Reliability and Diagnostic Accuracy of Semi-Automated and Automated Acoustic Quantification of Vocal Tremor Characteristics.
Youri Maryn, Kaitlyn Dwenger, Sidney Kaufmann and 1 others
PMID 40324153WHAT IT FOUND
Semi-automated acoustic analysis matched manual vocal tremor measurements closely, with correlations above .9 across all properties.
Fully automated analysis varied widely, reaching only .5 for pitch rate. Diagnostic accuracy was acceptable for both automated and semi-automated extent measures, but thresholds require confirmation in larger samples.
Key findings
01The semi-automated method showed strong correlations with manual measurements across all four tremor properties (pitch rate, pitch extent, intensity rate, intensity extent), whereas the fully automated method ranged from moderate to strong.
02Diagnostic accuracy for detecting vocal tremor was acceptable for all automated and semi-automated measures of modulation extent, with area under the curve values between 0.888 and 0.920.
03Fully automated analysis produced errors in cases with varying modulation rates, unvoiced intervals, or voice breaks, which the semi-automated method corrected by allowing user intervention.
STILL TO COME
How it was doneWhat they foundWhat it means for SLPs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study included only 15 participants, which limits the generalizability of the findings to the broader population of people with vocal tremor. Only two raters performed the manual measurements, and only one user applied the semi-automated tool, so the reliability of these methods across different clinicians is not yet established. The study did not include a control group of healthy individuals or patients with other voice disorders, so it is unclear how well these acoustic thresholds distinguish essential vocal tremor from other conditions. The comparison of intensity modulation extent was confounded by different units of measurement (Pascals for manual, decibels for automated), leading to large numerical differences that may not reflect true measurement disagreement.
Declared interests
The authors declared no conflicts of interest. The work was funded by the National Institute on Deafness and Other Communication Disorders.
The easy way to misread this
Do not interpret the high correlation of the semi-automated method as evidence that fully automated acoustic analysis is ready for clinical use. The fully automated method showed moderate correlation for some measures and produced significant errors in complex signals. The semi-automated method requires a trained user to identify and correct errors, meaning it is not a 'push-button' solution.