Riede, Tobias : Vocal changes in animals during disorders


Chapter 1. Introduction

Acoustic communication plays an important role in most animal taxa (e.g. reviews: mammals: Tembrock 1996; birds: Kroodsma, D.E., Miller, E.H. 1996; anura: Gerhardt 1991; insects: Bailey 1991). Communication involves two individuals, a sender and a receiver. The sender produces a signal which conveys information. This signal is transmitted through the environment and is detected by the receiver. The receiver uses the information to help make a decision about how it should respond. The receiver's response affects the fitness of the sender as well as its own (Hauser 1996; Bradbury, Vehrencamp 1998).

Playback experiments have proven that acoustic signals can deliver useful information about the sender to the recipient, such as identifying the sender (e.g. Hammerschmidt, Todt 1995), or passing on information about different types of predators seen by the sender (Cheney, Seyfarth 1990; Fischer 1996).

It has long been of interest if information about the sender, like emotional state or hormonal state, can be evaluated by studying behavioral parameters (review e.g. in Stamp-Dawkins, 1982; Tembrock 1990; Puppe 1996). The acoustic approach, i.e. using the acoustic utterance of the sender looking for correlates of these internal parameters, has been repeatedly documented (e.g. signaling stress in pigs: Schrader 1996; signaling hunger in piglets: Weary et al. 1997; signaling pain in piglets: White et al. 1995; signaling need for warmth in pelican chicks: Evans 1994). It however remains unclear whether vocal changes occur due to disorders in the sender. If the sender's physiology or merely the sound generating apparatus is affected by a disease, what impact on voice does it have? How can this vocal change be described? Those questions were the central issue in this work, consequently this work is focussed on the sender's side - the acoustic signal and the mechanism of sound production.

One can imagine several contexts in which the information about a sender's state of health might be important for the receiver. Attracting and stimulating a mate is a typical context in which the sender's body condition or state of health are relevant for further behavioral processes. Choosy females are predisposed to work as 'veterinarians' and the voice might than be a good indicator for evaluation if other communicative channels like the visual one are lacking. In birds such relationships have already been found out. There is a significant negative correlation between the parasitic load of a singing bird and its vocal repertoire size (Møller 1991; Buchanan et al. 1999) and a positive relationship between the male's vocal


repertoire size and its reproductive success is known for a long time (e.g. Catchpole 1980).

Vocal changes can be caused by several reasons. It can be assigned to an ontogenetic change as have already been shown for instance in the wolf (Frommolt et al. 1988), in the rhesus macaque (Hammerschmidt et al. in press), and in bushbabies (Zimmermann 1995). Apart from physiological vocal changes, alterations may also be expected in diseased conditions, because the larynx may be involved (primarily or secondarily) in the pathological process. Vocal changes have already been mentioned for domestic animals as symptom during several diseases (e.g. Bagley et al. 1993 and see chapter 2.3 of this study).

Under experimental conditions vocal changes have been provoked by destruction of the nerval supply of the larynx (Jürgens et al. 1978) or by brain lesions (Ploog 1988, Jürgens 1995), in squirrel monkeys (Saimiri sciureus). In birds several experimental studies have been accomplished to study the sound generating mechanisms by experimentally affecting the different anatomical structures (for instance destroying the tympanal membranes of the syrinx) which were assumed to be involved in that process and looking what effect those affections have on the vocal product (e.g. Suthers, Goller, 1997; Goller, Larson, 1997).

However, the objective of such experimental studies has largely been the evaluation of the mechanism of sound production in animals rather than simulating natural disease states which may affect the animals. Studies of acoustic alterations in diseased animals are inconclusive as vocal changes were subjectively evaluated by the unaided ear and lacked a detailed sound analysis.

The main goal of this work was the quantification of vocal changes during disorders. For that purpose primarily signal analysis was applied. Additionally, post-mortem investigation of the larynx delivered insights into probable generation mechanisms of the vocal change in one of the case studies. As we were mainly studying canine vocalization, anatomical measurement of the vocal tract delivered insights into the vocal tract's role in the acoustic of the domestic dog vocalization.

