noticeable differences in distribution as shown by the outlined boxes in Fig. 7. It can be seen that while few adult units had thresholds greater than 100 dB SPL, 20% of the low-frequency selective fibres in the froglets had BEFs less than 800 Hz and thresholds greater than 100 dB SPL (box 1). These high thresholds cannot be attributed to a decrease in the physiological condition of the frog during the recording session, since units having lower thresholds were often subsequently encountered. High-frequency selective fibres in adults, having BEFs between 1000 and 1700 Hz, had thresholds clustered between 20 and 60 dB SPL. However, thresholds of froglet auditory fibres having BEFs in the same frequency range were widely distributed between 20–105 dB SPL, and 39 % of these fibres had thresholds exceeding 60 dB SPL (box 2). Finally, box 3 in Fig. 7 shows the distribution of thresholds for the high-frequency selective fibres that had BEFs between 1700 –2500 Hz in the froglet. This population is absent in the adult. Note that the thresholds of the high-frequency population in adults distributed over a 40 dB range, whereas in the froglets the high-frequency selective fibres had thresholds distributed over a 70 dB range.

The suppression of the auditory response of a primary fibre to an excitatory pure tone by the addition of a second tone has been defined as two-tone inhibition (Sachs & Kiang, 1968). A total of 90 auditory fibres from 6 froglets and 52 fibres from 3 adults was tested for two-tone inhibition (Fig. 3). Units of the low-frequency population were all inhibitable in adults as well as in froglets with a few exceptions. These exceptions were low-frequency selective fibres having high thresholds of excitation. Note that in the froglets, low-frequency selective fibres with BEFs extending to 675 Hz, but mostly below 500 Hz, exhibited two-tone inhibition. Data from the adults showed that generally only units with BEFs below 200 Hz, but a few extending to 600 Hz, exhibited two-tone inhibition. On the other hand, mid-and high-frequency selective units in both groups of frogs, regardless of their thresholds of excitation, did not show two-tone inhibition.

The best inhibitory frequencies of individual units were examined, and data from the adults and froglets showed some differences. The excitatory tone was fixed at 10 dB above threshold at the unit’s BEF, and the frequency and intensity of the second tone was varied to find the inhibitory tuning curve and the best inhibitory frequency (BIF), i.e. the frequency with the lowest threshold to reduce the excitatory response by 50%. The BIF was always above the BEF of each unit, and generally the inhibitory tuning curve was outside of the excitatory tuning curve at the high frequency side. For example, Fig. 8(a) illustrates excitatory and inhibitory tuning curves of an auditory unit obtained from a 28 mm froglet that had a BEF of 340 Hz and a threshold of 54 dB SPL. The auditory response to this tone at 64 dB SPL (10 dB above threshold) was inhibited with the addition of a second tone. The BIF was 1370 Hz at a threshold of 92 dB SPL, which was 28 dB (ΔI) above the intensity of the 340 Hz tone. The excitatory and inhibitory responses of this unit are shown by the post-stimulus time histograms in Fig. 8. The excitatory response at 340 Hz at 64 dB SPL (10 dB above threshold) was tonic and phase-locked. The addition of a second tone at 1370 Hz at 92 dB SPL suppressed the excitatory response by half and completely inhibited it at 97 dB SPL (Fig. 8). The distributions of BIFs are shown in Fig. 9. The BIFs of the low-frequency selective fibres recorded from froglets ranged from 700 to 1700 Hz, which corresponded to the BEF range of mid-frequency selective fibres in the froglets (and the high-frequency selective fibres in the adults). On the other hand, BIFs in the adults ranged from 485 to 990 Hz, corresponding to the mid-frequency BEF range for adults. The ΔI values for two-tone inhibition ranged from 3 to 40 dB and 8 to 42 dB in froglets and adults, respectively.

