The role of two-tone suppression in song coding by ventral cord neurones in the cricket Teleogryllus oceanicus (Le Guillou)

1.The calling, courtship and aggression songs ofT. oceanicus have distinctive temporal patterns but their spectral compositions are very similar: the carrier frequency occurs around 4.5 kHz and harmonics occur which extend above background noise level up to at least 40 kHz (Fig. 1).

2.Units were recorded in the cervical connectives which responded to sound frequencies over the range from 1 kHz to at least 40 kHz. The sensitivities of some of the high frequency units were comparable to, or exceeded, the sensitivities of the units responding to the carrier frequencies of the species songs (Fig. 2).

3.At the level of the neck connectives, units responding to the range from 10 kHz to 40 kHz could be divided, on the basis of their prothoracic morphology and physiology, into Through Neurones and Ascending Neurones. All the recorded neurones showed two-tone suppression effects but they varied in their degree of susceptibility to suppression. Frequencies in the range 4–5 kHz were most effective in producing suppression of the responses to high frequency tones (Figs. 5, 6).

4.The two identified Ascending Neurone types had similar morphologies in the prothoracic ganglion (Fig. 3). ANA showed tonic responses to pure tones in the frequency range from 10 kHz to 40 kHz. Below 10 kHz high intensity inhibition of the response occurred and the latency increased from about 15 ms for an 18 kHz tone pulse, to about 55 ms for a 4.5 kHz tone (Fig. 4). ANB responded with 1–3 spikes to stimuli in the frequency range from 7 kHz to 40 kHz. The response latency to an 18 kHz tone pulse was about 15 ms. No responses were obtained to 50 ms stimuli in the range from 2 kHz to 7 kHz (Fig. 6). However, considerable rebound activity was produced in response to 4.5 kHz stimuli of 200 ms or more; the latency of the rebound to 200 ms stimuli was about 400 ms (Fig. 6C).

5.ANA neurones responded to the syllables of the natural songs (Fig. 7), clearly coding the temporal patterns of the three songs. Cutting the contralateral leg nerve resulted in no detectable change in the neurone's response pattern. ANB neurones responded only during the intervals between chirps and during pauses within chirps. A long period of bursting activity occurred at the end of each song sequence (i.e. 3 chirps). This rebound activity may be the result of extreme inhibition caused by the CF of the songs (Fig. 7).

6.Synthesised natural songs (Fig. 1 E) with the correct temporal parameters but filled with 4.5 kHz alone, 18 kHz alone or a combination of 18 kHz and 4.5 kHz, were produced to investigate the role of two-tone suppression phenomena in song coding. The responses of both ANA and ANB (Fig. 8) showed the interaction of excitatory and inhibitory inputs to these units during the coding of the species songs.

7.Two-tone inhibition effects are advanced as the basic mechanism for the accurate coding of temporal characteristics in the songs ofT. oceanicus. Two different strategies for the recognition of species song types in crickets are advanced.