Gerald S. Pollack

Temporal Pattern as a Cue for Species-Specific Calling Song Recognition in Crickets

Female crickets can recognize conspecific calling song from its temporal pattern alone. In Teleogryllus oceanicus, the song pattern consists of three classes of interpulse intervals arranged in a stereotyped sequence. Females recognize a model song in which the sequential order of intervals is random. This argues against the hypothesis that recognition results from matching auditory input to an internal template of the song.

Neural Processing of Acoustic Signals

Like all sensory systems, auditory systems have been shaped by the stimuli that carry meaning for the animals they serve (see Hoy, Chapter 1; Michelsen, Chapter 2; Romer, Chapter 3; Robert and Hoy, Chapter 6; Barth, Chapter 7; Fullard, Chapter 8). It thus comes as no surprise that auditory neurons and neural circuits are specialized to detect and analyze those sounds that carry behaviorally important information. The strong effect of selective pressure is particularly evident among insects, where hearing has evolved independently many times (Fullard and Yack 1993; Hoy, Chapter 1), and often seems to be a “special-purpose” modality that serves restricted and obvious behavioral functions. Because of this close relationship between biological function and auditory neurophysiology, the first section of this chapter focuses on the behavioral functions of sound and on how biologically meaningful information is represented by the physical parameters of acoustic signals. Subsequent sections examine how this information is analyzed by the nervous system.

Discrimination of calling song models by the cricket,Teleogryllus oceanicus: the influence of sound direction on neural encoding of the stimulus temporal pattern and on phonotactic behavior

SummaryRecordings were made from an identified auditory neuron, the omega neuron, in the cricketTeleogryllus oceanicus. Models of the conspecific calling song and of the song of another species were presented either singly or simultaneously, and the degree to which the temporal pattern of the conspecific model was encoded in the neurons spike train was determined. When a single stimulus was presented alone, its temporal pattern was faithfully reflected by the cellss spiking activity, no matter what the azimuth of the broadcasting loudspeaker (Fig. 3). When two stimuli were presented simultaneously from opposite sides, encoding of the pattern ipsilateral to the recorded neuron was interfered with only slightly by the contralateral pattern, as long as the two loudspeakers were sufficiently separated (Figs. 2, 3, 4). When the loudspeakers were each 15° from the crickets midline, however, the encoding of the temporal pattern of the ipsilateral song model was severely disrupted (Figs. 3, 4). Bilateral interactions are important in determining the response level of the neuron, but do not appear to contribute to the direction-selective encoding of the stimulus temporal pattern (Figs. 5, 6).Phonotactic steering movements of tethered, flying crickets were recorded under stimulus conditions similar to those used in the neurophysiological tests. Under one-stimulus conditions, crickets attempted to turn towards the conspecific model for all tested speaker locations. The heterospecific model elicited reliable steering behavior when it was broadcast from azimuths of 90° and 60°, but often failed to elicit consistent responses when the speaker was positioned closer to the crickets midline (Figs. 7, 8A and 8B). Responses to the heterospecific pattern were smaller in amplitude than those to the conspecific song model (Figs. 7, 8B). Under two-stimulus conditions, the conspecific model was consistently preferred over the heterospecific song for all tested speaker locations in half the tested crickets. In the remaining animals, preference for the conspecific pattern was only evident for the larger loudspeaker azimuths (Figs. 7, 8C).These results demonstrate that simultaneouslypresented stimuli can be represented separately in the nervous system as a consequence of auditory directionality. It is postulated that the crickets ability to choose between these stimuli may result from the interactions between two bilaterallypaired song recognizers, each of which may be driven primarily by sound stimuli from one side.

Who, what, where? Recognition and localization of acoustic signals by insects.

Insects, like all hearing animals, must analyze acoustic signals to determine both their content and their location. Neurophysiological experiments, together with behavioral tests, are beginning to reveal the mechanisms underlying these signal-analysis tasks. Work summarized here focusses on two issues: first, how insects analyze the temporal structure of a single signal in the presence of other competing signals; and second, how the signals location is represented by the binaural difference in neural activity.

Phonotaxis to individual rhythmic components of a complex cricket calling song

SummaryThe calling song ofTeleogryllus oceanicus consists of a chirp section and a trill section. We tested the phonotactic attractiveness to femaleT. oceanicus of the chirp and trill components of the song by permitting crickets to choose between synthetic song models in two-choice behavioral assays.In a flight phonotaxis assay tethered flying crickets were simultaneously offered two different song models from loudspeakers located to the left and right. Song preference was indicated by postural changes associated with steering attempts. In the flight assay females preferred a song derived from the chirp component of the normal song pattern to one derived from the trill component (Figs. 3, 4A). They also preferred a song which consisted of 100% chirp over the normal song pattern, which is only 16% chirp (Fig. 6A). And, they preferred the normal song pattern to a chirp-free, trill-derived model (Fig. 5A). Two other song models, in which the percent of time occupied by the chirp was approximately twice normal, were also preferred to the normal pattern (Fig. 7). Thus, the chirp section of the song was more attractive than the trill section. Although the two components differed in their relative attractiveness, either was preferred to a heterospecific song model (Figs. 8A, 9A).Five of these experiments were also performed using a walking phonotaxis assay. Crickets were released in an arena in which two loudspeakers broadcast different song models. Song preferences were indicated by the relative number of times each loudspeaker was approached. For three experiments (Figs. 4, 5, 8) the results obtained with the walking assay were qualitatively similar to those obtained with the flight assay. In two experiments there was no apparent discrimination of the songs in the walking assay, although the songs were discriminated in the flight assay (Figs. 6, 9). Possible reasons for the differences observed between the assays are discussed.

