Acoustic Hemifields in the Spatial Release from Masking of Speech by Noise

Abstract

The Hearing-in-Noise Test (HINT) is able to measure the benefit to speech intelligibility in noise conferred when the noise masker is displaced 90 degrees in eccentricity from a speech source located at zero degrees azimuth. Both psychoacoustic and neurophysiological data suggest that the perceptual benefit of the 90-degree azimuth separation would be greatest if the speech and noise were presented in different acoustic hemifields, and would be smallest if the two sources were in the same acoustic hemifield. The present study tested this hypothesis directly in ten normal-hearing adult listeners. Using the HINT stimuli, we confirmed the hypothesis. Release from masking scores averaged 8.61 dB for "between-hemifield" conditions, 6.05 dB for HINT conditions, and 1.27 dB for "within-hemifield" conditions, even though all stimulus configurations retained a 90-degree angular separation of speech and noise. These data indicate that absolute separation of speech and noise alone is insufficient to guarantee a significant release from masking, and they suggest that what matters is the location of the stimulus elements relative to the left and right spatial perceptual channels.

Psychophysical evidence for adaptation of central auditory processors for interaural differences in time and level

“…The outcomes of these experiments carry theoretical significance. There is evidence from human psychophysics (Boehnke and Phillips, 1999;Phillips et al, 2003) and animal neuroscience (Phillips and Irvine, 1981;Jenkins and Masterton, 1982;Jenkins and Merzenich, 1984;Phillips and Brugge, 1985;McAlpine et al, 2001;Stecker and Middlebrooks, 2003) that the processing of auditory azimuth is based on left and right spatial channels with broadly hemifield tuning that overlaps at midline azimuths. Accordingly, one model of sound localization mechanisms posits that each side of the brain (at least rostral to the superior olivary complex) is independently capable of localizing sounds in the contralateral hemifield.…”

The effects of lateralized adaptors on lateral position judgements of tones within and across frequency channels

“…Direct measurements of the spatial receptive fields of neurons in the brains of animals are the basis of this model (Middlebrooks and Pettigrew, 1981;Rajan et al, 1990;Clarey et al, 1995;Brugge et al, 1996;Stecker et al, 2005), although it has even earlier origins in studies of the coding of dichotic sound localization cues (Phillips and Irvine, 1981;Phillips and Brugge, 1985) and behavior-lesion studies (e.g., Jenkins and Masterton, 1982;Jenkins and Merzenich, 1984). It receives expression in human psychophysical studies using gap detection (Boehnke and Phillips, 1999), spatial release from masking (Phillips et al, 2003) and selective adaptation paradigms (Phillips and Hall, 2005;Phillips et al, 2006). In this model, sound localization processing for sources near the midline likely depends on the relative outputs of the two channels, since the channels have overlapping medial borders; for far-lateral azimuths, spatial processing is presumably mediated by mechanisms operating within the hemifield channel.…”

“…There are some exceptions in the form of cells with high preferred frequencies whose receptive fields are located on the contralateral pinna axis (Middlebrooks and Pettigrew, 1981) or near the midline (Rajan et al, 1990). Human free-field psychophysical studies exploiting temporal gap detection (Boehnke and Phillips, 1999) and spatial release from masking paradigms (Phillips et al, 2003) suggest the existence of hemifieldtuned perceptual channels for sound azimuth.…”

A perceptual architecture for sound lateralization in man

“…A subsequent experiment (Phillips et al, 2003) examined the spatial release from masking afforded by a 90°separa-tion of speech and noise, as a function of whether both speech and noise were in the same or different hemifields. The logic of the experiment was that if speech and noise fall in different perceptual channels, then they are differentially accessible to attentional and other cognitive processors.…”