Improving jitter and shimmer measurements in normal voices

Abstract

Instrumental acoustic voice analysis is a widely used clinical assessment technique to assist differential diagnosis, documentation and evaluation of treatment for voice disorders. However recent reports criticise an unsatisfactory reliability and validity of acoustic assessments. The present work examines confounding factors associated with the usual clinical measurement procedure and how their influence might be reduced. Further, it was investigated what jitter and shimmer indicate, and how this could be applied in voice clinics. In a routine clinical voice assessment the individuals` speaking voice SPL and F0, gender and the vowel significantly influence both jitter and shimmer. Differences in habitual voice SPL have by far the strongest influence, which may even underlie gender effects. It was shown for the first time, that clinical jitter and shimmer measurements might be considerably improved when patients phonate at a predefined level of 85dBA (10cm distance) without control of F0, and always use the vowel /a/. In healthy adults jitter and shimmer were not associated with perceptual voice irregularity. However clinical measurements in a variety of voice tasks showed that jitter and shimmer were always lower in higher voice intensities. Also in vocally healthy teachers increased voice SPL and F0 after a working day were associated with lower jitter and shimmer. Higher voice SPL is accompanied by increased vocal fold tone, which might result in more regular voice vibration patterns and thereby in lower jitter and shimmer values. From a clinical perspective this would be highly relevant information, especially in patients with impaired vocal fold tone regulation such as in functional or neurogenic voice disorders. Future research should clarify sources of jitter and shimmer and revise current normative values considering the proposed assessment protocol and gender. This might establish the clinical potential of instrumental acoustic voice analysis.