Encyclopedia of Computational Neuroscience

Living Edition
| Editors: Dieter Jaeger, Ranu Jung

Auditory Processing in Insects

  • R. Matthias HennigEmail author
  • Bernhard Ronacher
Living reference work entry
DOI: https://doi.org/10.1007/978-1-4614-7320-6_321-1

Keywords

Spike Train Modulation Transfer Function Auditory Processing Interaural Time Difference Spike Rate 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Synonyms

Definition

Auditory processing in insects serves to extract relevant information from acoustic signals about identity and location of acoustic objects, usually in the context of mate attraction and predator avoidance. For this goal, insects process spectral information from the carrier frequency of a signal and obtain temporal information from the sound’s amplitude modulation pattern (the envelope). Auditory processing in insects is constrained by size in several aspects: first, the signal often contains ultrasonic frequencies due to small sender size; second, for localization, insects have only poor directional cues because of the small distance between their ears; and third, due to their small brains, the auditory processing capacity is limited to a small number of neurons.

Detailed Description

Overview and Background

Large Diversity of Hearing Insects and Their Ears

The sense of hearing has evolved in vertebrates and arthropods, and hearing organs are known from several orders of insects. Ears have evolved in rather different locations of their body, not only on the head but also on the thorax, abdomen, and even wings and legs (Fig. 1a, Fullard and Yack 1993). Consequently, numerous neuronal substrates for the processing of acoustic information are known that evolved from a mechanosensory modality employed in a different context (proprioception by chordotonal organs: Meier and Reichert 1990; van Staaden and Römer 1998). In insects, two types of ears are known that are sensitive to two different physical attributes of sound, that is, the movement of particles in an elastic medium: tympanal ears that are specialized to sense changes in sound pressure and antennae or filiform hairs that are sensitive to particle velocity (Fig. 1bd, Michelsen 1979).
Fig. 1

Insect ears and sound transduction. (a) Evolution of insect ears in different body locations (b–d). Types of ears: (b) sound pressure receiver (mammals and humans); (c) pressure difference receiver (insects); (d) sound velocity receiver (filiform hair of insects); S sensory cells, p sound pressure, pd sound pressure difference, and v particle velocity. (e) Sensory transduction at a tympanic ear: 1, sound wave; 2, vibration of membrane and movement of mechanosensitive ion channels due to sound pressure; and 3–4, membrane potential and elicited action potential in sensory cell. (f, g) Intensity-response curves of an auditory neuron in crickets for different background intensities and different carrier frequencies (f, 3 kHz; g, 16 kHz). At increasing background intensities as indicated by the top symbols in (f, g), the response curves shift to higher intensities ((a) Modified from Fullard and Yack 1993, with permission; see also there for explanation of numbers. (bd) From Penzlin 2005, with permission. (e) From Gollisch and Herz 2005, with permission. (f, g) From Hildebrandt et al. 2011, with permission)

Goals of Hearing and Auditory Processing

The ability to hear sound and to extract relevant information from an acoustic signal has evolved in three functional contexts. Notably, the production of sound alone is not a sufficient indicator of hearing ability as many insects produce defensive sounds when threatened (e.g., Bura et al. 2011). Conversely, numerous insects can hear sound but are themselves mute (Riede 1987; Riede et al. 1990).

Acoustic Communication

In many species the perception of acoustic signals occurs in the context of acoustic communication for mate recognition and mate localization. The most prominent examples stem from grasshoppers, crickets, and bushcrickets (orthoptera), but cicadas and moths have also evolved elaborate acoustic signals. A major goal of intersexual acoustic communication is to discriminate conspecific from heterospecific signals which helps to avoid fitness losses (Fig. 2). Sexual selection by female choice for song signals of particularly attractive mating partners is also well known (von Helversen and von Helversen 1994; Andersson and Simmons 2006). Insects employ very stereotyped signals for acoustic communication, the production and recognition of which has a strict innate basis (Bentley and Hoy 1972, von Helversen and von Helversen 1975a,b, 1987). For this reason, the acoustic communication of grasshoppers, crickets, and bushcrickets has served as a model system to study the mechanisms of neuronal processing within the auditory pathway by carefully designed behavioral experiments and recordings of neuronal activity of single cells (Schildberger and Elsner 1994; Huber et al. 1989; Gerhardt and Huber 2002).
Fig. 2

Acoustic communication in grasshoppers and crickets. (a, f) A sender, the male, produces a specific sound signal that is transmitted and perceived by a receiver, the female. (a) In some grasshoppers, the female responds to the male with her own song (upper trace). Songs of males and females have different pulse shapes (lower traces). (b, c) The power spectra differ between males (b) and females (c) in relative content of low (L)- and high (H)- frequency components. (d, e) Females and males respond selectively to the relative content of low- and high-frequency components in a signal. (d) Females prefer the combination of low and high components as they occur in the male song (as in b). (e) Males prefer signals with only low-frequency components as they are typical for a female song (as in c). (g) The spectral content of a song of a cricket as in (f) is narrowly tuned to a peak frequency (5.0 kHz). (h) Auditory tuning for the specific carrier frequency of a conspecific song for different species of crickets (frequency axis was normalized to the respective best frequency; P.p. Paroecanthus podagrosus, G.b. Gryllus bimaculatus, G.c. G. campestris). Note the differences in the sharpness of tuning ((a–e) From von Helversen and von Helversen 1997; (f, g) modified from Huber 1992; and (h) from Schmidt et al. 2011, with permission)

Since the nervous system of insects is small, all computations must be performed by relatively few neurons. This likely imposes a strong pressure for an efficient and sparse representation of the external world. A big advantage of investigating nervous system processes via the study of communication signals is that an animal’s “external world” can be reduced to a manageable set of relevant stimuli, the stimulus space. This rather small set of highly relevant stimuli allows one to characterize how the stimuli are represented within the nervous system and how these representations are transformed at different stages of processing.

To allow for successful communication, signal properties must be matched to the sensory characteristics of receivers. A straightforward example is the frequency tuning of a cricket female’s ears to the narrowband signals of the males (see Fig. 2g, h). However, since in many species the temporal pattern of communication signals conveys the species-specific information, sender and receiver should be matched not only for carrier frequency but also for temporal characteristics. Here, a new problem arises: temperature. Since the function of neurons and muscles is strongly temperature-dependent (Janssen 1992; Robertson and Money 2012) and in general insects’ body temperatures vary directly with ambient temperature, a signal’s temporal pattern will vary with ambient temperature. This creates a recognition problem for the receiver’s auditory system, in particular, if sender and receiver differ in their body temperatures (von Helversen and von Helversen 1987; Gerhardt and Huber 2002; Ronacher et al. 2004)

Detection and Avoidance of Predators

Although hearing in insects evolved more than 200 million years ago (MYA), the appearance of bats about 60 MYA sparked an explosion of insect species with ears (Fig. 1a, Fullard and Yack 1993; Hoy et al. 1998; Stumpner and von Helversen 2001). While some species modified already-existing ears for the detection of ultrasound, numerous other insects that had previously lacked hearing evolved ultrasonic ears at this time (Miller and Surlykke 2001). For moths this acquired sensory ability may even have served as a starting point for the evolution of intraspecific acoustic communication. Within their auditory pathways, insects employ different neuronal substrates for the processing of intraspecific signals and sounds from predators (see below: categorical perception in crickets and parallel processing of information). Some species perform stream segregation within individual neurons by spectral and temporal cues (Schul and Sheridan 2006). Numerous species of grasshoppers have ears and hearing abilities, but produce no sound during mate attraction. For these species, hearing likely serves for predator detection only, and predation was the selection pressure under which ears evolved or were conserved (Riede 1987; Riede et al. 1990; Lehmann et al. 2007).

Host Finding

Hearing has evolved independently at least twice in parasitic flies in the context of finding hosts for their eggs. The performance of flies in localizing their host by its sound signal is most impressive, in view of their tiny ears located underneath the head (Robert et al. 1996; Lakes-Harlan et al. 1999). With these tympanal ears flies can detect and localize not only the broadband sounds of bushcrickets and cicadas but also the pure tone signals emitted by crickets. For that goal flies possess sharply tuned hearing, as is evident from the sensitivity of their sensory receptors and auditory interneurons (Stumpner and Lakes-Harlan 1996; Robert and Hoy 1998; Robert and Göpfert 2002). Another example of host finding via auditory cues is a bloodsucking corethrellid fly that is attracted by frog calls (Bernal et al. 2006).

