The effect of experience on environmental sound identification

Abstract

The perception of environmental stimuli was compared across normal hearing (NH) listeners exposed to an eight-channel sinewave vocoder and experienced bilateral, unilateral, and bimodal cochlear implant (CI) users. Three groups of NH listeners underwent no training (control), one day of training with environmental stimuli (exposure), or four days of training with a variety of speech and environmental stimuli (experimental). A significant effect of training was observed. The experimental group performed significantly better than exposure or control groups, equal to bilateral CI users, but worse than bimodal users. Participants were divided into low, medium and high-performing groups using a two-step cluster algorithm. High-performing members were only observed for the CI and experimental conditions, and significantly more low-performing members were observed for exposure and control conditions, demonstrating the effectiveness of training. A detailed item-analysis revealed that the most accurately identified sounds were often temporal in nature or contained iconic repeating patterns (e.g., a horse galloping). Easily identified stimuli were common across all groups, with experimental subjects identifying more short or spectrally driven stimuli, and CI users identifying more animal vocalizations. These data demonstrate that explicit training in identifying environmental stimuli improves sound perception, and could be beneficial for new CI users.