Evidence from a simple two-channel model for asymptotic gap detection

Abstract

Detection of silent temporal gaps is characterized by two prominent features when measured as a function of frequency separation between two sinusoids that mark the silent gap. First, over a range of about an octave separation, gap detection thresholds (GDTs) routinely increase as the frequency difference is increased between the two sinusoids. Second, GDTs become asymptotic for greater sinusoidal frequency separations. This characteristic GDT pattern probably reflects two different processes. The first process almost certainly reflects gap detection measured within a single auditory filter or channel. The nature of the second process is less certain, but may reflect across-channel processing of the silent gap stimulus in two or more independent frequency channels. To evaluate the idea that asymptotic GDTs can be explained with a simple two-channel model, GDTs were measured as a function of frequency separation between a pre-gap sinusoid presented to the left ear (channel 1) and a post-gap sinusoid presented to the right ear (channel 2). The resulting GDTs for standard pre-gap sinusoids from 250 to 2000 Hz correspond closely with asymptotic GDTs measured for three listeners for the same set of stimulus conditions presented monaurally. This correspondence of the data supports a two-channel hypothesis. [Research supported by NIH.]