Two Types of Ears in Insects and Their Constraints by Size and Signals

Generally, two principal types of ears are known: particle velocity receivers and sound pressure receivers (Fig. 1bd, Michelsen 1979; Faure et al. 2009). (1) Particle velocity receivers exploit the vector component of sound particles close to the sound source. The antennae (arista) of flies, especially of Drosophila and mosquitoes, are a well-investigated model system for sound perception, auditory transduction, and active sensing (Robert and Göpfert 2002). Many insects, as well as arthropods in general, employ filiform hairs (located on the abdomen, cerci, legs, or other parts of the body, Fig. 1d) of different length to detect sounds of predators and prey (Barth 2002). A general property of particle velocity receivers is their limitation to lower frequency ranges (< 500 Hz). Low frequency signals from small senders commonly have low amplitude and thus small range (e.g., wing movements by Drosophila), and the perception of particle velocity is then usually limited to the near field of the sender (a distance of a few wavelengths). Since the vector component of sound is perceived, these ears also provide directional information (Michelsen 1979; Faure et al. 2009). Their sensitivity matches or even surpasses that of tympanal ears (Robert and Göpfert 2002). (2) The perception of sound pressure is mediated by tympanal ears and follows the same principles as in vertebrates including humans (Fig. 1b, c, Montealegre et al. 2012). Tympanal ears in insects arose from the cuticular surface of their exoskeleton under which large air sacs derived from tracheal tubes, the respiratory system of insects, were located. Since insects possess an abundance of mechanosensory proprioreceptors for monitoring the strain and movement of their cuticle, auditory organs were prone to evolve from chordotonal organs in almost any part of the body (e.g., the legs, thorax, abdomen, and wings, Fig. 1a, Fullard and Yack 1993; van Staaden and Römer 1998). Tympana are usually small, and the sensitivity of the ears often extends into the ultrasonic frequency range. Sound localization via tympanal ears in larger vertebrates (Fig. 1b) requires the computation of interaural intensity and time differences as pressure receivers do not respond to the vector component of sound (Brown 1994; Yost 2000). Due to the small size of insects, intensity differences and in particular interaural time differences are too small to be exploited directly. Insects circumvent this problem by using tympanic pressure difference receivers, in which an internal connection by tracheal tubes exists between both ears (Fig. 1c). Small frogs, lizards, and birds face similar problems and also have developed pressure difference receivers. In these ears, the vibration amplitude of the tympanum is determined not only by the sound pressure at the outer side but also by the sound traveling through the body to the inside of the tympanum (Autrum 1942). The resulting pressure difference between the inside and outside will then determine the tympanal vibration. The amplitude of these vibrations depends on sound direction, because the phase angle of a given sound frequency at both sides of the tympanum is also dependent on the direction of the incident sound wave (Michelsen et al. 1994).

Themes of Auditory Processing in Insects

A major challenge in summarizing the capacities for auditory processing of insects results from the overwhelming diversity of hearing species, of functional ears, and of the different designs of auditory pathways. The subchapters below therefore give only a brief overview of the computational capabilities of insects for different tasks and under different constraints.

For auditory processing, insects exploit numerous general principles of sensory processing that are well known from other modalities and from vertebrates, including mammals. Among these are the capacity for sound frequency analysis by a traveling wave, tonotopic representations and formation of internal neuronal maps, parallel processing of information, the timing and balance of excitation and inhibition for feature extraction, lateral and contralateral inhibition for contrast enhancement, transformation of coding from a temporal code to a place code, burst coding, resonant properties of neurons, selective attention, and even stream segregation.

However, insects face constraints on the computational power provided by their small brains. The concept of identified neurons, in which individual neurons could be identified by their morphology and physiology, arose from neurobiological research in insects and other arthropods (Huber and Markl 1983). Computations performed by thousands of neurons in mammals may find their counterpart in a single identifiable neuron of an insect, which illustrates an impressive compression of function (e.g., contralateral inhibition for directional hearing is mediated by the lateral superior olive in mammals and by a single local interneuron, ON1, in crickets and bushcrickets (Grothe 2000; Selverston et al. 1985; Römer and Krusch 2000).

Notably, the processing and coding capacity of insect nervous systems is restricted to relevant tasks. Peripheral filters and computations for instance aid in reducing the required processing power. Therefore, the ears of insects and their auditory pathways are by no means all-purpose devices, and processing is rather specific to function. Examples include crickets that distinguish only 2 categories of sound (mate signals and predators, Hoy 1989) and the tympanic ear of a moth that is equipped with only 2 sensory cells for bat detection (Boyan and Fullard 1988). Generally, insects also employ simple algorithms for processing at the cost of acuity, for example, by computing acoustic hemispheres rather than localizing the angle of a sound source (von Helversen 1997; Römer and Krusch 2000). Nevertheless, insects are by no means imprecise. At least in some species their performance for temporal resolution in a gap detection task has a precision in the millisecond range, rivaling that of humans (von Helversen 1972; Prinz and Ronacher 2002).

Basic Steps of Auditory Processing: Transduction and Information Coding by Sensory Neurons

In both tympanal ears and particle velocity receivers, sound induces the vibration of a structure (a thin membrane or a lever on a flexible pivot, Michelsen 1979; Robert and Göpfert 2002). The mechanical oscillation distorts the dendrite of a scolopidial cell, which is attached to the lever, the tympanum, or a tracheal tube (Fig. 1e1, e2). This distortion opens mechanosensitive ion channels, which are not yet fully characterized, and the resulting current transduces the movement into a change of the cell’s membrane potential (Fig. 1e3; Gollisch and Herz 2005). Evidently, the membrane potential of sensory neurons cannot follow the fast vibrations of the tympanal membrane in the kilo-Hertz range; rather, the membrane potential depends on the instantaneous sound pressure level. Similar to vertebrates, frequency discrimination occurs according to a frequency-place transformation (Montealegre et al. 2012). Either in the sensory neuron itself or in a downstream neuron, the membrane depolarization is then translated into a series of action potentials, a spike train, which encodes the sound envelope by modulations of the spike rate (Fig. 1e4). Spike rates of auditory afferents can be high, up to 400 Hz or more (Römer 1976).

Although insect and vertebrate ears obviously evolved independently, and the insect mechanoreceptors (scolopidia) differ from auditory hair cells of vertebrates, in the last decade many unexpected commonalities were detected. Recently, active processes were found in insect ears that serve to attain and adjust their formidable sensitivity, similar to the function of outer hair cells in the mammalian cochlea (Robert and Göpfert 2002; Nadrowski et al. 2011). In both locusts and bushcrickets, traveling waves were observed to play an essential role in frequency discrimination. As in the mammalian cochlea, traveling waves exhibit peaks at different locations, depending on sound frequency (Windmill et al. 2005; Hummel et al. 2011; Montealegre et al. 2012). In addition, in central projections of auditory afferents, there is a tonotopic representation of frequencies (Römer 1983; Römer et al. 1988; Stumpner 1996; Stölting and Stumpner 1998). Finally, the development of the auditory organ of Drosophila depends on similar genes, e.g., of the atonal family, as the vertebrate ear (Senthilan et al. 2012).

Although the temporal resolution of insect ears can match that of vertebrates in some respects (see below), the frequency resolution of insects is generally inferior to that of vertebrates. In crickets we find a kind of “categorical response,” by which the frequency scale is segmented into two parts: sound pulses with carrier frequencies above 15–20 kHz (up to 100 kHz) evoke an avoidance response, whereas sound pulses with frequencies between 3 and 10 kHz are attractive and evoke a positive steering response when flying (Moiseff et al. 1978; Hoy 1989). Remarkably, however, in an interneuron of a bushcricket, a sharpening of the broader frequency tuning of auditory receptors by frequency-dependent “lateral” inhibition similar to vertebrates has been observed (Stumpner 1997). The sharpness of frequency tuning of auditory neurons is commonly described by the Q10dB score. This gives the ratio between the characteristic frequency (i.e., the frequency of the neuron’s lowest threshold) and the width of the tuning curve 10 dB above the lowest threshold. Typical values for insects are between 0.5 and 2.5 and only rarely extend to 3.5 (Hennig et al. 2004), whereas in vertebrates, we find much higher values, between 1 and 25 (and up to 400 in the acoustic foveae of bats, Suga et al. 1997).

The dependency of spike rate on sound intensity is commonly depicted by the f-I curve (firing rate vs. intensity, also rate-intensity curve; Fig. 1f, g). The dynamic range between threshold and saturation indicates the region of discriminable sound intensities. The steepness of the f-I curve determines how well small sound pressure differences can be discriminated – which is important for directional hearing and when fine modulation details of a signal have to be assessed (compare Fig. 1f, g). A steep f-I curve, however, has the disadvantage that it covers only a small part of the relevant sound intensity range which may extend to 100–120 dB SPL (Fig. 1f). (Note that the dB SPL scale is a logarithmic scale that expresses sound pressure relative to 20 μPa; a 100 dB sound has a 105 larger sound pressure compared to the human hearing threshold at 1 kHz.) As in other sensory systems, the intensity range problem may be solved by adaptation, by which the f-I curve can be adjusted to the average ambient sound pressure levels (Benda and Herz 2003; Benda and Hennig 2008). Spike frequency adaptation as early as the level of sensory neurons helps to attain a certain degree of intensity invariance that is important for object identification and behavioral decisions (Benda and Hennig 2008). However, in different neuron types we may find different biophysical realizations, ranging from cell-intrinsic spike-triggered adaptation currents to inhibitory inputs and presynaptic adaptation mechanisms (Gollisch and Herz 2004; Hildebrandt et al. 2009). A complementary way to cope with the large range of encountered sound intensities may be a kind of “range fractionation.” In locusts and bushcrickets, for example, we find auditory receptors with a rather limited dynamic range of 20–30 dB. Since different receptors exhibit a similarly broad frequency tuning but very different thresholds, intensity discrimination is possible over a broad range (Römer 1976; Römer et al. 1998).