Besides an introduction into the basic anatomy of the larynx and the physiology of sound production, Chapter 2 provides an overview of disorders in animals associated with vocal changes. Additionally, techniques of signal analysis used in this study are explained.

In Chapter 3, three case studies of vocal changes are presented, a Japanese macaque (Macaca fuscata) infant with a metabolic disease, a domestic cat (Felis catus) infant with


craniocerebellar trauma and an adult dog-wolf-hybrid female with peculiarities in the larynx anatomy. All three cases have in common that the vocal change was first recognized by the unaided ear and subsequently confirmed with signal analysis techniques.

The case study is a commonly used method in the medical sciences for the presentation of special and rarely occuring (clinical) cases. Those cases can usually not be repeated under acceptable (and comparable) conditions with a statistically sufficient high number of subjects. However, case studies can help to understand theoretically derived hypotheses and in that way, they help to come to generalizing conclusions.

It is hypothesized, that the amount of nonlinear phenomena in the vocal repertoire increases during voice disorders. This hypothesis is based on the assumption that the elastic tissue of the vocal folds can be considered as a system of coupled oscillators. During harmonic oscillation, all oscillation modes of the vocal folds are synchronized. Under certain conditions, voice instabilities should be observed. In particular, vocal fold lesions, paralysis, and other pathological conditions may induce subharmonic vocalization, biphonation, and deterministic chaos, which are considered as nonlinear phenomena.

These relationships are well known for the human voice. In normal human phonation, there is a certain amount of nonlinear phenomena occuring in the acoustic utterances (for instance in newborn cries: Sirviö and Michelsson, 1976; Mende et al. 1990; in non-cry vocalization of infants: Robb and Saxman, 1988; in normal conversational speech: Dolansky and Tjernlund, 1968; Kohler 1996). The amount of nonlinear phenomena increases under pathological conditions (Herzel, Wendler, 1991; Herzel et al. 1994). Further in non-human mammals nonlinear phenomena seem to be normal to a certain extent in normal vocalization (Wilden et al. 1998). Three case studies in the present work show that the amount can increase in disordered animals.

Chapter 4 presents data of the application of the harmonic-to-noise-ratio (HNR) to dog barks. In the case studies from Chapter 3, only harmonic vocalization was considered. Counting nonlinear phenomena is a useful method to define a vocal change when dealing with harmonic vocalization. However, in dog barks, subharmonic and chaotic oscillation modes occur originally. The spectrogram shows harmonic energy and noise energy to various extent in the bark. Vocal changes of the bark, for instance hoarseness, seem to be founded on a shifting of the ratio of the energy between harmonic and noisy elements. Additionally, other authors suggest different communicative relevance according to this energy ratio (Tembrock 1976; Feddersen-Petersen 1996, Wilden 1997).

Thus, a moving average procedure for calculation of the harmonics-to-noise-ratio (HNR) was applied and tested. First, synthetic sounds with defined HNR confirmed the applicability of the


procedure. Second, human subjects evaluated dog barks as predicted by the HNR measure. Third, using the ranking of the animals according their HNR values, it was possible to reproduce the HNR ranking applying multivariate statistics to a parameter set measured on the barks of the same dogs. The results suggest that dysphonia in dog barks can be quantified applying the HNR.

In Chapter 5, the role of the vocal tract in the acoustics of dog vocalization was investigated. Domestication and selective breeding has resulted in a high variability in head size and shape in the dog (Canis familiaris). This suggests that there might be large differences in the vocal tract length which could result in a formant behaviour influencing the interbreed communication. Modern investigations of the animal's vocal tract look at the relationship between resonance characteristic and body size parameters. We took this question as a starting point for a vocal tract investigation to see if any, and to what extent did domestication affect this relationship.

Generally, describing vocal changes in animals during diseases was the main goal of this thesis. Since the description of a pathological state is best done by understanding the physiological mechanisms, the major results of this thesis were obtained by applying computer modelling approaches of sound production to experimental ('true') data.

[Front page] [Acknowledgements] [Attached Audiofiles] [1] [2] [3] [4] [5] [6] [7] [Bibliography] [Declaration]

© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.

DiML DTD Version 2.0
Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML - Version erstellt am:
Thu Sep 28 14:01:51 2000