1. The interaural-phase-difference (IPD) sensitivity of single neurons in the primary auditory (AI) cortex of the anesthetized cat was studied at sitmulus frequencies ranging from 120 to 2,500 Hz. Best frequencies of the 43 AI cells sensitive to IPD ranged from 190 to 2,400 Hz. 2. A static IPD was produced when a pair of low-frequency tone bursts, differing from one another only in starting phase, were presented dichotically. The resulting IPD-sensitivity curves, which plot the number of discharges evoked by the binaural signal as a function of IPD, were deeply modulated circular functions. IPD functions were analyzed for their mean vector length (r) and mean interaural phase (∅). Phase sensitivity was relatively independent of best frequency (BF) but highly dependent on stimulus frequency. Regardless of BF or stimulus frequency within the excitatory response area the majority of cells fired maximally when the ipsilateral tone lagged the contralateral signal and fired least when this interaural-phase relationship was reversed. 3. Sensitivity to continuously changing IPD was studied by delivering to the two ears 3-s tones that differed slightly in frequency, resulting in a binaural beat. Approximately 26% of the cells that showed a sensitivity to static changes in IPD also showed a sensitivity to dynamically changing IPD created by this binaural tonal combination. The discharges were highly periodic and tightly synchronized to a particular phase of the binaural beat cycle. High synchrony can be attributed to the fact that cortical neurons typically respond to an excitatory stimulus with but a single spike that is often precisely timed to stimulus onset. A period histogram, binned on the binaural beat frequency (f(b)), produced an equivalent IPD-sensitivity function for dynamically changing interaural phase. For neurons sensitive to both static and continuously changing interaural phase there was good correspondence between their static (∅(s)) and dynamic (∅(d)) mean interaural phases. 4. All cells responding to a dynamically changing stimulus exhibited a linear relationship between mean interaural phase and beat frequency. Most cells responded equally well to binaural beats regardless of the initial direction of phase change. For a fixed duration stimulus, and at relatively low f(b), the number of spikes evoked increased with increasing f(b), reflecting the increasing number of effective stimulus cycles. At higher f(b), AI neurons were unable to follow the rate at which the most effective phase repeated itself during the 3 s of stimulation. 5. Changing stimulus intensity equally at both ears at BF resulted in relatively little systematic change in the interaural-phase sensitivity when data obtained near threshold and at nonmonotonic levels were excluded from the analysis. In a few cells where the effects of intensity change were studied at a number of excitatory frequencies the results were the same. 6. The IPD functions were converted to interaural-time-difference (ITD) functions. Approximately 85% of the curves had their peaks or troughs within the physiological range of ± 400 μs, and most (91%) of the peaks occurred when the ipsilateral tone lagged the contralateral (average: +142 μs), whereas most (79%) of the troughs occurred at an ipsilateral lead (average: -144 μs). The same trends were seen when a neuron's ITD sensitivity was represented by a composite function. For the majority of cells analyzed in this way both the composite peak and trough were within the physiological range of ITD. 7. The slope of the linear regression line fitted to the plot of mean interaural phase-versus-frequency yielded the characteristic delay (CD) for that neuron, and the γ-intercept its characteristic phase. When ITD functions from a single cell obtained at different frequencies were plotted on a common time axis, they rarely showed an exact registration near their peaks or troughs; the CDs could be located anywhere along the spike count-versus-ITD function. Of the 14 cells for which such analysis was possible, 12 of them had CDs that fell within the ± 300 μs range with values almost equally distributed about zero ITD (average: +49 μs).
