Abstract
Recent studies suggest that automatic speech analysis is promising for diagnosing movement and cognitive disorders. However, the reliability of acoustic and linguistic measures receives insufficient research efforts. The present study examines the test-retest reliability of commonly used acoustic and linguistic features in a healthy Chinese adult sample. Forty healthy young adults participated in the study and received seven frequently used speech tests twice, separated by 2–3 days. Fifty-six acoustic and linguistic features were extracted for each participant. The test-retest reliabilities of those features were then estimated using the intra-class correlation (ICC) method. The frequency-related, spectral, diadochokinetic (diadochokinetic rate, voice onset time), and content length features showed acceptable absolute agreement between the two test sessions. Articulation tasks (sustained vowel pronunciation, diadochokinetic rate) provided more reliable features than cognitive tasks such as picture descriptions. However, approximately half of those features failed to reach a moderate reliability requirement. There were no gender differences in the reliability estimates. We suggest increasing trial numbers or using multiple tests to increase the reliability of automated speech tests.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
