In addition, in echolocation systems and passive localization systems, the
structure of the vocal signal, or the parameters of the signal to be analyzed,
are strongly and directly constrained by physical characteristics of the
signal-generating mechanism, the medium for signal transmission, and the
physical characteristics of the target. This means that sensory mechanisms
for localization are influenced principally by natural selection. In all localization systems, there is compelling evidence for behavioral sensitivity to
microsecond timing and considerable evidence for specialized neural
systems to accurately represent information on such time scales. To date,
there is no behavioral or physiological evidence to suggest that receiver sensitivity to microsecond timing is important outside the context of autocommunication. Vocal communication, on the other hand, generally arises
in the context of reproductive behaviors such as courtship and territorial
defense. Thus, vocal communication is influenced principally by sexual
selection, although physical constraints can also influence the design of
communication signals (Konishi 1970; Ryan and Kime, Chapter 5). The differences in selection pressures on different types of signaling may also have
profound influences on CNS organization.
2. Organizational Principles for Neural Representations
of Conspecific Vocalizations
Internal representations of information in the world are a function of both
the form of the input and state of the neural system. Neuroethological
studies using conspecific vocalizations bridge these “input” and “state” components in a biologically plausible way and have made important contributions to theories of representation and brain organization at multiple
levels of analysis. In this section, we focus on the organizational principles
of representation that underlie the discrimination and classification of vocal
communication signals. These neuroethological analyses provide insight
into behavioral mechanisms at both proximate and ultimate levels. At the
proximate level, such studies relate to questions of neural coding, such as
the forms of hierarchical and parallel representation, the representation of
spectrally and temporally complex acoustic signals, and representational
plasticity. Insights into ultimate processes that derive from neuroethological studies of conspecific vocalizations include coupling mechanisms
between motor and sensory systems and the relationship between vocal
recognition at inter- and intra-specific levels.
2.1. The Feature Detector Hypothesis
An understanding of the hierarchical and parallel organization of neural
activity patterns is fundamental to the study of integrative brain mechanisms. Behavioral phenomena such as sensory perceptions may manifest
330
T.Q. Gentner and D. Margoliash
structure of the vocal signal, or the parameters of the signal to be analyzed,
are strongly and directly constrained by physical characteristics of the
signal-generating mechanism, the medium for signal transmission, and the
physical characteristics of the target. This means that sensory mechanisms
for localization are influenced principally by natural selection. In all localization systems, there is compelling evidence for behavioral sensitivity to
microsecond timing and considerable evidence for specialized neural
systems to accurately represent information on such time scales. To date,
there is no behavioral or physiological evidence to suggest that receiver sensitivity to microsecond timing is important outside the context of autocommunication. Vocal communication, on the other hand, generally arises
in the context of reproductive behaviors such as courtship and territorial
defense. Thus, vocal communication is influenced principally by sexual
selection, although physical constraints can also influence the design of
communication signals (Konishi 1970; Ryan and Kime, Chapter 5). The differences in selection pressures on different types of signaling may also have
profound influences on CNS organization.
2. Organizational Principles for Neural Representations
of Conspecific Vocalizations
Internal representations of information in the world are a function of both
the form of the input and state of the neural system. Neuroethological
studies using conspecific vocalizations bridge these “input” and “state” components in a biologically plausible way and have made important contributions to theories of representation and brain organization at multiple
levels of analysis. In this section, we focus on the organizational principles
of representation that underlie the discrimination and classification of vocal
communication signals. These neuroethological analyses provide insight
into behavioral mechanisms at both proximate and ultimate levels. At the
proximate level, such studies relate to questions of neural coding, such as
the forms of hierarchical and parallel representation, the representation of
spectrally and temporally complex acoustic signals, and representational
plasticity. Insights into ultimate processes that derive from neuroethological studies of conspecific vocalizations include coupling mechanisms
between motor and sensory systems and the relationship between vocal
recognition at inter- and intra-specific levels.
2.1. The Feature Detector Hypothesis
An understanding of the hierarchical and parallel organization of neural
activity patterns is fundamental to the study of integrative brain mechanisms. Behavioral phenomena such as sensory perceptions may manifest
330
T.Q. Gentner and D. Margoliash