Category:Evolution of Communication Sounds
From CNBH Acoustic Scale Wiki
|
The size-shift-covariant auditory image (sscAI), and the description of how it might be constructed, have been presented largely from the perspective of human speech perception, and what makes it robust to changes in acoustic scale. We attempted to motivate the need for shift covariance by explaining that the communication sounds of a population of animals must vary in acoustic scale because the sound generators are integral parts of the body, and that perceptual robustness requires the appropriate processing of scale information. Nevertheless, the proposed perceptual solution – a scale shift covariant auditory image – might seem rather fantastic to some readers, and they might wonder how something so unusual could possibly have developed. In point of fact, from the information processing point of view, the evolution of pulse-resonance communication would appear to be relatively straightforward, and the prevalence of pulse-resonance communication in the natural world suggests that it probably evolved separately, a number of times, in widely divergent groups of animals.
Although it is not necessary to explain the evolution of hearing in order to study auditory perception, it might be reassuring to know that it is quite simple to implement size-shift covariance in neural processing. Accordingly, the category Evolution of Communication Sounds is intended to provide a brief overview of how acoustic communication might have evolved. The text is currently under construction.
Moreover, it is an interesting topic in its own right, and it might be expected to provide a useful framework for understanding the relative importance of perceptual phenomena and the relative importance of the structures used in processing the sounds; specifically, the wavelet nature of cochlear filtering and the fact that hair cells phase lock to BMM. Both of these aspects of auditory processing are readily observable but their importance is largely ignored in spectrographic models of acoustic processing. It is assumed that if airplanes can fly on inflexible metal wings then the processing of sound can be satisfactorily achieved from a summary time-frequency representation like the spectrogram.
We also think it is much more difficult to imagine how communication based on a spectrographic representation of sound could have evolved, given that the resonators that produce the sounds have to grow with the animal. The spectrogram destroys a significant part of the scale information and it does nothing to assist with the processing of the scale information that it does preserve. It leaves the problem to more central mechanisms to solve. (???Littob???)
The following is a hypothetical list of the main evolutionary developments in acoustic bio-communication from the perspective of information processing; that is, what kind of signals are produced and how are they detected and analysed to provide the listener with information about acoustic sources, both animate and inanimate.
The hair cell and the information recorded by primary nerve fibers
The transduction mechanism in vertebrate hearing is a highly specialized form of hair cell; Fig. 3.1 shows a row or outer hair cells from a ???. The ‘hairs’ are actually stiff cilia and each cell has a w-shaped, or v-shaped wedge of these cilia attached to one large kinocilium at the vertex of the wedge (Do the w-shaped ones have two kinocilia?). The wedges of adjacent hair cells are aligned in the same direction. Within a hair bundle, the length of the cilium decreases progressively with the distance from the vertex. There is a link from the tip of each cilium to a point just below the tip on the adjacent cilium as they proceed up towards the kinocilium. The tip links are single molecules that stretch.
The structures that support and cover hair cells are designed to direct the acoustic pressure of a sound wave onto the surface of the wedge, which causes the bundle of cilia to bend as a unit in the direction of the kinocilium. This stretches the surface of the bundle and opens ion channels in the cilia where the tip links attach to the next cilium. This allows calcium in the surrounding fluid to flow into the cell in proportion to the pressure on the surface of the wedge. This depolarizes the hair cell, causing it to release neural transmitter into the synapse at the base of the cell, again in proportion to the pressure on the wedge.
The structure of the hair bundle means that the hair cell is a uni-directional device, and the information that it produces about acoustic pressure is mono-polar. The transmitter is produced in response to rising pressure on the front of each cycle of the wave. The transmitter accumulates until the pressure peaks; then, when the wave reverses direction, the tip links relax, the ion channels close, and the flow of transmitter into the synapse ends abruptly. It does not start to flow again until the rising portion of the next cycle passes the neutral point and begins to stretch the tip links. There is also a fast-acting compressive mechanism inside the cilia of the hair bundle; the flow of calcium ions into the cilia causes stiffening of the cilium bundle. As a result, the resistance of the bundle increases with acoustic pressure and ever more pressure is required to push the wedge further. Thus, the geometry and structure of the hair cell and its surrounding structures mean that the hair cell produces a burst of transmitter that accumulates to a maximum just as the wave reaches its positive pressure peak, and the compression mechanism concentrates the transmitter release in time.
Hair cells are not nerve cells; they do not fire. The neural impulses are produced by primary, auditory nerve fibers whose butons cluster around the base of the hair cell. They fire when the volume of transmitter collected from their butons exceeds a threshold level. This means that the combination of hair-cell and primary nerve fibre produce neural impulses that record the times of pressure peaks, firing typically just before the peak pressure is reached. This is referred to as neural phase locking; the neural impulses are locked to one phase of the acoustic pressure wave. Phase locking is one of the prime characteristics of vertebrate hearing. The accuracy of phase locking decreases as the frequency of the acoustic wave increases, but in many animals phase locking persists to frequencies in excess of 1000 Hz where the cycle time is less than 1 millisecond!
So, what kind information does the hair cell with its primary fibers extract from communication sounds like the /e/ vowels in Fig. 1? The description above indicates that a population of hair cells connected to primary auditory nerve fibres would respond to these sounds with a stream of nerve impulses that are phase locked to the positive pressure peaks in the wave, and the number of fibres that fire in response to each peak would be a compressive function of the pressure at the peak. A simulation of the aggregate firing rate, or the probability of firing, as a function of time would resemble a half-wave rectified version of the waves with a little smoothing of the fine timing details by the loss of phase locking at high frequencies. The important point to note is that the system appears to be designed to preserve information about the timing of the peaks in the wave. It is not designed to extract the envelope of the wave, or the energy of the wave. If the desired information were the envelope, it would be better to apply full-wave rectification and temporal smoothing, neither of which are evident in the cochlea or early stages of neural processing. If the desired information were the energy, there should be evidence that the cochlea is somehow squaring the amplitude of the wave, which there is not. Squaring is an expansive transform; in contrast, the cochlea applies compression to the amplitude of the wave. Together these observations suggest that the auditory system goes to considerable trouble to preserve the fine timing information of the sound, and that we might expect to find that the time-interval information plays an important role in auditory perception.
[Include a picture of the neat rows of cilia wedges in the mammalian cochlea. Argue that the degree of organisation is an indication of the importance of phase locking and that it should not be ignored. Point out that if the auditory system wanted the best representation of the envelope from an oscillating BM, it would have hair cells with the opposite orientation which would capture the times of the negative peaks, and fill in the gaps in the envelope at exactly the most important instant. And point out the Yost et al., 1998 show that it is only half wave rectification that can predict the PSR and Irn(-G).]
Scale information in bursts of neural impulses
The important information in the stream of neural impulses flowing from the cochlea is contained in the pattern of time-intervals rather than the individual, absolute firing times.
Amphibians and fish do not have a cochlea or any means of performing a mechanical frequency analysis, but they do have hair cells much like those of the amniotes (animals whose eggs have an amniotic membrane including reptiles, birds and mammals), and fish and frogs do communicate using pulse-resonance sounds. It suggests that the reception of sound with hair cells and the production of pulse-resonance sounds evolved in tandem to support bio-communication long before the development of mechanical frequency selectivity by the amniotes.
Evolutionary heuristic – an auditory fairy tale in which we string together the things you need to know in a seemingly logical temporal sequence that is memorable and emphasized what are the important components of communication and how they work together.
This category currently contains no pages or media.
