ABSTRACT
We review some of the major trends in binaural modeling, particularly with regard to models based on the interaural cross-correlation of the auditory-nerve responses to the stimuli. Emphasis is placed on providing an intuitive understanding of cross-correlation-based binaural models, combined with an appreciation of their capabilities and limitations in describing a variety of binaural phenomena. We focus on the seminal theory of binaural processing by Jeffress and its later elaboration and quantification by Colburn. This theory describes and predicts binaural phenomena in terms of the putative activity of central units that record interaural coincidences of firing from matched pairs of auditory-nerve fibers (one from each ear). The input auditory-nerve fibers are matched in characteristic frequency with a fixed time delay inserted on one side. The response of a number of such central units at a given characteristic frequency, plotted as a function of internal delay, is an approximation to the interaural cross-correlation function of the sound as processed by the auditory periphery. We discuss predictions for many of the simple and complex binaural stimuli that are commonly used in psychoacoustical experiments. These experiments include measurements of subjective lateral position, interaural discrimination, binaural detection, and dichotic pitch.