Obviously, the spike trains of sensory neurons are the only information any central nervous system has about events in the external world. In other words, the brain has to infer the structure of the outer world from the spike trains arriving from various sense organs. The very successful stimulus reconstruction methods seek to understand signal processing from the viewpoint of the central nervous system and to infer information about an external stimulus from the spike trains of sensory neurons (Rieke et al. 1997). The basic idea is to reconstruct the stimulus envelope from spike trains that have been recorded in response to that stimulus. By this procedure we can estimate how much information the CNS can obtain about a sensory stimulus and what aspects of the stimulus are lost. For example, from the spike trains of the locusts’ auditory afferents, stimuli with large modulation amplitudes can be reconstructed more accurately than stimuli with small modulation depths (Machens et al. 2001). Remarkably, grasshopper songs seem to be matched to this feature of the receiver’s auditory periphery. This investigation has further shown that in the very periphery of the auditory pathway, a single sensory neuron transmits a high amount of information, up to 180 bits/s (Machens et al. 2001).

Processing of Signal Envelopes Within the Auditory Pathway

Temporal Resolution and Temporal Integration

The communication signals of many species contain fast amplitude modulations that are evaluated by females to assess the attractiveness of potential mates. Hence, we must ask whether there are neuronal constraints that determine the behavioral limits of temporal resolution. Two widely used paradigms to investigate these limits are modulation transfer functions (MTF) and gap detection (for reviews see Green 1985; de Boer 1985; Michelsen 1985; Viemeister and Plack 1993; Joris et al. 2004). In the MTF paradigm either random modulations in a certain frequency band or sine wave modulations are used. The latter paradigm can be applied in neurophysiological as well as in behavioral experiments. Stimuli with appropriate carrier frequencies are presented that exhibit sinusoidal amplitude modulations of different frequencies and reveal what range of modulation frequencies the system is able to represent, and if there are specific modulation frequencies to which a system responds particularly well, see Fig. 3a for an example. The black dots in this diagram show the attraction of cricket females to a 4.5 kHz tone that was amplitude modulated at frequencies between 1 and 50 Hz; two regions of enhanced attractiveness around 3 and 30 Hz are obvious (Wendler 1989; Hennig 2009). In spike train recordings one can determine the average spike count (rate, r-MTF) or evaluate how well the spikes are locked to a period of the stimulus envelope (temporal, t-MTF). r-MTF reveals a neuron’s filter properties, e.g., high-pass, low-pass, band-pass, or band-reject features for sound pulse rates. The t-MTF, in contrast, indicates how well fast modulations can be resolved by a neuron. One should be aware, though, that the construction principle of MTFs is based on a large amount of averaging, and therefore, the information provided by single spike trains may be lower than that suggested by a MTF (Wohlgemuth et al. 2011).
Fig. 3

Temporal resolution of amplitude modulations in grasshoppers and crickets. (a) Behavioral modulation transfer function of crickets (filled symbols, stimuli as in 1 and 2 at right). Best responses are obtained if lower and higher modulation frequencies for pulse and chirp (i.e., groups of pulses) are combined in one stimulus as in patterns 3 and 4 at right (open symbols). (b) Gap detection in a grasshopper measured in behavioral experiments (black curve) and neuronal response (red curve, AN4). (c) Response of the AN4 neuron to uninterrupted and gap containing sound syllables. This neuron responds to sound onset first with a deep inhibition, an IPSP (arrows), followed by excitation and spikes (upper traces). In the interrupted stimuli, each onset after a gap triggers the IPSP anew which leads to an effective suppression of spiking (lower traces, adapted from Ronacher & Stumpner 1988). (d) Gap detection in crickets (behavioral data) ((a) From Hennig 2009; (b) modified from von Helversen 1972 and Franz and Ronacher 2002; (c) modified from Ronacher and Stumpner 1988; and (d) from Schneider and Hennig 2012, with permission)

The second paradigm, gap detection, has been applied in behavioral experiments to grasshoppers and crickets (von Helversen 1972; von Helversen and von Helversen 1997; Schneider and Hennig 2012). Grasshoppers detect gaps of 2–3 ms duration and in this respect are not inferior to vertebrates (Fig. 3b, Prinz and Ronacher 2002). This high resolution of gaps seems to be mediated by the specific interactions between inhibition and excitation in one identified interneuron (Fig. 3c; Ronacher and Stumpner 1988). The temporal resolution of cricket ears is lower compared to grasshoppers; minimal detectable gap widths are between 6 and 8 ms (Fig. 3d, Schneider and Hennig 2012). One reason for this may stem from the males’ sound production system: cricket songs are produced by a resonant mechanism which precludes very fast amplitude changes (Bennet-Clark 1998). Hence, on the receiver’s side there is no need to push the temporal resolution to extremes. In addition, compared to broadband signals, which are typical in many grasshoppers and bushcrickets, pure tone signals, such as those produced by many crickets, tend to be more strongly affected by random amplitude fluctuations when traveling through the habitat, which also sets limits for temporal resolution (Römer and Lewald 1992).

Temporal integration refers to the time-intensity trading paradigm. In these experiments the minimal audible threshold was found to depend on the duration of the stimuli used. To detect very short stimuli, e.g., of 5 ms duration, higher sound intensities are necessary than for longer, e.g., 100 ms, stimuli. The product of sound intensity and stimulus duration determines the threshold up to durations around 200–300 ms, whereas for longer stimuli, the threshold stays constant (Green 1985). Hence, we are confronted with an apparent discrepancy, the temporal integration-resolution paradox (de Boer 1985): gap detection and MTF paradigms yield time constants in the order of 1–6 ms, whereas from the time-intensity trading paradigm, we are left with time constants in the range of 150–300 ms (for a discussion of this paradox and possible solutions, see de Boer 1985; Viemeister and Wakefield 1991; Tougaard 1998; Pohl et al. 2013).

Transformation of Coding Along the Auditory Pathway

In insects and other arthropods, many neurons can be uniquely identified on the basis of their characteristic morphology. The peripheral stage of a grasshopper’s auditory pathway comprises sensory neurons (afferents), local neurons whose processes are confined to the thoracic ganglia, and ascending neurons, whose axons reach the brain. Present knowledge indicates that this corresponds to a feedforward network (Fig. 4a, Vogel et al. 2005; Vogel and Ronacher 2007).
Fig. 4

Transformation of coding along the auditory pathway. (a) Scheme of a grasshopper’s auditory pathway (AFF auditory afferents, LN local neurons, AN neurons whose axons ascend to the brain). Numbers indicate approximate numbers of neurons at the respective levels. (b) Response (numbers of action potentials at right) of an ascending neuron (AN1) to five presentations of an identical stimulus. (c) Maximal spike rates at different processing levels. (d) Variability of interspike intervals (CV variation coefficient). (e) Variability of spike count (expressed as Fano factor FF). Axis in (ce) AFF afferents, LN local neurons, and AN ascending neurons; numbers indicate sample size ((a) From Ronacher 2013; (be) modified from Vogel et al. 2005, with permission)

Auditory afferents exhibit high firing rates, up to several hundred Hertz (Fig. 4c). Their tonic spike responses represent the amplitude modulation patterns of auditory stimuli by a modulation of the firing rate. The variability of their responses is rather low (see Fig. 4d, e). The precise responses of sensory neurons allow for a good discrimination and classification of auditory stimulus ensembles (Machens et al. 2003; Wohlgemuth and Ronacher 2007).

Along the auditory pathway the maximal spike rates decrease, whereas the spike train variability increases (both spike count and inter-spike-interval variability, Fig. 4ce). In accordance with the larger variability of higher-order neurons, the temporal resolution and the classification success for similar stimuli decrease markedly among the ascending neurons. At the level of afferents and among primary-like local neurons, we find a high classification success that depends almost exclusively on the timing of spikes. In contrast, among ascending neurons the classification success based on a single neuron’s responses decreases, and spike count differences between stimuli become more important (Wohlgemuth and Ronacher 2007). Among ascending neurons, the information appears to be distributed among several neurons and to be represented as a labeled-line population code (Clemens et al. 2011, 2012). A similar reduction in spike rates from ascending to brain neurons is observed within the auditory pathways of crickets (Schildberger 1984; Kostarakos and Hedwig 2012).

