Gehr et al. 2000). Different magnitudes of the nonlinear components of the
combination response, and different temporal windows of integration, may
reflect different mechanisms that give rise to combination sensitivity.
2.1.3.1. Spectral Combination Sensitivity
Neurons with spectral combination sensitivity are reported in a broader
range of species than are neurons with temporal combination sensitivity.
This may reflect, at a neurobiological level, the fundamental nature of the
process of integration of spatial information arranged across a sensory
epithelium and, at a behavioral level, that sensitivity to temporal structure
is probably a secondary adaptation. Neurons that require two or more spectral lines before exhibiting a facilitated, typically excitatory response have
been well-characterized in bats (Suga et al. 1979), birds (Langner et al. 1981;
Margoliash and Fortune 1992), cats (Sutter and Schreiner 1991; Nelken et
al. 1994a, 1994b), and frogs (Mudry et al. 1977; Fuzessery and Feng 1982,
1983), and there is weaker evidence in other systems as well. The responses
have almost always been linked to specific spectral components of vocalizations, whereas spectral combination sensitivity was not observed in
the auditory cortex of monkeys trained on a missing-fundamental task
(Schwarz and Tomlinson 1990). Spectral combination-sensitive neurons
can be common throughout the auditory system in some animals (bats), or
locally common within a specific neural structure (frogs, cats) or within a
specific neural pathway (songbirds). Because these response properties can
be difficult to identify without the adequate stimulus, the failure to find
spectral (or temporal) combination-sensitive neurons cannot be taken as
proof that such response properties are absent.
The mechanisms for producing spectral combination-sensitive neurons
are probably best described from extracellular recordings conducted in the
echolocating CF-FM mustached bat, Pteronotus parnelli. Members of this
species produce a biosonar “pulse” for orientation composed of a constantfrequency (CF) and frequency-modulated (FM) component, each of which
has four harmonics (Fig. 7.2). The returning echoes of these pulses are
delayed temporally, reflecting target distance, and have a Doppler shift in
frequency based on the relative velocity of the target, typically an insect.
Periodic frequency (and amplitude) modulations in the echo may result
from insect wing beats. Many types of combination-sensitive neurons have
been described in this system, and the auditory cortex can be divided into
multiple subregions based on the particular combinations of pulse and echo
components that give rise to facilitated responses. For example, neurons in
the CF 1 /CF 3 subregion are tuned to the first harmonic of the pulse and the
third harmonic of the CF component in the returning echo (see Fig. 7.2).
Within this region, specific cells are tuned to specific frequency differences
between the pulse and echo and/or phase-locked to frequency modulation
of the harmonic echo, representing a specific relative target velocity or
336
T.Q. Gentner and D. Margoliash
combination response, and different temporal windows of integration, may
reflect different mechanisms that give rise to combination sensitivity.
2.1.3.1. Spectral Combination Sensitivity
Neurons with spectral combination sensitivity are reported in a broader
range of species than are neurons with temporal combination sensitivity.
This may reflect, at a neurobiological level, the fundamental nature of the
process of integration of spatial information arranged across a sensory
epithelium and, at a behavioral level, that sensitivity to temporal structure
is probably a secondary adaptation. Neurons that require two or more spectral lines before exhibiting a facilitated, typically excitatory response have
been well-characterized in bats (Suga et al. 1979), birds (Langner et al. 1981;
Margoliash and Fortune 1992), cats (Sutter and Schreiner 1991; Nelken et
al. 1994a, 1994b), and frogs (Mudry et al. 1977; Fuzessery and Feng 1982,
1983), and there is weaker evidence in other systems as well. The responses
have almost always been linked to specific spectral components of vocalizations, whereas spectral combination sensitivity was not observed in
the auditory cortex of monkeys trained on a missing-fundamental task
(Schwarz and Tomlinson 1990). Spectral combination-sensitive neurons
can be common throughout the auditory system in some animals (bats), or
locally common within a specific neural structure (frogs, cats) or within a
specific neural pathway (songbirds). Because these response properties can
be difficult to identify without the adequate stimulus, the failure to find
spectral (or temporal) combination-sensitive neurons cannot be taken as
proof that such response properties are absent.
The mechanisms for producing spectral combination-sensitive neurons
are probably best described from extracellular recordings conducted in the
echolocating CF-FM mustached bat, Pteronotus parnelli. Members of this
species produce a biosonar “pulse” for orientation composed of a constantfrequency (CF) and frequency-modulated (FM) component, each of which
has four harmonics (Fig. 7.2). The returning echoes of these pulses are
delayed temporally, reflecting target distance, and have a Doppler shift in
frequency based on the relative velocity of the target, typically an insect.
Periodic frequency (and amplitude) modulations in the echo may result
from insect wing beats. Many types of combination-sensitive neurons have
been described in this system, and the auditory cortex can be divided into
multiple subregions based on the particular combinations of pulse and echo
components that give rise to facilitated responses. For example, neurons in
the CF 1 /CF 3 subregion are tuned to the first harmonic of the pulse and the
third harmonic of the CF component in the returning echo (see Fig. 7.2).
Within this region, specific cells are tuned to specific frequency differences
between the pulse and echo and/or phase-locked to frequency modulation
of the harmonic echo, representing a specific relative target velocity or
336
T.Q. Gentner and D. Margoliash