Effect of musical experience on learning to understand vocoded speech

Abstract

This study explores whether musicians have an advantage over non-musicians in processing and comprehending 8-channel vocoded speech, spectrally degraded speech that imitates cochlear implant output (see Loebach, Bent, & Pisoni, 2008). Musicians and non-musicians completed a pre-test in which they were asked to transcribe a number of vocoded sentences and words. Afterwards, they completed training on either vocoded or natural stimuli. After training, participants completed post-tests during which they transcribed vocoded speech, including items from the original test and from new speakers and 4-channel vocoded stimuli. In preliminary studies, musicians showed no advantage over non-musicians, and contrary to expectations, participants in natural-speech training condition improved more from pre- to post-test vocoded word recognition than participants in the vocoded training condition. Current studies are examining the relation of rhythmic and melodic perception (Musical Ear Test, Wallentin et al., 2010) to performance on the vocoded speech learning; the vocoded and natural speech training tasks are also compared to a control training task (testing sine-wave tone detection). Results of these tests will be used to evaluate the extent to which musical ability and training may be related to perception of degraded speech.