Central Processing in the Frequency or Time Domain?

With Fourier analysis, a signal’s temporal structure can be broken down into sine waves, each having a particular amplitude and phase (see, e.g., Yost 2000). If both the resulting spectra for amplitude and phase are known, the original signal can be fully reconstructed. Therefore, there are two principal means of processing a periodic signal: an analysis in the frequency domain, i.e., of the amplitude spectrum without phase and thus without temporal information, and an analysis in the time domain by the computation of temporal parameters such as durations and periods of events. For instance, the processing of the sound carrier by a traveling wave is equivalent to the computation of an amplitude spectrum for a frequency analysis.

The envelopes of the communication signals of many insect species have a highly regular and repetitive structure and consist of a series of stereotyped subunits. Remarkably, several experiments have shown that crickets and grasshoppers accept song signals with randomized or shuffled patterns as conspecific (Fig. 5, Pollack and Hoy 1979; von Helversen and von Helversen 1998; Schmidt et al. 2008). This suggested that central processing may be restricted to the frequency domain and serve to compute an amplitude spectrum of the song envelope. A crucial experiment to determine whether a signal is processed in the frequency or time domain is to present reversed or inverted versions of an attractive signal (see Fig. 5e, f). Inverted variants exhibit the same amplitude spectrum as the original but differ in their phase components and thus temporal qualities. If such modified song models show the same attractiveness as the original, a spectral analysis of the signal envelope is very likely as opposed to temporal processing. However, evidence from experiments as shown in Fig. 5e, f and others revealed large differences in attractiveness which suggests that insects process the amplitude modulations of sound stimuli in the time domain (von Helversen and von Helversen 1998; Schmidt et al. 2008; Hennig 2009).
Fig. 5

Regular song patterns and the importance of temporal order for recognition. (a) Regular calling song of the cricket Teleogryllus oceanicus. (bd) Song models (top) for tests of preference of females and relative frequency of different pulse intervals (bottom). Note the differences in the interval distributions between (b, c, and d). (b) Song model of T. oceanicus. (c) Shuffled song model with the same relative frequency of pulse intervals as in (a). (d) Song model of T. commodus. Song models of (b) and (c) are attractive for females of T. oceanicus, although the temporal order in c is randomized. The song pattern in (d) from the sibling species is not attractive. Although this experiment suggested central processing of the signal envelope in the frequency domain, the bulk of experimental evidence demonstrates processing in the time domain by crickets and other Insects. (e, f) Envelopes of reversed song models for behavioral tests with grasshoppers. The pattern in (e) is very attractive, but the reversed version in (f) is rejected (von Helversen and von Helversen 1998) ((a–d) From Pollack and Hoy 1979 with permission; (e, f) from von Helversen and von Helversen 1998, with permission)

Global Algorithms of Coding

Which global algorithms of coding are implemented in the auditory pathways of insects to achieve the goals of hearing (see Overview and Background)? In the context of acoustic communication as well as predator avoidance, insects have to recognize specific sound signals for which they possess an innate, internal representation. Presently there is no evidence for a maplike neural representation of specific features in the auditory pathways of insects, except for the tonotopic frequency maps in the periphery (Hildebrandt 2014). For predator avoidance ultrasonic cues and the detection of strong onsets as a typical effect of intense bat calls appear to be most important. Burst coding in the auditory pathways of crickets and grasshoppers shows how onsets are detected (see below, Marsat and Pollack 2006; Creutzig et al. 2009; see also Krahe and Gabbiani 2004). For the recognition of conspecific signals, several global schemes were proposed: autocorrelation, cross-correlation with a template, or combinations of high-, low-, and band-pass filters. All these schemes are derived from Fourier transformation and represent mathematical and technical solutions, for which evidence from several investigations exists (Schildberger 1994; Weber and Thorson 1989; Hennig 2003). However, only a few studies have been able to reproduce the preferences exhibited by female insects over a wider range of signals (see Fig. 6ac for response profiles from several species of crickets, bushcrickets, and grasshoppers upon presentation of stereotyped sound patterns composed of subunits built by regular pulses and pauses (von Helversen and von Helversen 1994)). In a recent alternative approach, the recognition of sound signals was examined using linear-nonlinear (LN) models adapted from computational neuroscience (Clemens and Hennig 2013; Clemens and Ronacher 2013). For crickets as well as grasshoppers, Gabor functions emerged as filters that responded best to particular subunit shapes common to the conspecific song signal (combinations of pulse and pause or pairs of pulses, Fig. 6d, e). Gabor functions (a sine wave multiplied with a Gaussian function) are well known from sensory pathways in vertebrates (visual, Daugman 1984; Simoncelli and Olshausen 2001; Priebe and Ferster 2012; auditory, Smith and Lewicki 2006). However, the most notable property of the proposed LN model was the independence of the computation from the exact timing of occurrence of the template within a larger time window. Therefore, this LN-model approach offers an elegant solution to the attractiveness of shuffled and irregular song signals (see Fig. 5 and subchapter: processing in the frequency or time domain). Small modifications of Gabor functions are capable of reproducing response profiles known from several species of insects (Fig. 6ac, Clemens and Hennig 2013). Although these filters by default describe the output of the whole recognition system, present evidence suggests that certain identified neurons in crickets may represent a neuronal correlate of specific LN features (compare Fig. 6eg, Zorović and Hedwig 2011; Kostarakos and Hedwig 2012; Clemens and Hennig 2013).
Fig. 6

Preference profiles of female insects for song signals and LN models. (a–c) Preference profiles for song signals in the time domain by crickets, bush crickets, and grasshoppers (A: oce oceanicus, com commodus; B: cau caudata, can cantans, vir viridissima; C: br brunneus, big biguttulus, mo mollis). (d–e): LN models account for preference functions of the cricket Gryllus bimaculatus. (d) Two LN models (red, green) that predict the preferences for pulse patterns by female crickets (G. bimaculatus): linear filters (left panels) and the respective nonlinearities (central panel) that predict behavioral responses of female crickets (right panel). Note that the duration of the linear filters is only 64 ms. The red dot in the right panel of (d) refers to the pattern in (e). (e) Output of the LN models (d) in response to a song model (upper trace). Response of the linear filters to the song model before (left panels) and after passing through the nonlinearity (right panels). The upper filter (red in d and e) responds like an onset detector; the lower filter resembles a Gabor-function and selectively responds to pairs of pulses (green in d and e). (f–g) The response pattern of an auditory interneuron in the cricket brain resembles the output of the onset detector (red in d, e). The neuronal discharges in (g) illustrate onset responses to sound patterns with different pulse rates ((a–c) From Hennig et al. 2004; (d, e) modified from Clemens and Hennig 2013; and (f, g) from Zorović and Hedwig 2011, with permission)

Local Mechanisms of Coding

For insects there exist several prominent examples of how global algorithms of coding are at least in part implemented by identified neurons. For instance, the detection of bat predators by crickets is mediated by a specific auditory neuron (AN2), whose bursting is crucial for a behavioral response (Fig. 7, Nolen and Hoy 1984; Marsat and Pollack 2006). Similarly, the AN12 in grasshoppers codes for a specific song feature by bursts (Creutzig et al. 2009). Specific combinations of excitation and inhibition account for feature extraction in grasshoppers (gap detection by AN4; Ronacher and Stumpner 1988 – see temporal resolution) and serve as a basis for pulse rate detection by the identified neuron B-LI4 in crickets (Kostarakos and Hedwig 2012). Similarly, resonant properties within the auditory pathway of bushcrickets mediate the detection of specific pulse rates and can be modeled as a property of single neurons (Bush and Schul 2006; Webb et al. 2007). Short time constants of single identified neurons allow them to act as feature detectors for ultrasonic pulses in the auditory pathway of moths (Boyan and Fullard 1988).
Fig. 7

Burst coding by an identified interneuron (AN2) and avoidance behavior in crickets to bat like stimuli. (a) Bursts in AN2 in response to amplitude-modulated stimuli (carrier frequency: 30 kHz); burst spikes are marked as black dots in raster plot. (b) Response of AN2 to a sound stimulus (as in a) and behavioral response (bottom trace: abdominal movements away from the sound source). (c) Amplitude of abdominal movements after isolated action potentials (gray) or bursts (black). Positive values indicate abdomen flexion away from the sound source (From Marsat and Pollack 2006, with permission)

Processing Under the Constraints of Noise: A Twofold Problem

Noise poses unavoidable problems for all sensory systems. There are two classes of noise, external and internal. Various types of external noise influence sound waves on their way from sender to receiver. Hence, as a rule, receivers have to cope with signals that are masked and degraded in their temporal structure (e.g., Michelsen and Larsen 1983; Römer et al. 1989; Römer 2001; Schmidt et al. 2011; see also Brumm and Slabbekoorn 2005; Wiley 2006).

