Theory of binaural interaction based on auditory-nerve data. IV. A model for subjective lateral position

Abstract

A model for the subjective lateral position of 500-Hz tones is presented and compared with experimental lateralization data. Previous papers in this series have explicitly described the auditory-nerve response to these stimuli and proposed a binaural displayer that interaurally compares the auditory-nerve firing times. The outputs of the displayer are postulated to represent the only information about detailed firing times that is available to the brain. In the present paper, lateral-position predictions are obtained by a central nonoptimal weighting of these outputs that depends on the interaural intensity difference of the tone. These predictions describe the results of lateralization-matching experiments more accurately and over a wider range of stimulus conditions than previous theories, except for those results which suggest that low-frequency binaural tones can generate multiple perceptual images. The predictions of our model are also consistent with the results of centering and laterality-comparison experiments. It is argued that the data discussed in this paper are generally incompatible with theories that propose a peripheral interaction of interaural timing and intensity information such as the latency hypothesis.

Neural and Behavioral Sensitivity to Interaural Time Differences Using Amplitude Modulated Tones with Mismatched Carrier Frequencies

“…Binaural neurons act as coincidence detectors where maximal firing occurs when inputs from each side arrive at the same time, as in the original Jeffress (1948) model (see Joris et al 1998 for a review). Modern implementations of this model are based on binaural elements that receive matched frequency inputs from each side (Colburn 1973;Stern and Colburn 1978). Presentation of bilaterally mismatched frequencies can stimulate these coincidence detectors because of the spread of excitation into bilaterally matched channels.…”

Lateralization of two-transient stimuli

“…The basic bandpass correlation model implemented here incorporates the features of cross-correlation mod- els that have evolved over the years to account for a variety of steady-state lateralization data (e.g., Colburn, 1996;Licklider, 1959;Sayers & Cherry, 1957;Stern & Colburn, 1978). The core of the processing is the calculation of cross-correlation functions between counterpart left and right bandpass signals.…”

“…A weighting function is applied to each bandpass cross-correlation function along the correlation delay dimension. This function, p, applies uniform maximal weighting over the range between 2300 and +300 msec, with exponential tapering beyond those limits (Stern & Colburn, 1978). A rectangular window is also applied to the cross-correlation functions from 21 to +1 msec to eliminate contributions outside of that range.…”

Multi-microphone correlation-based processing for robust speech recognition

“…Nevertheless, most previous studies have used cross-correlation-based processing to identify the direction of a desired sound source, rather than to improve the quality of input for speech recognition (e.g. [7,8]). …”