The structure of spatial receptive fields of neurons in primary auditory cortex of the cat

Abstract

Transient broad-band stimuli that mimic in their spectrum and time waveform sounds arriving from a speaker in free space were delivered to the tympanic membranes of barbiturized cats via sealed and calibrated earphones. The full array of such signals constitutes a virtual acoustic space (VAS). The extra-cellular response to a single stimulus at each VAS direction, consisting of one or a few precisely time-locked spikes, was recorded from neurons in primary auditory cortex. Effective sound directions form a virtual space receptive field (VSRF). Near threshold, most VSRFs were confined to one quadrant of acoustic space and were located on or near the acoustic axis. Generally, VSRFs expanded monotonically with increases in stimulus intensity, with some occupying essentially all of the acoustic space. The VSRF was not homogeneous with respect to spike timing or firing strength. Typically, onset latency varied by as much as 4-5 msec across the VSRF. A substantial proportion of recorded cells exhibited a gradient of first-spike latency within the VSRF. Shortest latencies occupied a core of the VSRF, on or near the acoustic axis, with longer latency being represented progressively at directions more distant from the core. Remaining cells had VSRFs that exhibited no such gradient. The distribution of firing probability was mapped in those experiments in which multiple trials were carried out at each direction. For some cells there was a positive correlation between latency and firing probability.

Fig. 1.

VSRFs obtained from two AI neurons at six intensity levels showing monotonically expanding and unbounded receptive field properties. In this and all other figures illustrating VSRFs, VAS is represented by a globe, opened and hinged at coordinates 0° elevation (EL) and 90° azimuth (AZ). Frontal acoustic hemifield is on the left; the rear acoustic hemifield is on the right. Contralateral acoustic space is represented in the contiguous right frontal and left rear quadrants. Stimulus intensity is given in decibels of attenuation (ATTN). SA, Spherical area of the VSRF, in spherical degrees; LI, laterality index. CF-tone binaural interaction: D9130M9, EI; D9130M7, EE. See text for further details.

Fig. 2.

VSRFs obtained from two AI neurons at different intensities, showing expanding but bounded receptive field properties. Neuron on the right exhibits a “fractured” pattern. CF-tone binaural interaction: D9064M6, EI; D9136M5, EI. See text and legend to Figure 1 for further details.

Fig. 3.

Frontal VSRFs obtained from two AI neurons at different intensities. Neuron on the left exhibits a “fractured” pattern. CF-tone binaural interaction: D9054M7, EE; D9139M8, EE. See text and legend to Figure 1 for further details.

Fig. 4.

Frontal VSRFs obtained from two AI neurons at different intensities. Neuron on the right exhibits a “fractured” pattern. CF-tone binaural interaction: D9224M2, PB; D9130M8, PB. See text and legend to Figure 1 for further details.

Fig. 5.

Laterality index and VSRF area plotted as a function of stimulus intensity for 14 neurons. A, B, Unbounded VSRFs; C, D, bounded VSRFs; E, F, frontal VSRFs. See text for further details.

Fig. 6.

Complex VSRFs from two neurons at five different intensities. Both neurons exhibit a “fractured” pattern. CF-tone binaural interaction: D9016M5, EE; D9054M13, EE. See text and legend to Figure 1 for further details.

Fig. 7.

A, Global representation of the distribution of the acoustic axis (filled circles) and the spatial median (open circles) for 65 contralateral and ipsilateral VSRFs. B, Scatter plot of acoustic axis versus VSRF median azimuth at CF. Upper right quadrantrepresents data in the contralateral acoustic hemifield; lower left quadrant represents data in the ipsilateral acoustic hemifield. C, Scatter plot of acoustic axis versus VSRF median elevation at CF. Upper right quadrant represents data in the upper right acoustic hemifield.

Fig. 8.

Distribution of first-spike latency for six AI neurons at different intensities. Number of spikes (N), mean first-spike latency (MEAN), and standard deviation (SD) given on each panel.

Fig. 9.

Distribution of first-spike latency for five AI neurons at different intensities. See legend to Figure 8 for further details.

Fig. 10.

“Ordered” VSRFs based on first-spike latency from one AI neuron obtained at six different intensities. Histograms show the frequency distribution of latency at each intensity.Left column, VSRFs color-coded for an absolute latency within a fixed binwidth of 1 msec. Colors in the VSRFs correspond to colors of bins in accompanying color bar and latency histogram.Right column, VSRFs based on the same data but plotted such that binwidth represents an equal (quintile) proportion of latency values. Colors in the VSRFs correspond to the colors of the accompanying color bar. CF-tone binaural interaction: D9130M9, EI. See text for further details.

Fig. 11.

“Ordered” VSRFs based on first-spike latency from five AI neurons, illustrating the range of VSRFs within that category. Intensity shown below each panel. CF-tone binaural interactions: D9054M12, EE; D9064M6, EI; D9064M9, EE; D9139M8, EE; D9054M9, EE. See legend to Figure 10 for additional details.

Fig. 12.

“Disordered” VSRFs based on first-spike latency from one AI neuron obtained at six different intensities. CF-tone binaural interaction: D9130M7, EE. See legend to Figure 10 and text for further details.

Fig. 13.

“Disordered” VSRFs based on first-spike latency from five AI neurons, illustrating the range of VSRFs within that category. Intensity shown below each panel. CF-tone binaural interactions: D9054M8, EE; D9054M13, EE; D9425M1, PB; D9064M2, EI; D9130M8, PB. See legend to Figure 11 for additional details.

Fig. 14.

VSRFs based on proportional latency (right column) and proportional firing probability (left column) in response to 15 stimulus trials at each VAS direction. Scatter plots show relationship between mean first spike latency and firing probability at each effective point in the corresponding VSRFs. Only a portion of the standard 1650 loci were tested in each case (A, n = 525; B, n = 654; C, n = 819; D,n = 575; E, n = 554). CF-tone binaural interactions: D93011M4, PB; D92100M2, unknown; D92103M6, EE; D9305M5, EE; D92100M4, EE. See text for further details.