1. The functional organization of neuronal tuning to the azimuthal location and sound pressure level (SPL) of noise bursts was studied in high-frequency primary auditory cortex (AI) of barbiturate-anesthetized cats. Three data collection strategies were used to map neural responses: 1) electrode penetrations oriented normal to the cortical surface provided information on the radial organization of neurons' responses; 2) neurons' responses were examined at a few points in the middle cortical layers in multiple normal penetrations across AI to produce fine-grain maps of azimuth and level selectivity; and 3) electrode penetrations oriented tangential to the cortical surface provided information on neurons' responses along the isofrequency dimension. 2. An azimuth-level data set was obtained for each single- or multiple- (multi-) unit recording; this consisted of responses to noise bursts at five SPLs (0-80 dB in 20-dB steps) from seven azimuthal locations in the frontal hemifield (-90 to +90 degrees in 30 degrees steps; 0 degrees elevation). An azimuth function was derived from these data by averaging response magnitude over all SPLs at each azimuth tested. A preferred azimuth range (PAR; range of azimuths over which the response was greater than or equal to 75% of maximum) was calculated from the azimuth function and provided a level-independent measure of azimuth selectivity. Each PAR was assigned to one of four azimuth preference categories (contralateral-, midline-, ipsilateral-preferring, or broad/multipeaked) according to its location and extent. A level function obtained from the data set (responsiveness averaged over all azimuths) was classified as monotonic if it showed a decrease of less than or equal to 25% (relative to maximum) at the highest SPL tested (usually 80 dB), and nonmonotonic if it showed a decrease of >25%. The percentage reduction in responsiveness, relative to maximum, at the highest lever tested (termed nonmonotonic strength) and the preferred level range (PLR; range of SPLs over which responsiveness was greater than or equal to 75% of maximum) of each response was also determined. 3. Normal penetrations typically showed a predominance of one azimuth preference category and/or level function type. The majority of penetrations (26/36: 72.2%) showed statistically significant azimuth preference homogeneity, and approximately one-half (17/36: 47.2%) showed significant level function type homogeneity. Over one-third (13/36) showed significant homogeneity for both azimuth preference and level function type. 4. Mapping experiments (n = 4) sampled the azimuth and level response functions at two or more depths in closely spaced normal penetrations that covered several square millimeters of AI. Cortical sites containing neurons with the same azimuth preference and/or level function type were grouped together. The most extensive map revealed a contralateral-preferring band of cortex that was elongated parallel to the frequency gradient and covered a frequency range from 8 to 20 kHz. Immediately ventral to this was a midline-preferring region that covered the same frequency range. 5. Azimuth and level response functions were obtained at closely spaced sites along 20 tangential penetrations oriented parallel to isofrequency contours, and most penetrations showed sequential responses that exhibited similar azimuth preferences and/or level function types. The majority (15/20: 75%) of penetrations showed significant clustering of neurons with the same azimuth preference. There were often multiple, discontinuous groups of neurons with similar azimuth preferences (usually contralateral) along isofrequency bands. Systematic shifts in the azimuth preference of successive neurons or groups of neurons were not commonly seen. Over one-half (12/20: 60%) the penetrations showed significant clustering of monotonic and nonmonotonic response types. 6. The organization of neurons with differing degrees of azimuth sensitivity, as measured by azimuth function modulation, was also examined. Neurons' responses were categorized as either high directional (HD; greater than or equal to 75% modulation) or low directional (LD; <75% modulation). Only a minority of normal and tangential penetrations showed significant homogeneity or clustering of neurons with HD or LD responses. 7. In an attempt to validate the use of the multiunit technique to study functional organization, comparisons of the azimuth and level response properties of single and multiunits were made. This was achieved by 1) comparing the responses of multi/single-unit pairs (n = 114) recorded at the same location and single-unit pairs (n = 16) recorded at the same location; and 2) comparing the distribution of responses falling into the different azimuth and level categories from a large multiunit sample and a comparable single-unit sample collected during these experiments. Multi- and single-unit recordings showed differences in nonmonotonic strength and azimuth function modulation. However, measures of azimuth preference and selectivity (best azimuth) or level selectivity (best SPL) that were derived from the range of maximum responsiveness of the azimuth or level function showed significant correlations.