Crickets with their pure tone songs have found a solution to reduce the impact of external noise. The females’ hearing system is tuned to the carrier frequency of male songs and becomes increasingly less sensitive to frequencies farther from the carrier frequency (Fig. 2g, h). The sharpness of the tuning curve may depend on ecological conditions (Schmidt et al. 2011). A cricket species (Paroecanthus podagrosus) living in very noisy tropical rainforests exhibits an exceptionally narrow tuning (Q10dB ~4) that allows for an efficient suppression of ambient noise (Figs. 2h, 8a). Thus, in this species we find a narrow peripheral filter that is perfectly matched to the carrier frequency of conspecific signals and dismisses signals from other species, at the expense of reducing the range of perceivable sounds (Fig. 8a). Many bushcrickets use broadband communication signals which yield a twofold advantage. First, broadband signals are less likely to be degraded in the biotope compared to narrowband signals (Michelsen and Larsen 1983; Römer and Lewald 1992). In addition, the receiver’s nervous system can compare the neuronal signals from differently tuned auditory receptors and thereby reduce intrinsic neuronal noise (see below). Note that the hearing systems of vertebrates (mammals, owls) also use broadband signals to reduce noise or to resolve localization ambiguities (Konishi 1990).
Fig. 8

Processing under the constraints of noise: a twofold problem. (a) Song of a tropical cricket Paroecanthus podagrosus (upper trace). Lower traces. 1: Song envelope. 2: Song envelope under ambient noise levels as recorded in the habitat. 3: Song plus noise less efficiently filtered with the broader tuning curve of a European cricket (Gryllus bimaculatus). 4: Song plus noise filtered with the narrow tuning curve of P. podagrosus, compare Fig. 2h (Adapted from Schmidt et al. 2011). (b) Effects of intrinsic noise and external signal degradation on spike train dissimilarities of three representative neurons, assessed with the van Rossum metric and corrected for spike rate differences. The black arrow at “orig” indicates the average spike train distance found for repeated presentation of the original song pattern, i.e., the result of trial-to-trial variability. The open arrow indicates the additional distance caused by the most strongly degraded signal. AFF sensory neuron, TN1, and SN1 two primary-like local neurons. ((a) Modified from Schmidt et al. 2011. (b) Modified from Neuhofer et al. 2011 and Ronacher 2013, with permission)

In addition to external noise, auditory systems face a second noise problem. Neuronal signals are inherently noisy due to the stochastic opening and closing of ion channels. This intrinsic noise becomes evident as trial-to-trial variability of spike trains in response to repeated presentations of an identical stimulus (see Fig. 4b). Animals with a large nervous system may alleviate this problem by averaging responses from many neurons with similar properties. However, due to size constraints insects probably cannot afford this solution. Under some conditions noise may play a beneficial role and improve neural computations, for example, by stochastic resonance (for review see McDonnell and Ward 2011).

To quantify the trial-to-trial variability or to compare neuronal responses to different stimuli, one can apply a spike train metric (see, e.g., van Rossum 2001). This metric describes the similarity of two spike trains by a single number, which is in line with our intuition of distance: small values indicate high similarity. The metric uses an adjustable parameter to take into account both differences in the spike count as well as in the timing of spikes. This metric approach was used to determine the relative impacts of external and intrinsic noise on the encoding of envelope-degraded stimuli by auditory neurons of grasshoppers (Neuhofer et al. 2011). Unexpectedly, the contribution of external signal degradation to the overall spike train distances was low: even for the highest degradation level, its amount did not exceed that of intrinsic noise (Fig. 8b).

As long as rather different signals have to be analyzed, neuronal noise may not be a very serious problem. However, with communication signals that serve to attract mates, it will be necessary to discriminate between similar signals and to detect small deviations from a species-specific pattern. This is especially the case if quality cues from the sound signals of potential partners are to be extracted. Remarkably enough, the spike trains of even a single auditory afferent allow for an almost perfect discrimination of songs of different males of one species (Machens et al. 2003). At higher stages in the auditory pathway, however, the discrimination deteriorates and intrinsic noise cannot be neglected as a limiting factor. Grasshoppers appear to have circumvented this intrinsic noise problem by changing the coding scheme at a rather peripheral stage of processing (see Fig. 4a): among the ascending neurons information is distributed according to a labeled-line code (Clemens et al. 2011, 2012; see also Clemens and Ronacher 2013).

Directional Hearing

The biophysical qualities of the pressure difference receivers equip the ears of insects with a directional dependence of the vibrational amplitudes of their tympana (Michelsen 1998). Sensory neurons reflect this dependence in spike numbers, and for some species, this difference in response strength is also translated into timing differences (enhanced by ramps Krahe and Ronacher 1993; Ronacher and Krahe 2000). The contrast in response magnitude of sensory neurons for left and right differences is enlarged by contralateral inhibition from local interneurons (ON neurons in crickets, Selverston et al. 1985; bushcrickets, Römer and Krusch 2000; LN in grasshoppers, Marquart 1985). The representation of sound from one side is therefore enhanced, because the responses to sound from the opposite side are suppressed, and this takes place already at early levels of auditory processing of sound direction, usually at the first synapse after sensory receptors. This mechanism of contralateral inhibition in auditory pathways transforms the acoustic environment of insects into acoustic hemispheres. These hemispheres enable a rather accurate distinction of left and right sound signals at the cost of accuracy in determining the angle of the sound source (lateralization in grasshoppers, von Helversen and Rheinlaender 1988). A corollary of the computation of acoustic hemispheres is that sound signals are selectively represented on one side of the insect or the other. This computation resembles the phenomenon of selective attention (Pollack 1988) and allows insects to discriminate sound sources in much the same way as the cocktail party effect well known from humans (Fig. 9, Römer and Krusch 2000). Notably, contralateral inhibition and acoustic hemispheres can exert a considerable influence on mate choice as females of bushcrickets prefer males that take a leader role among singing males, since the representation of the song signal of a follower is suppressed in their auditory pathway (Hartbauer et al. 2005; Siegert et al. 2011). These local computations for enhanced directional responses can also affect the singing behavior of males and promote synchronization and thus chorusing among males (Greenfield and Roizen 1993; Greenfield 1994; Hartbauer et al. 2005). Due to the computational conflict between representation of sound pattern and sound source, grasshoppers split the sensory pathways for processing of cues for pattern and direction by parallel processing (von Helversen 1984; Ronacher et al. 1986). However, in the evolutionarily older communication systems of crickets and bushcrickets, serial processing of pattern and direction appears to prevail (Wendler 1989; Stabel et al. 1989; von Helversen and von Helversen 1995; Schul et al. 1998).
Fig. 9

Acoustic hemispheres and selective representation of sound patterns in bush crickets. (a, b) Left and right specimens of the local interneuron ON1 were recorded simultaneously, while different sound signals (pattern 1 and 2) were presented from either side. (c) Both interneurons selectively represent only the sound pattern from one side (inset: experimental arrangement, correlation coefficient of the spike train with the sound pattern as a measure for copying fidelity) (From Römer and Krusch 2000, with permission)

Conclusions

Auditory processing in insects is constrained by small body size and a relatively small number of neurons. Thus, the hearing capacities are focused on highly relevant tasks, such as predator detection, mate attraction, and, in some cases, eavesdropping on prey or host signals. Insect ears evolved from cuticular mechanosensory proprioreceptors and may respond to different aspects of sound waves – particle velocity or sound pressure; ears can be found in almost any part of the body, and as a consequence, different neural structures are involved in auditory processing. In the auditory pathways, we find similar computational principles and transformations of sensory representations as in vertebrates, however, with an important difference: complex computations are often performed by single neurons or very small populations of identifiable neurons. This size constraint and the focus on a few relevant tasks facilitate experimental approaches.

Notes

Acknowledgment

We want to thank the members of our lab who contributed to several of the cited studies. Special thanks are due to Dr. Michael Reichert who gave helpful advice on the English style and substantially improved the manuscript, as well as to an anonymous reviewer whom we owe many helpful suggestions.

