Login / Signup

Tipping the Scales: Indiscriminate Use of Interval Scales to Rate Diverse Dysarthric Features.

Kaila L StipancicBrooke-Mai WhelanLauren LaurYunxin ZhaoAndrea RohlInyong ChoiMili Kuruvilla-Dugdale
Published in: Journal of speech, language, and hearing research : JSLHR (2024)
Overall, results of this work suggest that DME is the best fit for scaling several hypokinetic dysarthria features, and not the conventionally used EAI scale. Prothetic dimensions best scaled by DME include overall speech severity, articulatory imprecision, reduced loudness, and short rushes of speech. Monotony was the only feature found to be a metathetic dimension and would be best scaled using EAI or DME. Findings call for rethinking the widespread use of EAI scales for rating perceptual features as part of the assessment and treatment of motor speech disorders.
Keyphrases
  • machine learning
  • hearing loss
  • working memory
  • deep learning