The results of the present study reveal some of the basic differences and similarities in the response characteristics of primary auditory fibres in adults and early post metamorphic bullfrogs. In each group studied, three populations of auditory fibres were found. The shapes and sharpness of the tuning curves of these populations as well as the temporal firing patternswere similar, but the distributions of BEFs differed. The BEF range of mid-frequency selective fibres in the froglets corresponded to that of the adult high-frequency selective fibres. The range of BEFs of the froglet high-frequency selective fibres extended well beyond the range of BEFs of the adult high-frequency population. The upper limit of BEFs was 1700 Hz in adults and 2500 Hz in froglets, and the upper limit of the auditory response at 100 dB SPL was 3500 Hz and 6000 Hz in adults and froglets, respectively. These results clearly indicate that early post-metamorphic bullfrogs respond to higher frequencies than do adults. This is in agreement with trends observed in comparative studies (Loftus-Hills & Johnstone, 1970; Loftus-Hills, 1973), where smaller species were shown to be more responsive to higher frequencies than larger species. In addition, it is interesting to note that the sensitivity of the high-frequency population in froglets (threshold range of 30100 dB SPL) was poorer in comparison to the sensitivity of the high-frequency population in adults (threshold range of 20–60 dB SPL). This pattern is consistent with the trends observed from comparative studies, i.e. higher frequency selectivity in smaller species is associated with higher thresholds (Loftus-Hills, 1973; Capranica et al. 1973).

In addition to the differences in high-frequency selectivity observed, there were notable differences in the sensitivity and distribution of the low-frequency population of primary fibres between the two groups of frogs. Although the thresholds of the low-frequency selective fibres were widely distributed in both groups, high threshold units were more commonly observed in froglets. Whereas the BEFs of the great majority of low-frequency selective fibres of the adults fell within a range of 100 –450 HZ.

The low-frequency population in the froglets had a broader range of 100–800 Hz. Only low-frequency selective fibres from the two groups of frogs showed two-tone inhibition. The BIFs of the adults ranged from 485 to 990 Hz and were in close agreement with previous studies on the bullfrog (Frishkopf et al. 1968; Liff & Goldstein, 1970; Feng et al. 1975), whereas the range of BIFs of the froglets was 700–1700 Hz. It is intriguing that in each group, the BIF range corresponded to the range of BEFs of mid-frequency selective fibres.

These results are in contrast to earlier work by Frishkopf et al. (1968) in which no correlation between the body size and frequency selectivity of the bullfrog peripheral auditory system was observed. More recently, Capranica & Moffat (1980) also failed to observe any significant change in the distribution of BEFs from American toads ranging from 10 to 50 g body weight. We have, however, additionally studied the BEFs of 117 single auditory fibres from the VHIth nerve of six intermediate-size bullfrogs (62–67 nnn) and found that the high-frequency population was centered around 1800–2000 Hz (which was intermediate between froglets and full-size adults) with no fibres having BEFs above 2050 Hz. The low-and mid-frequency populations of these intermediate frogs were practically identical to those of the adults. Thus, it appears that the change in the distribution of BEFs during the post-metamorphic growth of the bullfrog is a continuous and gradual process. Therefore, the degree to which the frequency selectivity of the auditory system can be correlated with body size may reflect the degree to which the peripheral structures change with body size. It is noteworthy that Narins & Capranica (1976) have also shown that in the Puerto Rican treefrog, where a sexual dimorphism in body size exists, the distribution of high-frequency selective fibres differed between adult males and females. Correlations between body size and frequency selectivity in a growing anuran may have gone undetected in previous studies if only subtle differences existed in the peripheral auditory structures among the animals used.

Behavioural studies have advocated the functional significance of the mating call structure in adult bullfrogs (Capranica, 1965). The effect adult calls have on froglets, however, has not been demonstrated. The froglets are not reproductively mature (Howard, 1978), and they obviously would not be participating in mating. However, it could be advantageous for froglets to detect the adult mating calls, since larger adults are potential predators. Interestingly enough, the low-and mid-frequency populations of auditory fibres found in the froglet corresponded to the dominant spectral energies present in the adult mating call (Capranica, 1965). In addition, the BIFs observed in the froglets were in the frequency range of 1000–1700 Hz, which corresponded to the dominant high-frequency peak of adult mating calls (Capranica, 1965). Thus, it is possible that the presence of two-tone inhibition in the froglet primary afferent fibres may provide a peripheral basis for predator avoidance. The biological significance of the high-frequency population of auditory fibres in early post-metamorphic bullfrogs is also unclear. It is not known whether froglets produce sounds.