References

  1. Andersson M, Simmons LW (2006) Sexual selection and mate choice. Trends Ecol Evol 21:296–302PubMedGoogle Scholar
  2. Autrum HJ (1942) Schallempfang bei Mensch und Tier. Naturwissenschaften 5:69–85Google Scholar
  3. Barth FG (2002) A spider’s world: senses and behavior. Springer, Berlin/Heidelberg/New YorkGoogle Scholar
  4. Benda J, Hennig RM (2008) Spike-frequency adaptation generates intensity invariance in a primary auditory interneuron. J Comput Neurosci 24:113–136PubMedGoogle Scholar
  5. Benda J, Herz AVM (2003) A universal model for spike-frequency adaptation. Neural Comput 15:2523–2564PubMedGoogle Scholar
  6. Bennet-Clark HC (1998) Size and scale effects as constraints in insect sound communication. Phil Trans R Soc Lond B 353:407–419Google Scholar
  7. Bentley DR, Hoy RR (1972) The genetic control of the neuronal network generating cricket (Teleogryllus gryllus) song pattern. Anim Behav 20:478–492PubMedGoogle Scholar
  8. Bernal XE, Rand AS, Ryan MJ (2006) Acoustic preferences and localization performance of blood-sucking flies (Corethrella Coquillett) to tungara frog calls. Behav Ecol 17:709–715Google Scholar
  9. Boyan GS, Fullard JH (1988) Information processing at a central synapse suggests a noise filter in the auditory pathway of the noctuid moth. J Comp Physiol A 164:251–258PubMedGoogle Scholar
  10. Brown CH (1994) Sound localization. In: Fay RR, Popper AN (eds) Comparative hearing: mammals. Springer, New York/Berlin, pp 57–96Google Scholar
  11. Brumm H, Slabbekoorn H (2005) Acoustic communication in noise. Adv Stud Behav 35:151–209Google Scholar
  12. Bura VL, Rower VG, Martin PR, Yack JE (2011) Whistling in caterpillars (Amorpha juglandis, Bombycoidea): sound-producing mechanism and function. J Exp Biol 214:30–37PubMedGoogle Scholar
  13. Bush SL, Schul J (2006) Pulse-rate recognition in an insect: evidence of a role for oscillatory neurons. J Comp Physiol A 192:113–121Google Scholar
  14. Clemens J, Hennig RM (2013) Computational principles underlying the recognition of acoustic signals in insects. J Comput Neurosci 35:75–85PubMedGoogle Scholar
  15. Clemens J, Ronacher B (2013) Feature extraction and integration underlying perceptual decision making during courtship in grasshoppers. J Neurosci 33:12136–12145PubMedGoogle Scholar
  16. Clemens J, Kutzki O, Ronacher B, Schreiber S, Wohlgemuth S (2011) Efficient transformation of an auditory population code in a small sensory system. Proc Natl Acad Sci USA 108:13812–13817PubMedCentralPubMedGoogle Scholar
  17. Clemens J, Wohlgemuth S, Ronacher B (2012) Nonlinear computations underlying temporal and population sparseness in the auditory system of the grasshopper. J Neurosci 32:10053–10062PubMedGoogle Scholar
  18. Creutzig F, Wohlgemuth S, Stumpner A, Benda J, Ronacher B, Herz AVM (2009) Time-scale invariant representation of acoustic communication signals by a bursting neuron. J Neurosci 29:2575–2580PubMedGoogle Scholar
  19. Daugman JG (1984) Spatial visual channels in the Fourier plane. Vision Res 24:891–910PubMedGoogle Scholar
  20. de Boer E (1985) Auditory time constants: a paradox? In: Michelsen A (ed) Time resolution in auditory systems. Springer, Berlin/Heidelberg, pp 141–157Google Scholar
  21. Faure PA, Mason AC, Yack JE (2009) Invertebrate ears and hearing. In: Binder MD, Hirokawa N, Windhorst U, Hirsch MC (eds) Encyclopedia of neuroscience. Springer, Berlin, pp 2035–2042Google Scholar
  22. Franz A, Ronacher B (2002) Temperature dependence of temporal resolution in an insect nervous system. J Comp Physiol A 188:261–271Google Scholar
  23. Fullard JH, Yack JE (1993) The evolutionary biology of insect hearing. Trends Ecol Evol 8:248–252PubMedGoogle Scholar
  24. Gerhardt HC, Huber F (2002) Acoustic communication in insects and anurans. University of Chicago Press, ChicagoGoogle Scholar
  25. Gollisch T, Herz AVM (2004) Input-driven components of spike-frequency adaptation can be unmasked in vivo. J Neurosci 24:7435–7444PubMedGoogle Scholar
  26. Gollisch T, Herz AVM (2005) Disentangling sub-millisecond processes within an auditory transduction chain. PLoS Biol 3(e8):1–11Google Scholar
  27. Green DM (1985) Temporal factors in psychoacoustics. In: Michelsen A (ed) Time resolution in auditory systems. Springer, Berlin/Heidelberg, pp 122–140Google Scholar
  28. Greenfield MD (1994) Synchronous and alternating choruses in insects and anurans: common mechanisms and diverse functions. Am Zool 34:605–615Google Scholar
  29. Greenfield MD, Roizen I (1993) Katydid synchronous chorusing is an evolutionary stable outcome of female choice. Nature 364:618–620Google Scholar
  30. Grothe B (2000) The evolution of temporal processing in the medial superior olive, an auditory brainstem structure. Prog Neurobiol 61:581–610PubMedGoogle Scholar
  31. Hartbauer M, Kratzer S, Steiner K, Römer H (2005) Mechanisms for synchrony and alternation in song interactions of the bushcricket Mecopoda elongata (Tettigoniidae, Orthoptera). J Comp Physiol A 191:175–188Google Scholar
  32. Hennig RM (2003) Acoustic feature extraction by cross-correlation in crickets? J Comp Physiol A 189:589–598Google Scholar
  33. Hennig RM (2009) Walking in Fourier’s space: algorithms for the computation of periodicities in song patterns by the cricket Gryllus bimaculatus. J Comp Physiol A 195:971–987Google Scholar
  34. Hennig RM, Franz A, Stumpner A (2004) Processing of auditory information in insects. Microsc Res Tech 63:351–374PubMedGoogle Scholar
  35. Hildebrandt KJ (2014) Neural maps in insect versus vertebrate auditory systems. Curr Opin Neurobiol 24:82–87PubMedGoogle Scholar
  36. Hildebrandt KJ, Benda J, Hennig RM (2009) The origin of adaptation in the auditory pathway of locusts is specific to cell type and function. J Neurosci 29:2626–2636PubMedGoogle Scholar
  37. Hildebrandt KJ, Benda J, Hennig RM (2011) Multiple arithmetic operations in a single neuron: the recruitment of adaptation processes in the cricket auditory pathway depends on sensory context. J Neurosci 31:14142–14150PubMedGoogle Scholar
  38. Hoy RR (1989) Startle, categorical response, and attention in acoustic behavior of insects. Ann Rev Neurosci 12:355–375PubMedGoogle Scholar
  39. Hoy RR, Popper AN, Fay RR (eds) (1998) Comparative hearing: insects. Springer, New YorkGoogle Scholar
  40. Huber F (1992) Verhalten und Neurobiologie von stimmbegabten Insekten. Naturwissenschaften 79:393–406Google Scholar
  41. Huber F, Markl H (1983) Neuroethology and behavioural physiology: roots and growing points. Springer, Heidelberg/New YorkGoogle Scholar
  42. Huber F, Moore TE, Loher W (eds) (1989) Cricket behavior and neurobiology. Cornell University Press, IthacaGoogle Scholar
  43. Hummel J, Kössl M, Nowotny M (2011) Sound-induced tympanal membrane motion in bushcrickets and its relationship to sensory output. J Exp Biol 214:3596–3604PubMedGoogle Scholar
  44. Janssen R (1992) Thermal influences on nervous system function. Neurosci Biobehav Rev 16:399–413PubMedGoogle Scholar
  45. Joris PX, Schreiner CE, Rees A (2004) Neural processing of amplitude-modulated sounds. Physiol Rev 84:541–577PubMedGoogle Scholar
  46. Konishi M (1990) Similar algorithms in different sensory systems and animals. Cold Spring Harb Symp Quant Biol 55:575–584PubMedGoogle Scholar
  47. Kostarakos K, Hedwig B (2012) Calling song recognition in female crickets: temporal tuning of identified brain neurons matches behavior. J Neurosci 32:9601–9612Google Scholar
  48. Krahe R, Gabbiani F (2004) Burst firing in sensory systems. Nat Rev Neurosci 5:13–24PubMedGoogle Scholar
  49. Krahe R, Ronacher B (1993) Long rise times of sound pulses in grasshopper songs improve the directionality cues received by the CNS from auditory receptors. J Comp Physiol A 173:425–434Google Scholar
  50. Lakes-Harlan R, Stölting H, Stumpner A (1999) Convergent evolution of insect hearing organs from a preadaptive structure. Proc R Soc Lond B 266:1161–1167Google Scholar
  51. Lehmann GUC, Strauß J, Lakes-Harlan R (2007) Listening when there is no sexual signalling? Maintenance of hearing in the asexual bushcricket Poecilimon intermedius. J Comp Physiol A 193:537–545Google Scholar
