Comparison of auditory spatial bisection and minimum audible angle in front, lateral, and back space

Abstract

Although vision is important for calibrating auditory spatial perception, it only provides information about frontal sound sources. Previous studies of blind and sighted people support the idea that azimuthal spatial bisection in frontal space requires visual calibration, while detection of a change in azimuth (minimum audible angle, MAA) does not. The influence of vision on the ability to map frontal, lateral and back space has not been investigated. Performance in spatial bisection and MAA tasks was assessed for normally sighted blindfolded subjects using bursts of white noise presented frontally, laterally, or from the back relative to the subjects. Thresholds for both tasks were similar in frontal space, lower for the MAA task than for the bisection task in back space, and higher for the MAA task in lateral space. Two interpretations of the results are discussed, one in terms of visual calibration and the use of internal representations of source location and the other based on comparison of the magnitude or direction of change of the available binaural cues. That bisection thresholds were increased in back space relative to front space, where visual calibration information is unavailable, suggests that an internal representation of source location was used for the bisection task.

Figure 1

Schematic of the layout for the bisection task (a) and MAA task (b), with examples of possible reference (shown by grey loudspeakers) and probe (shown by black loudspeakers) locations. For both tasks, the subject was oriented so that sounds were presented from in front, laterally, or from the back. For illustrative purposes, all angles have been expanded relative to their true values.

Figure 2

Example of a psychometric function from a typical subject for data collected in the bisection task. The standard deviation (σ) of the fit, which is proportional to the reciprocal of the slope of the psychometric function, was taken as the estimate of threshold/precision. The midpoint of the function was taken as the PSE, and the distance of the PSE from the physical center point was taken as the bias. The curve is a fitted cumulative Gaussian.

Figure 3

MAA and bisection (BIS) thresholds (left panel) and bias values (right panel). The thresholds/biases are shown for front (circle), back (rhombus) and lateral (triangle) positions for the bisection (BIS, violet symbols) and MAA (blue symbols) tasks. Large symbols represent means and small symbols represent data for single subjects. Brackets indicate significant differences at p < 0.05.

Figure 4

Illustration of how subjects could compare the magnitudes of binaural cues to perform the bisection task for frontal space (top) and lateral space (bottom). The lengths of the red open bars show schematically the magnitudes of the interaural cues. Cues favoring the left ear are plotted below the horizontal lines and cues favoring the right ear are plotted above the lines.