음성장애 선별검사 도구로서 스마트폰을 이용한 음성 녹음의 임상적 유용성

Abstract

Objectives: The purpose of the present study was to determine whether selected acoustic measures and estimates differ according to the recording devices, the Computerized Speech Lab (CSL) and a smartphone, and groups. In addition, correlation and reliability between the devices were explored. Lastly, the diagnostic ability of the two devices were compared. Methods: A total of 180 participants (90 patients and 90 controls) participated in the study. Vowel samples of the participants were recorded via the CSL and an android smartphone. For each sample, acoustic parameters including jitter percent (Jitt), shimmer percent (Shim), noise-to-harmonic ratio (NHR), cepstral peak prominence (CPP), L/H ratio, Cepstral Spectral Index of Dysphonia (CSID) were measured. Two-way repeated measures of ANOVA, Pearson correlation analysis, and ROC curve analysis were performed. Results: Interaction effect for Jitt, Shim, NHR, and σCPP between groups and devices were found. F0 measures of the smartphone were higher than those of the CSL. CPP and SR of the CSL and patients were higher than those of the smartphone and controls, respectively. In contrast, CSID and σSR were higher for the smartphone. All the parameters showed positive correlation between devices. There was no difference in the area under curve between the devices, although substantially different cut-off scores were obtained. Conclusion: The current data showed a significant correlation of acoustic measures and no difference in the diagnostic ability between the devices, although differences in several measures and higher cut-off scores of the smartphone were noted. In conclusion, smartphones can be used as a screening tool for voice disorders.