The differences in the frequency selectivities and sensitivities of the peripheral auditory systems of adults and froglets raise questions as to what morphological mechanisms are responsible for these observed changes during post-metamorphic development. A dramatic difference can be seen in the size of the tympanum between the two groups (Fig. 1), and differences in the sizes of the middle ear cavity and columellar bones are also obvious. These morphological changes presumably alter the transmission of acoustic energy to the inner ear. Unfortunately, the middle ear transfer functions have not been studied in the two groups of bullfrogs. However, mechanical measurements of the tympanum and middle ear displacements as a function of frequency from various anuran species indicate that these structures act as a low-pass filter and dictate the upper cut-off frequency of the peripheral auditory system (Saunders & Johnstone, 1972; Moffat & Capranica, 1978). Furthermore, smaller species are correlated with higher upper cut-off frequencies. Thus, it is likely that the extended frequency range of froglets is in part attributable to a higher upper cut-off frequency in the middle ear response than that found in adults. The increase in the mass of the middle ear with body size would primarily attenuate the transmission of high-frequency sounds.

In addition to the changes in the size of the tympanum and middle ear structures, the volume of the mouth cavity also shows a dramatic increase in size. Recently it has been suggested that the resonance property of the mouth cavity plays an essential role in determining the frequency selectivity of the peripheral auditory system (Chung, Pettigrew & Anson, 1978; Pettigrew, Chung & Anson, 1978). This hypothesis was later refuted by Moffat & Capranica (1978) and by Gerhardt & Mudry (1980). The results from the adult bullfrogs using the dorsal recording approach (with the mouth cavity closed) are in close agreement with those obtained using a ventral recording approach in which the mouth cavity was held opened (Feng et al. 1975). Thus, it is unlikely that the resonance characteristic of the mouth cavity is an important factor in determining the frequency selectivity of the auditory system. Therefore, observed differences between the frequency selectivities of the adult and froglet auditory peripheries cannot be attributed to the differences in the volumes of the mouth cavities. It is worth noting, however, that the mouth cavity does play an essential role in generating the directional cues of the peripheral auditory system (Chung et al. 1978; A. S. Feng & W. P. Shofner, in preparation).

The inner ear is another possible source for the variations in frequency selectivities and sensitivities observed between adults and froglets. Anurans are unique among the vertebrates in that they possess two auditory organs (Geisler, Van Bergeijk & Frishkopf, 1964) selective to different frequency ranges. It has been demonstrated in the adult bullfrog that the low-and mid-frequency selective fibres originate from the amphibian papilla, whereas the basilar papilla gives rise to the high-frequency selective fibres (Feng et al. 1975; Lewis, Leverenz & Koyama, 1980). It is probable that the low-frequency population in froglets (100–800 Hz) is derived from the amphibian papilla, since fibres selective to this frequency range also originate in the amphibian papilla in adults. However, the origins of the mid-frequency population of the froglet cannot be directly compared to the adult organization, since the range of froglet mid-frequency selective fibres was the same as that of the adult high-frequency (basilar papilla) population. Nevertheless, the observation that the range of BIFs corresponded to the range of mid-frequency selective fibres in both adults and froglets suggests that the froglet mid-frequency population is probably derived from the amphibian papilla. Therefore, the froglet high-frequency population pressure ably originates from the basilar papilla. If these suppositions were correct, however when the adult basilar papilla, which is believed to act as a simple resonating structure (Capranica & Moffat, 1977), would have different resonance characteristics from those of froglets. Thus, some morphological changes in the basilar papilla must occur during post-metamorphic development to account for the varying resonance characteristics. Interestingly, Li & Lewis (1974) have shown that the size of the basilar papilla is smaller in tadpoles than in adults. It is possible that this relationship also holds true for early post-metamorphic and adult bullfrogs but further investigations are necessary to extend these findings to froglets and to clarify the basis for the changes in frequency selectivities and sensitivities observed between the two groups of frogs.

The authors would like to thank Z. M. Fuzessery and P. M. Narins for their criticisms of this manuscript and N. J. Shofner for her assistance with the figures. This work was supported by NSF grant 79-12271, RIAS study grant SER 78-18244 from NSF and NIH training grant HEW PHS GM 07283-05.