acoustically complex vocal communication signals. Although we have
attempted to maintain a focus on perceptual mechanism, the nature of vocal
communication systems necessitates an integrative approach to production
and perception. Our discussion of motor-sensory linkages in the wellstudied oscine song system addresses these concerns from both empirical
and theoretical perspectives.
Much of the auditory cortex in mammals and auditory forebrain in birds
is mapped tonotopically, but a number of additional acoustic stimulus
parameter mappings have been described in birds and mammals, including
pitch periodicity, frequency, intensity, spatial location, duration, amplitude,
and frequency modulation (see Ehret 1997; Schulze and Langner 1997). The
structure of these mappings may change significantly in awake, alert animals
(Evans and Whitfield 1964; Pfingst et al. 1977; Dave et al. 1998a; Schmidt
and Konishi 1998; Capsius and Leppelsack 1999; cf. Recanzone et al. 2000).
Patterns of organization in numerous other systems—including barn owls,
bats, electric fish, songbirds, and others—suggest that a breakdown in
cochleotopic organization of responses often coincides with mappings for
more complex stimulus parameters. This may represent the convergence of
multiple acoustically simpler maps onto some regions (i.e., a stimulus reconstruction) or the extraction of encoded information along dimensions that
have not been represented in prior mappings (i.e., emergent properties).
The neuroethological approach provides a logical and theoretically sound
basis for investigation of such emergent properties in complex signals.
The neuroethological approach to vocal communication is also informative of the likely functional outputs of a perceptual system. These outputs form the basis for higher cognition. We have described how
vocal-recognition behavior, and more explicitly the organization of vocal
communication signals into behaviorally relevant classes or categories, can
be used to develop hypotheses about the perceptual and cognitive demands
placed on the central nervous system.
Principles in neuroethology have been difficult to identify. Most reviews
of neuroethological research are organized around a series of case
examples. This can incorrectly reinforce the conclusion that principles of
neuroethological research are not forthcoming. In contrast, for example,
systems neuroscience has identified a number of features of CNS organization, such as mappings, lateral interactions, population dynamics, and
network reconfiguration, which help to organize research and can be
considered principles. Neuroethological principles do exist, such as those
described in this paper, but they have been difficult to demonstrate because
of the scope of work required for a comparative analysis of brain and
behavior. Only now, after some 40 years of research, and only in the most
extensively studied behaviors, such as vocal communication, are these principles emerging. Somewhat akin to neurobiology, neuroethology aims to
explain how molecular, cellular, and systems-level phenomena result in
behaviors such as perception and cognition. But by the very nature of its
368
T.Q. Gentner and D. Margoliash
attempted to maintain a focus on perceptual mechanism, the nature of vocal
communication systems necessitates an integrative approach to production
and perception. Our discussion of motor-sensory linkages in the wellstudied oscine song system addresses these concerns from both empirical
and theoretical perspectives.
Much of the auditory cortex in mammals and auditory forebrain in birds
is mapped tonotopically, but a number of additional acoustic stimulus
parameter mappings have been described in birds and mammals, including
pitch periodicity, frequency, intensity, spatial location, duration, amplitude,
and frequency modulation (see Ehret 1997; Schulze and Langner 1997). The
structure of these mappings may change significantly in awake, alert animals
(Evans and Whitfield 1964; Pfingst et al. 1977; Dave et al. 1998a; Schmidt
and Konishi 1998; Capsius and Leppelsack 1999; cf. Recanzone et al. 2000).
Patterns of organization in numerous other systems—including barn owls,
bats, electric fish, songbirds, and others—suggest that a breakdown in
cochleotopic organization of responses often coincides with mappings for
more complex stimulus parameters. This may represent the convergence of
multiple acoustically simpler maps onto some regions (i.e., a stimulus reconstruction) or the extraction of encoded information along dimensions that
have not been represented in prior mappings (i.e., emergent properties).
The neuroethological approach provides a logical and theoretically sound
basis for investigation of such emergent properties in complex signals.
The neuroethological approach to vocal communication is also informative of the likely functional outputs of a perceptual system. These outputs form the basis for higher cognition. We have described how
vocal-recognition behavior, and more explicitly the organization of vocal
communication signals into behaviorally relevant classes or categories, can
be used to develop hypotheses about the perceptual and cognitive demands
placed on the central nervous system.
Principles in neuroethology have been difficult to identify. Most reviews
of neuroethological research are organized around a series of case
examples. This can incorrectly reinforce the conclusion that principles of
neuroethological research are not forthcoming. In contrast, for example,
systems neuroscience has identified a number of features of CNS organization, such as mappings, lateral interactions, population dynamics, and
network reconfiguration, which help to organize research and can be
considered principles. Neuroethological principles do exist, such as those
described in this paper, but they have been difficult to demonstrate because
of the scope of work required for a comparative analysis of brain and
behavior. Only now, after some 40 years of research, and only in the most
extensively studied behaviors, such as vocal communication, are these principles emerging. Somewhat akin to neurobiology, neuroethology aims to
explain how molecular, cellular, and systems-level phenomena result in
behaviors such as perception and cognition. But by the very nature of its
368
T.Q. Gentner and D. Margoliash