Perceptual Learning of Altered Vowel Space Improves Identification of Vowels Produced by Individuals With Dysarthria Secondary to Amyotrophic Lateral Sclerosis

Purpose: This study examines the efficacy of perceptual training for improving typical listeners’ identification of vowels produced by individuals with dysarthria. We examined whether training on a subset of vowels can generalize to (a) untrained vowels and (b) other speakers with similar overall intelligibility.

Method: Sixty naive listeners completed a pretest/posttest perceptual learning task. In the pretraining test and posttraining test, participants identified nine American English monophthongs produced by two speakers with dysarthria secondary to amyotrophic lateral sclerosis (ALS). In the 20-min training task, a two-alternative forced choice (2AFC) task with feedback trained listeners on a subset of the vowels and speakers presented in the pretraining test.

Results: Vowel identification accuracy improved overall as a function of training. However, patterns of generalization between speakers and vowel types were not symmetric. Specifically, listeners generalized training from front vowels to back vowels but not vice versa. Likewise, listeners generalized from one speaker to another but not in the opposite direction. Examination of confusion matrices for the pretraining and posttraining revealed complex patterns of vowel-specific improvement.

Conclusions: This study demonstrates that listeners benefit from a very simple training paradigm targeting vowels. Additionally, error patterns revealed that vowels are both resistant to and responsive to perceptual learning. Implications for future research and clinical training paradigms are discussed.

Files

Metadata

Work Title Perceptual Learning of Altered Vowel Space Improves Identification of Vowels Produced by Individuals With Dysarthria Secondary to Amyotrophic Lateral Sclerosis
Access
Open Access
Creators
  1. Annie Olmstead
  2. Jimin Lee
  3. Janice Chen
License In Copyright (Rights Reserved)
Work Type Article
Publisher
  1. Journal of Speech, Language, and Hearing Research
Publication Date May 27, 2022
Publisher Identifier (DOI)
  1. https://doi.org/10.1044/2022_jslhr-21-00567
Deposited April 03, 2024

Versions

Analytics

Collections

This resource is currently not in any collection.

Work History

Version 1
published

  • Created
  • Added Olmstead_Lee_Chen_2022_Perceltual_learning_of_altered_vowel_space_improves_identification_of_vowels.pdf
  • Added Creator Anne J Olmstead
  • Added Creator Jimin Lee
  • Added Creator Janice Chen
  • Published
  • Updated
  • Updated Description, Publication Date Show Changes
    Description
    • <p>Purpose: This study examines the efficacy of perceptual training for improving typical listeners’ identification of vowels produced by individuals with dysarthria. We examined whether training on a subset of vowels can generalize to (a) untrained vowels and (b) other speakers with similar overall intelligibility. Method: Sixty naive listeners completed a pretest/posttest perceptual learning task. In the pretraining test and posttraining test, participants identified nine American English monophthongs produced by two speakers with dysarthria secondary to amyotrophic lateral sclerosis (ALS). In the 20-min training task, a two-alternative forced choice (2AFC) task with feedback trained listeners on a subset of the vowels and speakers presented in the pretraining test. Results: Vowel identification accuracy improved overall as a function of training. However, patterns of generalization between speakers and vowel types were not symmetric. Specifically, listeners generalized training from front vowels to back vowels but not vice versa. Likewise, listeners generalized from one speaker to another but not in the opposite direction. Examination of confusion matrices for the pretraining and posttraining revealed complex patterns of vowel-specific improvement. Conclusions: This study demonstrates that listeners benefit from a very simple training paradigm targeting vowels. Additionally, error patterns revealed that vowels are both resistant to and responsive to perceptual learning. Implications for future research and clinical training paradigms are discussed.</p>
    • <p>Purpose: This study examines the efficacy of perceptual training for improving typical listeners’ identification of vowels produced by individuals with dysarthria. We examined whether training on a subset of vowels can generalize to (a) untrained vowels and (b) other speakers with similar overall intelligibility.
    • Method: Sixty naive listeners completed a pretest/posttest perceptual learning task. In the pretraining test and posttraining test, participants identified nine American English monophthongs produced by two speakers with dysarthria secondary to amyotrophic lateral sclerosis (ALS). In the 20-min training task, a two-alternative forced choice (2AFC) task with feedback trained listeners on a subset of the vowels and speakers presented in the pretraining test.
    • Results: Vowel identification accuracy improved overall as a function of training. However, patterns of generalization between speakers and vowel types were not symmetric. Specifically, listeners generalized training from front vowels to back vowels but not vice versa. Likewise, listeners generalized from one speaker to another but not in the opposite direction. Examination of confusion matrices for the pretraining and posttraining revealed complex patterns of vowel-specific improvement.
    • Conclusions: This study demonstrates that listeners benefit from a very simple training paradigm targeting vowels. Additionally, error patterns revealed that vowels are both resistant to and responsive to perceptual learning. Implications for future research and clinical training paradigms are discussed.</p>
    Publication Date
    • 2022-06-01
    • 2022-05-27
  • Renamed Creator Annie Olmstead Show Changes
    • Anne J Olmstead
    • Annie Olmstead