Mechanisms of inhibition and neuronal integration for signal processing in the primary auditory cortex of the Mongolian gerbil (Meriones unguiculatus)

A. A fundamental property of hearing is the decomposition of
complex sounds into perceptually distinct frequency components.
Each receptor cell in the cochlea and most centrally located
neurons respond only to a limited range of frequencies. The
individual frequency channels are spatially organized on the
cortical surface. This consistent topographical pattern provides a
framework for the investigation of other functional organization
principles, e.g., the functional properties of neurons in the six
cortical layers and the responsiveness of neurons to complex
sounds. The frequency specific features of inhibition should play
an important role in shaping a neuron’s response to complex
behaviorally relevant stimuli. Physiological and immunocytochemical
evidence indicates a layer-dependent organization of inhibitory
circuits in the neocortex. To investigate the contribution of
GABAergic inhibition to frequency tuning in the different cortical
layers, single and multi units were recorded in near-radial
penetrations before and during iontophoretic application of the
GABAA-receptor antagonist bicuculline in the auditory cortex of the
lightly anaesthetized gerbil (Meriones unguiculatus). Bicuculline
generally increased the spontaneous neuronal activity and enhanced
and prolonged onset responses to sound. Application of bicuculline
often resulted in a shift of the most sensitive frequency of the
neurons’ receptive fields and a decrease of threshold (5.5 dB). A
broadening of the frequency tuning evident by lower Q40dB values
was observed in 63% of the units. In units with several peaks in
their tuning curve or clearly separated response areas, bicuculline
application removed inhibitory gaps in the receptive fields and
created single-peaked tuning curves. The influence of bicuculline
on the receptive field size was not significantly layer-specific
but tended to be most pronounced in layers V and VI. In layer VI,
"silent" neurons were frequently found that responded to sound only
when GABAergic inhibition was antagonized. From the analysis of
postembedding GABA immunocytochemistry, the proportion of GABAergic
neurons was found to be maximal in layers I and V, and the number
of GABAergic perisomatic puncta (axon terminals) on cell somata
peaked in layer V. The influence of bicuculline was compared with
the effects of two-tone suppression. It was found that in some
units, the effects of suppression could be partially mediated by
intracortical GABAergic inhibition. In some units in layers IV, V,
and VI, additionally to the initial excitatory activity in response
to stimulus onset, a second, long-lasting excitatory response
occurred several hundred milliseconds after the stimulus. This late
response was not dependent on stimulus duration and could be
enhanced or elicited by GABAA blockade. The fact that several,
rhythmically occurring late responses were elicited by the
application of bicuculline suggests that recurrent excitatory
networks can become entrained by small modifications of inhibition.
B. In the natural environment, acoustic signals like animals’
communication sound or human speech is often masked by background
noise. Amplitude fluctuations are often superimposed upon
environmental sounds on their path of transmission which can lead
to a distinct temporal structure of the sound. Furthermore, many
natural background sounds are often temporally structured.
Vertebrates have evolved mechanisms to exploit amplitude
modulations in background noise to improve signal detection.
Psychophysical and behavioral experiments have shown that
amplitude-modulated background noise (comodulated noise) is less
effective as a masker than unmodulated noise bands of the same
bandwidth, a phenomenon called comodulation masking release (CMR).
This phenomenon has been extensively studied in human
psychoacoustics. However, the underlying neural mechanisms are
still debated. Animal models in which a direct comparison of the
neuronal response and the behaviorally measured performance is
possible could increase our understanding of the underlying
mechanisms. CMR could be demonstrated behaviorally and
neurophysiologically in a songbird, however, models for mammals are
still lacking. In behavioral experiments, Kittel et al. (2000)
demonstrated CMR in the gerbil. In the present study, using
acoustic stimuli that were identical with those of a behavioral
experiment, a neural correlate of CMR was described in the auditory
cortex of the gerbil and compared with the behavioral data. In this
study of neural mechanisms of masking release in the primary
auditory cortex of the anaesthetized gerbil, I determined neural
detection thresholds for 200-ms test tones presented in a
background of band-pass amplitude modulated (50 Hz) noise maskers
of different bandwidth (between 50 and 3200 Hz). Neural release
from masking caused by comodulated band-pass noise was evident at
the level of the gerbil’s primary auditory cortex. On average, the
largest masking release (median 6.9 dB) was found for a masker
bandwidth of 3200 Hz. This is less than the median masking release
of 15.7 dB observed in the behavioral study in the gerbil. For most
masker bandwidths, however, a small fraction of the neurons
exhibited a masking release that was close to or even larger than
the behavioral masking release. The observation that the release
from masking increased as a function of the masker bandwidth
indicates that spectral components remote from the signal frequency
enhance the signal detection. However, there was no correlation
between the neurons’ filter bandwidths and the amount of masking
release. Thus, neuronal masking release in the gerbil primary
auditory cortex could be attributed to both signalmasker
interactions across different frequency channels and also to
mechanisms that act within a single frequency channel. The gerbil
appears to be a suitable animal model for additional studies
comparing behavioral and physiological performance in the same
species. These studies could increase our understanding of the
perceptual mechanisms that are useful for the analysis of auditory
scenes.