because they link regional localization with profound yet restricted perceptual specializations (Damasio et al. 1990). Animal studies complement
this analysis by permitting controlled experimental manipulations. In electric fish, the magnitude and phase of the electric organ discharge are
processed separately by different “P” (probability) and “T” (timing) classes
of receptors that encode amplitude and phase, respectively, and distinct
pathways arising from these receptor classes ascend the CNS until they converge at higher levels (Heiligenberg 1991). The neurons in the highest levels
can be described as “recognition” neurons, whose activity is directly predictive of behavior (Rose et al. 1988). In bats, separate ascending systems
of projections and multiple cortical areas appear to mediate the differential processing of constant-frequency and frequency-modulated components of the echolocation calls (Olsen and Suga 1991a, 1991b; see Casseday
and Covey 1995), and lesion studies support this idea (Riquimaroux et al.
1991). Similarly, in barn owls, sound amplitude and timing information
appear to be encoded in separate ascending pathways prior to convergence
at the level of the midbrain (Moiseff and Konishi 1981; Sullivan and Konishi
1984). Again, lesions of each pathway produce specific behavioral deficits
associated with loss of discrimination of one dimension in the parameter
space but not the other (Takahashi et al. 1984). In frogs, the different frequency pathways that arise in the periphery are eventually combined at the
level of single neurons (Fuzessery and Feng 1982, 1983). A similar conclusion appears to obtain for the primate auditory system, where separate
ascending pathways exhibit differential sensitivity for stimulus morphology
and location (Romanski et al. 1999; Rauschecker and Tian 2000). Collectively, these and other data provide strong support for a hierarchical-,
modality-, and parameter-specific organization of vertebrate sensory
systems (Ulinski 1984).
2.1.3. Combination Sensitivity
Neurons can exhibit processing for complex sounds by virtue of elaborations in spectral or temporal components of their receptive fields. In audition, the sequence of discrete events is critical for perception, and neuronal
specializations for detecting sequences of events can be expected. One form
of neuronal sequence sensitivity is called combination sensitivity. A neuron
that is combination-sensitive responds with nonlinear summation (also
called facilitation) to a combination of sounds (either discrete spectral
components of the sound or two or more temporally discrete elements of
a sequence) as compared with the response of the neuron to subsets
(typically individual components) of the sound. There is no standard for
what a “sufficient” nonlinear response is for a neuron to be considered
combination-sensitive as long as the additional nonlinear component
reaches statistical significance. Neuronal combination sensitivity may be far
more common than has been appreciated (e.g., Brosch and Schreiner 2000;
7. Neuroethology of Vocal Communication
335
this analysis by permitting controlled experimental manipulations. In electric fish, the magnitude and phase of the electric organ discharge are
processed separately by different “P” (probability) and “T” (timing) classes
of receptors that encode amplitude and phase, respectively, and distinct
pathways arising from these receptor classes ascend the CNS until they converge at higher levels (Heiligenberg 1991). The neurons in the highest levels
can be described as “recognition” neurons, whose activity is directly predictive of behavior (Rose et al. 1988). In bats, separate ascending systems
of projections and multiple cortical areas appear to mediate the differential processing of constant-frequency and frequency-modulated components of the echolocation calls (Olsen and Suga 1991a, 1991b; see Casseday
and Covey 1995), and lesion studies support this idea (Riquimaroux et al.
1991). Similarly, in barn owls, sound amplitude and timing information
appear to be encoded in separate ascending pathways prior to convergence
at the level of the midbrain (Moiseff and Konishi 1981; Sullivan and Konishi
1984). Again, lesions of each pathway produce specific behavioral deficits
associated with loss of discrimination of one dimension in the parameter
space but not the other (Takahashi et al. 1984). In frogs, the different frequency pathways that arise in the periphery are eventually combined at the
level of single neurons (Fuzessery and Feng 1982, 1983). A similar conclusion appears to obtain for the primate auditory system, where separate
ascending pathways exhibit differential sensitivity for stimulus morphology
and location (Romanski et al. 1999; Rauschecker and Tian 2000). Collectively, these and other data provide strong support for a hierarchical-,
modality-, and parameter-specific organization of vertebrate sensory
systems (Ulinski 1984).
2.1.3. Combination Sensitivity
Neurons can exhibit processing for complex sounds by virtue of elaborations in spectral or temporal components of their receptive fields. In audition, the sequence of discrete events is critical for perception, and neuronal
specializations for detecting sequences of events can be expected. One form
of neuronal sequence sensitivity is called combination sensitivity. A neuron
that is combination-sensitive responds with nonlinear summation (also
called facilitation) to a combination of sounds (either discrete spectral
components of the sound or two or more temporally discrete elements of
a sequence) as compared with the response of the neuron to subsets
(typically individual components) of the sound. There is no standard for
what a “sufficient” nonlinear response is for a neuron to be considered
combination-sensitive as long as the additional nonlinear component
reaches statistical significance. Neuronal combination sensitivity may be far
more common than has been appreciated (e.g., Brosch and Schreiner 2000;
7. Neuroethology of Vocal Communication
335