COMPARISON OF NOISE AND TONE AZIMUTH TUNING OF NEURONS IN CAT PRIMARY AUDITORY-CORTEX AND MEDIAL GENICULATE-BODY

1. A comparison of the azimuth tuning of single neurons to broadband noise and to best frequency (BF) tone bursts was made in primary auditory cortex (AI: n = 173) and the medial geniculate body (MGB: n = 52) of barbiturate-anesthetized cats. Observations were largely restricted to cells located within the tonotopically organized divisions of the MGB (i.e., the ventral nucleus and the lateral division of the posterior nuclear group) and the middle layers of AT. All cells studied had BFs greater than or equal to 4 kHz. 2. The responses of each cell to sounds presented from seven frontal azimuthal locations (-90 to +90 degrees in 30 degrees steps; 0 degrees elevation) and at five sound pressure levels (SPLs: 0-80 dB or 5-85 dB in 20-dB steps) provided an azimuth-level data set. Responses were averaged over SPL to obtain an azimuth function, and a number of features of this function were used to describe azimuth tuning to noise and to tone stimulation. Azimuth function modulation was used to assess azimuth sensitivity, and cells were categorized as sensitive or insensitive depending on whether modulation was greater than or equal to 75% or <75% of maximum, respectively. The majority (88%) of cells in the sample were azimuth sensitive to noise stimulation, and statistical analyses were restricted to these cells, which are presumably best suited to encode sound source azimuth. Azimuth selectivity was assessed by a preferred azimuth range (PAR) over which azimuth function values exceeded 75% (PAR(75)) or 50% of maximum response. Cells were categorized according to the location and extent of their noise PARs. Unbounded cells had laterally located PARs that extended to the lateral pole (+/-90 degrees); bounded cells had PARs that were contained entirely within the frontal hemifield, and a subset of these had PARs centered on the midline (+/-15 degrees). A final group of cells exhibited multipeaked azimuth functions to noise stimulation. 3. Azimuth functions to noise were generally more selective and/or more sensitive than those to tones. Statistical analyses showed that these differences were significant for cells in each azimuth function category, and for the thalamic and cortical samples. With the exception of multipeaked cells, responsiveness to noise was significantly lower than that to tones in all categories, and for the thalamic and cortical samples. The slope of the azimuth function, defined by the range of azimuths over which the cell's response changed from 25 to 75% of maximum, tended to be steeper to noise than that to tones; this difference was significant in the midline and unbounded cell groups. The majority of cells (89.5%) showed best azimuths (midpoint of the PAR(75)) to the two stimuli that differed by less than or equal to 30 degrees. This indicates that although many cells were more narrowly tuned to the azimuth of noise than BF tones, they tended to have similarly located noise and tone PARs. 4. Azimuth-level data sets were averaged over azimuth to obtain a level function. The nonmonotonic strength of the level function was defined by the percentage reduction in responsiveness at the highest level tested. The effect of bandwidth on azimuth selectivity was slightly greater for neurons that showed strongly nonmonotonic level functions than for those with weakly nonmonotonic functions. There was no relationship between the nonmonotonic strength of the response to noise stimulation and differences observed in azimuth function modulation to the two stimuli. 5. Fifty cells were studied with reversible ear occlusion to obtain information on their binaural inputs and interactions, and this was related to differences in their azimuth tuning to noise and to tones. Cells were classified according to whether their azimuth tuning depended on monaural spectral cues [monaural directional (MD) cells] or binaural disparities [binaural directional (ED) cells]. Six MD cells received excitatory input from one ear with no evidence of input from the other (MD-EO), and these cells showed far broader and less modulated azimuth functions to tones than to noise. Their azimuth tuning was apparently derived from spectral cues present in broadband but not tonal stimuli. MD cells that received inhibitory input from the nonexcitatory ear (n = 11) showed the same trends as for MD-EO cells, although differences were less dramatic because the inhibitory input shaped the cell's response to tones. 6. The azimuth tuning of 18 cells that responded maximally to azimuths about the midline was found to be a product of binaural facilitation. The other class of ED cells (n = 15) received excitatory input from one ear (usually the contralateral ear) and inhibitory input (or mixed inhibitory-facilitatory input) from the other. These cells typically responded well throughout most or all of one lateral hemifield. A substantial proportion of cells within both groups showed greater selectivity and/or sensitivity to noise than to tones, although the differences between azimuth tuning to the two stimuli were less dramatic than those observed in either group of MD cells. These data suggest that some aspect of a broadband stimulus also contributes to azimuth tuning in binaural cells. 7. The consistency in azimuth preference and the narrower tuning observed in many cells' responses to noise compared with high-frequency pure tones is compatible with behavioral studies in cats, monkeys, and humans that have shown that both these stimuli can be localized in the horizontal plane but that performance is more accurate to noise than to tones.