Introduction:
For quantitative voice diagnostics, vocal fold dynamics can be examined during the phonation of a sustained vowel using digital high-speed (HS) imaging. The captured vibrational dynamics of the vocal folds allow conclusions about underlying pathological phonation mechanisms. The computer-aided quantification of the complex vibrational dynamics can be realized by so-called phonovibrograms (PVG), which are obtained by segmentation of the HS-videos. Despite long term experience in clinical HS-imaging, there is still a lack of quantitative data about the individual extend of the measurement inaccuracy caused by repeated measurements.
Methods:
In this prospective longitudinal study, a cohort of 5 healthy male subjects (nonsmokers, age ≥18 years) was examined over a period of 3 weeks on 4 different days by means of HS-imaging recorded during sustained phonation. The vocal fold dynamics were quantified using a wavelet-based PVG analysis approach, whereby a total of 40,000 single images were segmented.
Results:
The evaluation of the repeated measurements revealed that the healthy subjects show a reproducible scattering of the parameter values, which represents the extent of temporal perturbation. The vibration modes, however, changed significantly between the repeated recordings.
Conclusions:
Even in presence of highly similar examination conditions the patterns of vocal fold vibrations differ intra-individually between the repeated measurements. For inter-individual definition of normal ranges, perturbation measurements therefore allow a more robust interpretation due to their higher reproducibility between different recordings.
