As a neuroethologist (someone who studies the neural basis of natural, usually innate, actions) I am especially interested in how the brain produces behavior. How do the neurons that are devoted to song production actually generate the appropriate patterns? What do males and females hear and how does it affect their songs? Several research projects in the labs are tackling these problems and we are having some success.

With respect to production, we have been able to record from the laryngeal nerve of singing frogs (Yamaguchi and Kelley, 2000) and have learned that distinctive song patterns used by males and females are generated in the central nervous system (see Figure 6). We have also been examining vocal motor neurons in detail in a slice preparation (Yamaguchi et al., 2003) and have found striking differences in currents between the sexes; modeling suggests that these currents account for firing properties of song neurons. Since the vocal motor neurons do not, by themselves, generate vocal patterns, we are working on defining their afferent inputs, the constellation of interneurons responsible for different patterns (Zornik and Kelley, 2000). One of the long-term goals of this project is to be able to study song production in an isolated brain preparation so that we can determine how the ensemble activity of neurons generate song types. An allied goal is to determine how the program of sexual differentiation results in male- and female-specific patterns of pre-vocal activity and how sex differences in the expression of different ion channels contribute to sex differences in vocal expression. These ongoing studies are using a variety of approaches: single unit recordings in singing frogs, pathway tracing in the isolated brain, patch clamping in brain slices are some examples.

Figure 6 Recordings of nerve compound action potential while male and female frogs are singing (from Yamaguchi and Kelley, 2000). The patterns of nerve activity (lower traces) are an exact match to sound production (lower traces).

Figure 7 Auditory and vocal pathways in a schematic image of the Xenopus brain.

What about acoustic perception? For the moment we are focusing on the way the male perceives two female songs: ticking and rapping. Rapping is the female acoustic aphrodisiac we first recorded in South Africa (Tobias et al., 1998). A rapping female or a broadcast tape of rapping excites male calling and approach. In contrast, ticking silences males. These two song types are monotonous trills that vary only in click rate.

Rapping is a rapid trill and can be very intense; ticking is a slower, softer trill. Gary Rose and his colleagues have shown that neurons in the frog inferior colliculus are tuned to rate (Alder and Rose, 1998;2000) and that these neurons can count (Edwards et al., 2002). How rate tuning (the preference of neurons, expressed as a change in firing rate, for a particular rate of click production or amplitude modulation) emerges from response properties of auditory neurons is an interesting problem that will benefit from computational and modeling approaches. The laminar nucleus receives the most direct auditory input in Xenopus laevis (Edwards and Kelley, 2001) and we wish to determine whether we can find rate-tuned neurons in this or other toral nuclei. We expect that neurons tuned to slow rates will participate in the inhibition of calling produced by ticking, the slow trill; neurons tuned to rapid rates should play a role in stimulating calling as this is the effect of rapping. Understanding how these biologically critical signals are decoded will also require understanding how different vocal patterns are produced, as described above.

Working with Jakob Christiansen-Dahlsgaard at Southern Denmark University in Odense, Taffeta Elliott and I have begin to record from the auditory system of Xenopus laevis in order to address these questions. Our frogs hear underwater but as terrestrial mammals, we would like to record from their neurons in air! We have recently solved this problem of impedance matching and have characterized the properties of neurons in the eighth nerve and the dorsal acoustic medulla before tackling the torus itself.