  52. Machens CK, Stemmler MB, Prinz P, Krahe R, Ronacher B, Herz AVM (2001) Representation of acoustic communication signals by insect auditory receptor neurons. J Neurosci 21:3215–3227PubMedGoogle Scholar
  53. Machens CK, Schütze H, Franz A, Stemmler MB, Ronacher B, Herz AVM (2003) Auditory receptor neurons preserve characteristic differences between conspecific communication signals. Nat Neurosci 6:341–342PubMedGoogle Scholar
  54. Marquart V (1985) Local interneurons mediating excitation and inhibition onto ascending neurons in the auditory pathway of grasshoppers. Naturwissenschaften 72:42–44Google Scholar
  55. Marsat G, Pollack GS (2006) A behavioural role for feature detection by sensory bursts. J Neurosci 26:10542–10547PubMedGoogle Scholar
  56. McDonnell MD, Ward LM (2011) The benefits of noise in neural systems: bridging theory and experiment. Nat Rev Neurosci 12:415–425PubMedGoogle Scholar
  57. Meier T, Reichert H (1990) Embryonic development and evolutionary origin of the Orthopteran auditory organs. J Neurobiol 21:592–610PubMedGoogle Scholar
  58. Michelsen A (1979) Insect ears as mechanical systems. Am Sci 67:696–706Google Scholar
  59. Michelsen A (ed) (1985) Time resolution in auditory systems. Springer, Berlin/HeidelbergGoogle Scholar
  60. Michelsen A (1998) Biophysics of sound localization in insects. In: Hoy RR, Popper AN, Fay RR (eds) Comparative hearing: insects. Springer, Berlin/New York, pp 18–62Google Scholar
  61. Michelsen A, Larsen ON (1983) Strategies for acoustic communication in complex environments. In: Huber F, Markl H (eds) Neuroethology and behavioural physiology. Springer, Berlin/Heidelberg, pp 321–331Google Scholar
  62. Michelsen A, Popov A, Lewis B (1994) Physics of directional hearing in the cricket Gryllus bimaculatus. J Comp Physiol A 175:153–164Google Scholar
  63. Miller LA, Surlykke A (2001) How some insects detect and avoid being eaten by bats: tactics and countertactics of prey and predator. BioScience 51:571–582Google Scholar
  64. Moiseff A, Pollack G, Hoy R (1978) Steering responses of flying crickets to sound and ultrasound: mate attraction and predator avoidance. Proc Natl Acad Sci USA 75:4052–4056PubMedCentralPubMedGoogle Scholar
  65. Montealegre-Z F, Jonsson T, Robson-Brown KA, Postles M, Robert D (2012) Convergent evolution between insect and mammalian audition. Science 338:968–971PubMedGoogle Scholar
  66. Nadrowski B, Effertz T, Senthilan PR, Göpfert MC (2011) Antennal hearing in insects – new findings, new questions. Hearing Res 273:7–13Google Scholar
  67. Neuhofer D, Stemmler M, Ronacher B (2011) Neuronal precision and the limits for acoustic signal recognition in a small neuronal network. J Comp Physiol A 197:251–265Google Scholar
  68. Nolen TG, Hoy RR (1984) Initiation of behavior by single neurons: the role of behavioral context. Science 226:992–994PubMedGoogle Scholar
  69. Penzlin H (2005) Lehrbuch der Tierpyhsiologie. Elsevier, MünchenGoogle Scholar
  70. Pohl NU, Slabbekoorn H, Neubauer H, Heil P, Klump GM, Langemann U (2013) Why longer song elements are easier to detect: threshold level-duration functions in the Great Tit and comparison with human data. J Comp Physiol A 199:239–252Google Scholar
  71. Pollack GS (1988) Selective attention in an insect auditory neuron. J Neurosci 8:2635–2639PubMedGoogle Scholar
  72. Pollack GS, Hoy RR (1979) Temporal pattern as a cue for species-specific calling song recognition in crickets. Science 204:429–432PubMedGoogle Scholar
  73. Priebe NJ, Ferster D (2012) Mechanisms of neuronal computation in mammalian visual cortex. Neuron 75:194–208PubMedCentralPubMedGoogle Scholar
  74. Prinz P, Ronacher B (2002) Temporal modulation transfer functions in auditory receptor fibres of the locust (Locusta migratoria L.). J Comp Physiol A 188:577–587Google Scholar
  75. Riede K (1987) A comparative study of mating behaviour in some neotropical grasshoppers (Acridoidea). Ethology 76:265–296Google Scholar
  76. Riede K, Kämper G, Höfler I (1990) Tympana, auditory thresholds, and projection areas of tympanal nerves in singing and silent grasshoppers (Insects, Acridoidea). Zoomorphology 109:223–230Google Scholar
  77. Rieke F, Warland D, de Ruyter van Steveninck R, Bialek W (1997) Spikes – exploring the neural code. MIT Press, Cambridge, MAGoogle Scholar
  78. Robert D, Göpfert MC (2002) Novel schemes for hearing and orientation in insects. Curr Opin Neurobiol 12:715–720PubMedGoogle Scholar
  79. Robert D, Hoy RR (1998) The evolutionary innovation of tympanal hearing in Diptera. In: Hoy RR, Popper AN, Fay RR (eds) Comparative hearing: insects. Springer, New York, pp 197–227Google Scholar
  80. Robert D, Miles RN, Hoy RR (1996) Directional hearing by mechanical coupling in the parasitoid fly Ormia ochracea. J Comp Physiol A 179:29–44PubMedGoogle Scholar
  81. Robertson RM, Money TG (2012) Temperature and neuronal circuit function: compensation, tuning and tolerance. Curr Opin Neurobiol 22:724–734PubMedGoogle Scholar
  82. Römer H (1976) Die Informationsverarbeitung tympanaler Rezeptorelemente von Locusta migratoria. J Comp Physiol A 109:101–122Google Scholar
  83. Römer H (1983) Tonotopic organization of the auditory neuropile in the bushcricket Tettigonia viridissima. Nature 306:60–62Google Scholar
  84. Römer H (2001) Ecological constraints for sound communication: from grasshoppers to elephants. In: Barth FG, Schmid A (eds) Ecology of sensing. Springer, Berlin/Heidelberg/New York, pp 59–77Google Scholar
  85. Römer H, Krusch M (2000) A gain-control mechanism for processing of chorus sounds in the afferent auditory pathway of the bushcricket Tettigonia viridissima (Orthoptera, Tettigoniidae). J Comp Physiol A 186:181–191PubMedGoogle Scholar
  86. Römer H, Lewald J (1992) High-frequency sound transmission in natural habitats: implications for the evolution of insect acoustic communication. Behav Ecol Sociobiol 29:437–444Google Scholar
  87. Römer H, Marquart V, Hardt M (1988) Organization of a sensory neuropile in the auditory pathway of two groups of Orthoptera. J Comp Neurol 275:201–215PubMedGoogle Scholar
  88. Römer H, Bailey WJ, Dadour I (1989) Insect hearing in the field: III masking by noise. J Comp Physiol A 164:609–620Google Scholar
  89. Römer H, Spickermann M, Bailey W (1998) Sensory basis for sound intensity discrimination in the bushcricket Requena verticalis (Tettigoniidae, Orthoptera). J Comp Physiol A 182:595–607Google Scholar
  90. Ronacher B (2013) Processing of species-specific signals in the auditory pathway of grasshoppers. In: Hedwig B (ed) Insect hearing and acoustic communication. Springer, Berlin, Heidelberg, pp. 185–204Google Scholar
  91. Ronacher B, Krahe R (2000) Temporal integration vs. parallel processing: coping with the variability of neuronal messages in directional hearing of insects. Eur J Neurosci 12:2147–2156PubMedGoogle Scholar
  92. Ronacher B, Stumpner A (1988) Filtering of behaviourally relevant temporal parameters of a grasshopper´s song by an auditory interneuron. J Comp Physiol A 163:517–523Google Scholar
  93. Ronacher B, von Helversen D, von Helversen O (1986) Routes and stations in the processing of auditory directional information in the CNS of a grasshopper, as revealed by surgical experiments. J Comp Physiol A 158:363–374Google Scholar
  94. Ronacher B, Franz A, Wohlgemuth S, Hennig H (2004) Variability of spike trains and the processing of temporal patterns of acoustic signals–problems, constraints, and solutions. J Comp Physiol A 190:257–277Google Scholar
  95. Schildberger K (1984) Temporal selectivity of identified auditory neurons in the cricket brain. J Comp Physiol A 155:171–185Google Scholar
  96. Schildberger K (1994) The auditory pathway of crickets: adaptations for intraspecific acoustic communication. In: Schildberger K, Elsner N (eds) Neural basis of behavioural adaptations. G Fischer, Stuttgart, pp 209–225Google Scholar
  97. Schildberger K, Elsner N (1994) Neural basis of behavioural adaptations. G. Fischer, StuttgartGoogle Scholar
