example, in the case of repertoire–size preferences, it has been suggested
that the acoustic entropy in the signal may drive neuronal habituation
mechanisms (see Ryan 1998; Gentner et al. 2001) as opposed to more direct
measures of repertoire sizes employed by researchers, such as simple song
counts.
1.3. Technical Challenges
Given the attractiveness of vocal communication systems as research
objects, it is somewhat surprising that the underlying neural mechanisms
are relatively poorly understood. More than anything else, this lack of
understanding reflects the tremendous technical challenges involved in
studying these complex biological systems. Assessing perceptual responses
to conspecific stimuli often requires elaborate conditions difficult to stage
in the laboratory, and ethologists have only cautiously embraced conditioning procedures otherwise common in the study of perceptual phenomena (e.g., Dooling and Searcy 1980; Weary and Krebs 1992; Adret 1993). In
part, such caution has been in an effort to avoid potential confusions that
can arise when interpreting arbitrary conditioning paradigms in the context
of natural behavior.
A related technical limitation is that the proper analysis of vocal communication will often require one or more animals to be actively engaged
in communication as the experimenter records physiological activity.
Ideally, this should include both field and laboratory settings. For the most
part, this vision is more fanciful than real because the ability to reliably
conduct single-neuron recordings in awake, behaving vertebrate animals
of the small size commonly employed in neuroethological studies is
only beginning to emerge (Dave et al. 1998b; Nieder and Klump 1999;
Venkatachalam et al. 1999), so far mainly under relatively constrained laboratory conditions (Yu and Margoliash 1996; Dave et al. 1998a; Nieder and
Klump 1999).
Yet another reason for the relative paucity of analysis of vocal communication is in some sense accidental and historical. Three of the most
compelling vertebrate neuroethological systems studied from a vocal communication perspective—bats, weakly electric fish, and songbirds—involve
autocommunication in the form of echolocation or feedback-mediated
learning. Although some lessons and principles may emerge from the study
of autocommunication that are common to sender/receiver systems, some
clearly cannot. In autocommunication, the animal as the receiver has knowledge of the timing and structure of its own motor behavior (although an
efference copy of the final motor output may not be directly available to
the CNS—e.g., Heiligenberg 1977). An independent receiver cannot have
such detailed knowledge. Thus, the signal coding and subsequent processing mechanisms, along with the role of attentional systems, are likely to vary
between receivers engaging in autocommunication and those that are not.
7. Neuroethology of Vocal Communication
329
that the acoustic entropy in the signal may drive neuronal habituation
mechanisms (see Ryan 1998; Gentner et al. 2001) as opposed to more direct
measures of repertoire sizes employed by researchers, such as simple song
counts.
1.3. Technical Challenges
Given the attractiveness of vocal communication systems as research
objects, it is somewhat surprising that the underlying neural mechanisms
are relatively poorly understood. More than anything else, this lack of
understanding reflects the tremendous technical challenges involved in
studying these complex biological systems. Assessing perceptual responses
to conspecific stimuli often requires elaborate conditions difficult to stage
in the laboratory, and ethologists have only cautiously embraced conditioning procedures otherwise common in the study of perceptual phenomena (e.g., Dooling and Searcy 1980; Weary and Krebs 1992; Adret 1993). In
part, such caution has been in an effort to avoid potential confusions that
can arise when interpreting arbitrary conditioning paradigms in the context
of natural behavior.
A related technical limitation is that the proper analysis of vocal communication will often require one or more animals to be actively engaged
in communication as the experimenter records physiological activity.
Ideally, this should include both field and laboratory settings. For the most
part, this vision is more fanciful than real because the ability to reliably
conduct single-neuron recordings in awake, behaving vertebrate animals
of the small size commonly employed in neuroethological studies is
only beginning to emerge (Dave et al. 1998b; Nieder and Klump 1999;
Venkatachalam et al. 1999), so far mainly under relatively constrained laboratory conditions (Yu and Margoliash 1996; Dave et al. 1998a; Nieder and
Klump 1999).
Yet another reason for the relative paucity of analysis of vocal communication is in some sense accidental and historical. Three of the most
compelling vertebrate neuroethological systems studied from a vocal communication perspective—bats, weakly electric fish, and songbirds—involve
autocommunication in the form of echolocation or feedback-mediated
learning. Although some lessons and principles may emerge from the study
of autocommunication that are common to sender/receiver systems, some
clearly cannot. In autocommunication, the animal as the receiver has knowledge of the timing and structure of its own motor behavior (although an
efference copy of the final motor output may not be directly available to
the CNS—e.g., Heiligenberg 1977). An independent receiver cannot have
such detailed knowledge. Thus, the signal coding and subsequent processing mechanisms, along with the role of attentional systems, are likely to vary
between receivers engaging in autocommunication and those that are not.
7. Neuroethology of Vocal Communication
329