Representation of stimulus azimuth by low-frequency neurons in inferior colliculus of the cat

The responses to changes in stimulus azimuth of 204 neurons in the inferior colliculus of the cat with best frequencies of <3 kHz were studied. Three main unit classes were identified: omnidirectional units uninfluenced by speaker azimuth (39%); multipeaked units with several azimuthal regions at which peak firing occurred (9%); and azimuth-selective units that showed clear preferences for a discrete range of sound-source azimuths (52%). Alterations in stimulus intensity produced little change in the shapes of profiles relating firing rate to stimulus azimuth (azimuth functions), but the peaks of these functions could shift by up to 20°. Similar observations were made for a small sample of units, each of which was examined with a variety of stimulus frequencies. The pinnae were removed bilaterally in 11 cats, and azimuth functions for 35 units were measured both binaurally and with the ipsilateral or contralateral ear occluded. Evidence was found for facilitative or suppressive interactions, as a function of stimulus azimuth, between monaural inputs. The sharpness of an azimuth function was expressed by the half-width of the function, i.e., the number of degrees of azimuth between the peak of the function and the point at which 50% of maximum firing occurred on the medial side of the peak. When half-widths were plotted as a function of best frequency, it was found that the sharpest azimuth functions occurred for units with best frequencies between 1.1 and 1.5 kHz. Most units in the lowest two octaves (0.1-0.4 kHz) or having best frequencies between 2 and 3 kHz were omnidirectional. The relationship between half-width and the azimuth at which peak firing occurred (best azimuth) revealed that a range of best azimuths between 10 and 40° contralateral contained the sharpest azimuth functions. When best frequency was plotted against best azimuth, it was observed that the majority of units (93%) had best azimuths in the contralateral hemifield. For frequencies between 0.7 and 1.7 kHz, best azimuths occurred relatively evenly between 10 and 60° contralateral. These data suggested that for frequencies between 1.2 and 1.4 kHz, at least, the best azimuths of different units with the same best frequency collectively provide information about stimulus location across much of the contralateral hemifield. The best azimuths of azimuth-selective units were related to location within the inferior colliculus for 13 microelectrode penetrations; in three cats, multiple penetrations were made into the inferior colliculus. This reconstruction indicated a clear tendency for more peripheral best azimuths to be located rostrally and medially compared with best azimuths close to the midline, which occurred caudally and laterally. Thus there is a clear indication of a topographical organization of stimulus azimuth in the inferior colliculus, at least for low frequencies. Findings were discussed in relation to neural mechanisms for sound localization, sound localization capacities, and the auditory spatial topography observed in other studies.