  98. Schmidt A, Ronacher B, Hennig RM (2008) The role of frequency, phase and time for processing amplitude modulated signals by grasshoppers. J Comp Physiol A 194:221–233Google Scholar
  99. Schmidt AKD, Riede K, Römer H (2011) High background noise shapes selective auditory filters in a tropical cricket. J Exp Biol 214:1754–1762PubMedPubMedCentralGoogle Scholar
  100. Schneider E, Hennig RM (2012) Temporal resolution for calling song signals by female crickets, Gryllus bimaculatus. J Comp Physiol A 198:181–191Google Scholar
  101. Schul J, Sheridan RA (2006) Auditory stream segregation in an insect. Neuroscience 138:1–4PubMedGoogle Scholar
  102. Schul J, von Helversen D, Weber T (1998) Selective phonotaxis in Tettigonia cantans and T. viridissima in song recognition and discrimination. J Comp Physiol A182:687–694Google Scholar
  103. Selverston A, Kleindienst H-U, Huber F (1985) Synaptic connectivity between cricket auditory interneurons as studied by selective photoinactivation. J Neurosci 5:1283–1292PubMedGoogle Scholar
  104. Senthilan PR, Piepenbrock D, Ovezmyradov G, Nadrowski B, Bechstedt S, Pauls S, Winkler M, Möbius W, Howard J, Göpfert MC (2012) Drosophila auditory organ genes and genetic hearing defects. Cell 150:1042–1054PubMedGoogle Scholar
  105. Siegert ME, Römer H, Hashim R, Hartbauer M (2011) Neuronal correlates of a preference for leading signals in the synchronizing bushcricket Mecopoda elongata (Orthoptera: Tettigoniidae). J Exp Biol 214:3924–3934PubMedCentralPubMedGoogle Scholar
  106. Simoncelli E, Olshausen B (2001) Natural image statistics and neural representation. Annu Rev Neurosci 24:1193–1216PubMedGoogle Scholar
  107. Smith EC, Lewicki MS (2006) Efficient auditory coding. Nature 439:978–982PubMedGoogle Scholar
  108. Stabel J, Wendler G, Scharstein H (1989) Cricket phonotaxis: localization depends on recognition of the calling song pattern. J Comp Physiol A 165:165–177Google Scholar
  109. Stölting H, Stumpner A (1998) Tonotopic organization of auditory receptors of the bushcricket Pholidoptera griseoaptera (Tettigoniidae, Decticinae). Cell Tissue Res 294:377–386PubMedGoogle Scholar
  110. Stumpner A (1996) Tonotopic organization of the hearing organ in a bushcricket. Naturwissenschaften 83:81–84Google Scholar
  111. Stumpner A (1997) An auditory interneurone tuned to the male song frequency in the duetting bushcricket Ancistrura nigrovittata (Orthoptera, Phaneropteridae). J Exp Biol 200:1089–1101PubMedGoogle Scholar
  112. Stumpner A, Lakes-Harlan R (1996) Auditory interneurons in a hearing fly (Therobia leonidei, Ormiini, Tachinidae, Diptera). J Comp Physiol A 178:227–233Google Scholar
  113. Stumpner A, von Helversen D (2001) Evolution and function of auditory systems in insects. Naturwissenschaften 88:159–170PubMedGoogle Scholar
  114. Suga N, Zhang Y, Yan J (1997) Sharpening of frequency tuning by inhibition in the thalamic auditory nucleus of the mustached bat. J Neurophysiol 77:2098–2114PubMedGoogle Scholar
  115. Tougaard J (1998) Detection of short pure-tone stimuli in the noctuid ear: what are temporal integration and integration time all about? J Comp Physiol A 183:563–572Google Scholar
  116. van Rossum MCW (2001) A novel spike distance. Neural Comput 13:751–763PubMedGoogle Scholar
  117. van Staaden MJ, Römer H (1998) Evolutionary transition from stretch to hearing organs in ancient grasshoppers. Nature 394:773–776Google Scholar
  118. Viemeister NF, Plack CJ (1993) Time analysis. In: Yost WA, Popper AN, Fay RR (eds) Human psychophysics. Springer, Berlin/Heidelberg/New York, pp 116–154Google Scholar
  119. Viemeister NF, Wakefield GH (1991) Temporal integration and multiple looks. J Acoust Soc Am 90:858–865PubMedGoogle Scholar
  120. Vogel A, Ronacher B (2007) Neural correlations increase between consecutive processing levels in the auditory system of locusts. J Neurophysiol 97:3376–3385PubMedGoogle Scholar
  121. Vogel A, Hennig RM, Ronacher B (2005) Increase of neuronal response variability at higher processing levels as revealed by simultaneous recordings. J Neurophysiol 93:3548–3559PubMedGoogle Scholar
  122. von Helversen D (1972) Gesang des Männchens und Lautschema des Weibchens bei der Feldheuschrecke Chorthippus biguttulus (Orthoptera, Acrididae). J Comp Physiol 81:381–422Google Scholar
  123. von Helversen D (1984) Parallel processing in auditory pattern recognition and directional analysis by the grasshopper Chorthippus biguttulus L (Acrididae). J Comp Physiol A 154:837–846Google Scholar
  124. von Helversen D (1997) Acoustic communication and orientation in grasshoppers. In: Lehrer M (ed) Orientation and communication in arthropods. Birkhäuser, Basel, pp 301–341Google Scholar
  125. von Helversen D, Rheinlaender (1988) Interaural intensity and time discrimination in an unrestrained grasshopper: a tentative behavioural approach. J Comp Physiol A 162:333–340Google Scholar
  126. von Helversen D, von Helversen O (1975a) Verhaltensgenetische Untersuchungen am akustischen Kommunikationssystem der Feldheuschrecken (Orthoptera, Acrididae). I. Der Gesang von Artbastarden zwischen Chorthippus biguttulus und C. mollis. J Comp Physiol 104:273–299Google Scholar
  127. von Helversen D, von Helversen O (1975b) Verhaltensgenetische Untersuchungen am akustischen Kommunikationssystem der Feldheuschrecken (Orthoptera, Acrididae). II. Das Lautschema von Artbastarden zwischen Chorthippus biguttulus und C. mollis. J Comp Physiol 104:301–323Google Scholar
  128. von Helversen D, von Helversen O (1995) Acoustic pattern recognition and orientation in orthopteran insects: parallel or serial processing. J Comp Physiol A 177:767–774Google Scholar
  129. von Helversen D, von Helversen O (1997) Recognition of sex in the acoustic communication of the grasshopper Chorthippus biguttulus (Orthoptera, Acrididae). J Comp Physiol A 180:373–386Google Scholar
  130. von Helversen D, von Helversen O (1998) Acoustic pattern recognition in a grasshopper: processing in the frequency or time domain? Biol Cybern 79:467–476Google Scholar
  131. von Helversen O, von Helversen D (1987) Innate receiver mechanisms in the acoustic communication of orthopteran insects. In: Guthrie DM (ed) Aims and methods in neuroethology. Manchester Univ Press, Manchester, pp 104–150Google Scholar
  132. von Helversen O, von Helversen D (1994) Forces driving coevolution of song and song recognition in grasshoppers. In: Schildberger K, Elsner N (eds) Neural basis of behavioural adaptations. G. Fischer, Stuttgart, pp 253–284Google Scholar
  133. Webb B, Wessnitzer J, Bush SL, Schul J, Buchli J, Ijspeert A (2007) Resonant neurons and bushcricket behaviour. J Comp Physiol A 193:285–288Google Scholar
  134. Weber T, Thorson J (1989) Phonotactic behavior of walking crickets. In: Huber F, Moore TE, Loher W (eds) Cricket behavior and neurobiology. Cornell University Press, Ithaca, pp 310–339Google Scholar
  135. Wendler G (1989) Acoustic orientation in crickets in the presence of two sound sources. Naturwissenschaften 76:128–129Google Scholar
  136. Wiley RH (2006) Signal detection and animal communication. Adv Study Behav 36:217–247Google Scholar
  137. Windmill JFG, Göpfert MC, Robert D (2005) Tympanal travelling waves in migratory locusts. J Exp Biol 208:157–168PubMedGoogle Scholar
  138. Wohlgemuth S, Ronacher B (2007) Auditory discrimination of amplitude modulations based on metric distances of spike trains. J Neurophysiol 97:3082–3092PubMedGoogle Scholar
  139. Wohlgemuth S, Vogel A, Ronacher B (2011) Encoding of amplitude modulations by auditory neurons of the locust: influence of modulation frequency, rise time, and modulation depth. J Comp Physiol A 197:61–74Google Scholar
  140. Yost WA (2000) Fundamentals of hearing – an introduction. Academic, San Diego/New YorkGoogle Scholar
  141. Zorović M, Hedwig B (2011) Processing of species-specific auditory patterns in the cricket brain by ascending, local, and descending neurons during standing and walking. J Neurophysiol 105:2181–2194PubMedCentralPubMedGoogle Scholar

Copyright information

© Springer Science+Business Media New York 2013

Authors and Affiliations

  1. 1.Department of BiologyHumboldt-Universität zu BerlinBerlinGermany