Development and evaluation of the listening in spatialized noise test

The goal of this study was to design and develop an audiological test that provides an ecologically valid measure of speech understanding in background noise while minimizing the effects of between-listener variation in factors such as linguistic skills and attention on test performance. The Listening in Spatialized Noise Test (LISN) creates a three-dimensional auditory environment under headphones and was designed to be totally software driven, so that it can be delivered in any audiology clinic with the use of only a PC and an audiometer. The extent to which the LISN was able to simulate free-field conditions and the effect of learning on the test were also examined.

: In a three-alternative forced choice adaptive procedure, 20 adults with normal hearing were required to indicate the intelligibility level of target continuous discourse presented at 0 degrees azimuth in the presence of distracter sentences simultaneously presented at either 0 degrees azimuth (0 degrees condition) or at both +90 degrees and -90 degrees azimuth (+/-90 degrees condition). The target story was always spoken by female 1, whereas there were three conditions of speaker for the distracter sentences: the "same female speaker" as the target (same voice condition); two "different female speakers" (different female voices condition); and a "male speaker" (male voice condition). In a separate study, 16 adults with normal hearing who had not participated in the first study were assessed on the same voice and different female voices conditions, which were presented and then retested in the same order and test session to determine the effect of practice on performance on the LISN.

The 20 adults were able to understand the target story at a significantly lower threshold in the +/-90 degrees condition than the 0 degrees condition. The degree of this spatial separation advantage (SSA) decreased significantly as the vocal quality of speakers of the target and the distracter sentences became more different (10.4 dB in the same voice condition, compared with 5.6 dB in the different female voices condition, and only 3.3 dB in the male voice condition). The SSA for the different female voices and male voice conditions were comparable to measurements previously reported in a free-field environment. There was no significant difference in SSA between the first and second presentations for either the same voice condition (at 10.3 dB and 10.2 dB) or the different female voices condition (at 4.7 and 5.7 dB).

For adults with normal hearing, the ability to comprehend the story in the separate condition was facilitated by the use of binaural cues, such as interaural time differences, to distinguish the target from the spatially separated distracters. When a target and masker are distinguishable on the basis of features of the various speakers' voices (such as large differences in fundamental frequency), listeners are less reliant on spatial cues to recognize the target, and the SSA in dB is reduced. The stability of test scores with practice, the comparable levels of performance to those achieved in free-field environments, and the ability of the test to utilize difference scores to assess binaural processing while minimizing differences between participants in variables such as linguistic skills make the LISN a potentially valuable tool for assessing auditory processing disorders.