Measuring the Dysphonia Severity Index (DSI) in the Program Praat

Abstract

Background: The original Dysphonia Severity Index (ie, DSI) weighs and combines four voice markers in a single number to size dysphonia gradation in the clinic: percent jitter (from Multi-Dimensional Voice Program; KayPENTAX Corp., Montvale, NJ), softest intensity and highest fundamental frequency (both from Voice Range Profile; KayPENTAX Corp.), and maximum phonation time. To be more generally applied, however, implementation of DSI in the program Praat (Paul Boersma and David Weenink, Institute for Phonetic Sciences, University of Amsterdam, The Netherlands) would be advantageous for all voice clinicians. The presented project was therefore designed (1) to develop such a Praat application and (b) to test its concurrent validity.

Methods: The four voice markers for the original DSI, as well as ten additional voice markers in Praat, were administered on a total of 49 subjects in three different clinical voice centers. A crossover research design was implemented to counterbalance for possible exercise effects. First, stepwise multiple linear regression was applied to build a statistical model with the best combination of Praat predictors for the original DSI. Second, correlation statistics were applied to compare Praat's DSI with the original DSI.

Results: Both DSI versions correlated strongly. A custom script was therefore written for automated DSI determination in Praat.

Conclusion: With this script, every voice clinician can easily determine DSI in the Praat program.

Keywords: Dysphonia Severity Index; Multi-Dimensional Voice Program; Praat; Variability; Voice Range Profile.