vocalizations, and the specificity of the response properties shows increasing complexity between field L and NCM/cHV (Leppelsack and Vogt 1976;
Müller and Leppelsack 1985; see Section 3). In general, neurons responsive
to specific, behaviorally significant features of vocalizations form a plausible substrate for categorical representations (see Section 2.1). Very few
studies, however, have explored the neuronal responses to natural variation
within and between categories of stimuli (Freedman et al. 2001). This work
is only now beginning in primates and birds (Gentner and Margoliash 2001;
Tian et al. 2001; Wang and Kadia 2001). This is an area ripe for future
research. In Section 3, we provide a logical basis for understanding these
neural mechanisms in the context of vocal recognition.
3. Emerging Model Systems for the Neuroethology
of Vocal Recognition
The preceding sections discuss a number of ways in which neuroethological approaches using vocal communication signals can constrain both the
input to and state of the representational system as well as its output. The
biological plausibility of these constraints derives directly from the functional (i.e., adaptive) role served by vocal communication signals. Thus, the
extent to which these various constraints can be invoked for any one system
depends on our knowledge of how that organism uses particular signals
under natural conditions. Ideally, one wants to know how behaviors are
driven by acoustic variation in a given communication signal and then adapt
the critical components of such behaviors to laboratory procedures that are
amenable to physiological preparations. This requires sophisticated knowledge of both behavior and basic auditory physiology, and to date only a few
systems meet these requirements. In this section, we review the central auditory physiology and vocal-recognition behavior in two communication
systems, those of songbirds and primates. We concentrate on vocal recognition because it captures many features of communication systems described above (e.g., categorization) and is widespread among many taxa,
making the results amenable to comparative analyses.
3.1. Perception of Vocal Signals
Variation in communication signals can occur in the spectro-temporal properties of the signal itself and also in the spatial-temporal distribution of
signal sources. Together, this variation leads to at least two general classes
of receiver behavior. The first class derives from the fact that not all acoustic
events are of equal interest, so animals must be able to dissociate appropriate target signals from irrelevant/background noise, including nontarget
conspecific vocalizations. This has been studied in the context of the so354
T.Q. Gentner and D. Margoliash
Müller and Leppelsack 1985; see Section 3). In general, neurons responsive
to specific, behaviorally significant features of vocalizations form a plausible substrate for categorical representations (see Section 2.1). Very few
studies, however, have explored the neuronal responses to natural variation
within and between categories of stimuli (Freedman et al. 2001). This work
is only now beginning in primates and birds (Gentner and Margoliash 2001;
Tian et al. 2001; Wang and Kadia 2001). This is an area ripe for future
research. In Section 3, we provide a logical basis for understanding these
neural mechanisms in the context of vocal recognition.
3. Emerging Model Systems for the Neuroethology
of Vocal Recognition
The preceding sections discuss a number of ways in which neuroethological approaches using vocal communication signals can constrain both the
input to and state of the representational system as well as its output. The
biological plausibility of these constraints derives directly from the functional (i.e., adaptive) role served by vocal communication signals. Thus, the
extent to which these various constraints can be invoked for any one system
depends on our knowledge of how that organism uses particular signals
under natural conditions. Ideally, one wants to know how behaviors are
driven by acoustic variation in a given communication signal and then adapt
the critical components of such behaviors to laboratory procedures that are
amenable to physiological preparations. This requires sophisticated knowledge of both behavior and basic auditory physiology, and to date only a few
systems meet these requirements. In this section, we review the central auditory physiology and vocal-recognition behavior in two communication
systems, those of songbirds and primates. We concentrate on vocal recognition because it captures many features of communication systems described above (e.g., categorization) and is widespread among many taxa,
making the results amenable to comparative analyses.
3.1. Perception of Vocal Signals
Variation in communication signals can occur in the spectro-temporal properties of the signal itself and also in the spatial-temporal distribution of
signal sources. Together, this variation leads to at least two general classes
of receiver behavior. The first class derives from the fact that not all acoustic
events are of equal interest, so animals must be able to dissociate appropriate target signals from irrelevant/background noise, including nontarget
conspecific vocalizations. This has been studied in the context of the so354
T.Q. Gentner and D. Margoliash