Exploring the Impact of Two Feedback Types on Speech Intelligibility, Precision, and Naturalness.

Abstract

This pilot study examined the impact of feedback type on learning a novel speech task, as measured by listener ratings, and will inform procedures for future investigations within a larger sample size. Twenty-four native monolingual English-speaking college-aged adults participated in a single training session to learn novel Hindi phrases. Participants were randomly placed into one of three feedback groups: knowledge of performance (KP), knowledge of results (KR), or a combined KP + KR condition. Participant performance was assessed at 1 day and 1 week post-training. Participant responses were audio recorded and judged for intelligibility, precision, and naturalness by native Hindi speakers, blind to the feedback conditions, via rating scales. At 2 days post-training, participants in the KP and KP + KR feedback conditions were rated as performing better than participants in the KR condition on all three perceptual measures. At 1 week post-training, participants in the KP feedback condition were judged to be superior across all three perceptual measures. Preliminary findings suggest that augmented feedback enhances learning, especially when skills are considered novel and learners are unable to rely on their own internal feedback. These results may have implications for the application of motor learning principles into clinical practice for persons with motor speech disorders.