Frequency and temporal pattern-dependent phonotaxis of crickets (Teleogryllus oceanicus) during tethered flight and compensated walking

SummaryPhonotactic responses ofTeleogryllus oceanicus were studied with two methods. Tethered crickets were stimulated with sound while they performed stationary flight, and steering responses were indicated by abdominal movements. Walking crickets tracked a sound source while their translational movements were compensated by a spherical treadmill, and their walking direction and velocity were recorded.During both flight and walking, crickets attempted to locomote towards the sound source when a song model with 5 kHz carrier frequency was broadcast (positive phonotactic response) and away from the source when a song model with 33 kHz carrier frequency was used (negative phonotactic response) (Figs. 2, 4).One-eared crickets attempted, while flying, to steer towards the side of the remaining ear when stimulated with the 5 kHz model, and away from that side in response to the 33 kHz model (Fig. 3). While walking, one-eared crickets circled towards and away from the intact side in response to the 5 kHz and 33 kHz models, respectively (Fig. 6).Positive and negative responses differed in their temporal pattern requirements. Phonotactic responses were not elicited when a non-calling song pattern (2 pulses/s) was played with a carrier frequency appropriate for positive phonotactic responses (5 kHz), but this pattern did elicit negative responses with 33 kHz carrier frequency (Figs. 7–10). When an intermediate carrier frequency, 15 kHz, was used, the response type (positive or negative) depended on the stimulus temporal pattern; the calling song pattern elicited primarily positive responses, while the non-calling song pattern elicited negative responses (Figs. 11, 12, 14, 15). A curious phenomenon was often observed in the flight steering responses; while most responses to 15 kHz song pattern were primarily positive, they often had an initial negative component which was supplanted by the positive component of the response after approximately 2–5 s (Figs. 11, 12).In recent experiments onGryllus campestris, Thorson et al. (1982) described frequency-dependent errors in phonotactic direction (anomalous phonotaxis) and showed how such errors might arise from the frequency-dependent directional properties of the crickets auditory apparatus. Our findings, particularly the dependence of response type on temporal pattern when 15 kHz carrier frequency was used, argue that frequency-dependent directional properties alone cannot account for positive and negative phonotaxis inT. oceanicus. Rather, these represent qualitatively different attempts to locomote towards and away from the sound source, respectively.We discuss the possibility that central integration of these opposing tendencies might contribute to anomalous phonotaxis.

A Behavioral Role for Feature Detection by Sensory Bursts

Brief episodes of high-frequency firing of sensory neurons, or bursts, occur in many systems, including mammalian auditory and visual systems, and are believed to signal the occurrence of particularly important stimulus features, i.e., to function as feature detectors. However, the behavioral relevance of sensory bursts has not been established in any system. Here, we show that bursts in an identified auditory interneuron of crickets reliably signal salient stimulus features and reliably predict behavioral responses. Our results thus demonstrate the close link between sensory bursts and behavior.

Phonotactic specificity of the cricketTeleogryllus oceanicus: intensity-dependent selectivity for temporal parameters of the stimulus

SummaryWe have investigated the effects of alterations of several temporal parameters of auditory stimuli, as well as of stimulus intensity changes, on the attractiveness of these stimuli to femaleTeleogryllus oceanicus, as measured by monitoring sound-elicited flight steering responses. AlthoughT. oceanicus has a rhythmically complex calling song, females are attracted by a simpler model consisting of regularly repeating sound pulses. We have found that the two major temporal features of this model, sound pulse duration and pulse repetition rate, are both important for eliciting phonotactic steering responses.Stimuli with altered temporal features had intensity thresholds indistinguishable from the control stimulus (Fig. 3). The majority of crickets, however, ceased to respond to the altered stimuli when the stimulus intensity was sufficiently increased (Figs. 4–7). In some cases, intensity increases resulted in a reversal of the steering response from positive to negative (Fig. 10). Effects of altered temporal parameters were also apparent at lower stimulus intensities, where the amplitudes of steering responses to stimuli with altered parameters were smaller than those in response to the control stimulus (Figs. 8, 9).We considered the possibility that the cessation of responsiveness to stimuli with altered temporal features was due to a temporal pattern-specific diminution of binaural cues for sound localization at high intensities. Experiments performed with unilaterally deafened crickets (Fig. 11) led us to conclude that this was not the case, and that our findings instead reflect the properties of the song recognition mechanism.

Phonotaxis in flying crickets: Neural correlates

Abstract Tethered, flying crickets perform stereotyped steering movements when stimulated with sound. Electrophysiological recordings show that identifiable motoneurones respond to acoustic stimulation in a manner which reflects the phonotactic behaviour of the intact cricket. Both steering behaviour and its neural correlates reflect the fine temporal structure of the acoustic stimulus. Decapitation of the cricket eliminates steering movements in response to sound.

Directionality of acoustic orientation in flying crickets

SummaryAbdominal flexions associated with flight steering were measured in tethered flyingTeleogryllus oceanicus stimulated with a model of conspecific calling song presented at various intensities and from many directions.Flexions increased in size with stimulus intensity until a plateau level was reached. Flexion amplitude was then approximately constant over a range of 20–30 dB, and decreased at still higher intensities (Figs. 2, 3). The shape of this intensity function results from binaural processing; in unilaterally deafened crickets flexion amplitude increased monotonically with stimulus intensity (Fig. 4).Abdominal flexions were graded with respect to sound location; they were larger for laterally placed sound sources and smaller for sound sources near the midline (Figs. 5, 6).A model for the specification of flight steering movements is presented which accounts for our findings (Fig. 7).