gms | German Medical Science

Instrumental acoustic analysis is widely used to assess voice disorders and to describe intervention success [3], [4]. In indicating pitch (jitter) and amplitude (shimmer) perturbation, it has been suggested that jitter and shimmer objectively measure different aspects of perceptual dysphonia [5], [7], [8]. However previous reports disagree on the exact nature of the relation and have associated perceptual “hoarseness” with increased jitter [7] and/or shimmer [5]. Further other authors contradict this and suggest that there is no relation between instrumental acoustic and perceptual assessments [1], [2]. Therefore this study investigates if jitter and shimmer relate to the severity of perceptual hoarseness in voice patients.

In a cross-sectional single cohort study thirty-three patients with head and neck cancer were recorded before and after radiotherapy while saying a prolonged /i/ for 6 seconds. 61 recordings were assessed by instrumental and perceptual analysis. The instrumental acoustic parameters jitter (%) and shimmer (dB) were determined with PRAAT [9]. Perceptual voice analysis was done by 2 voice experts using the Grading-Roughness-Breathiness-Asthenia-Strain scale (GRBAS) [6]. Thereafter the mean overall impression of hoarseness (as indicated by “G“) of the 2 voice experts was correlated with jitter and shimmer.

Mean jitter (j) and shimmer (s) were lowest for voices rated GRBAS G0 [j0=0.47(0.36); s0=0.27(0.09)]. These patients define the jitter and shimmer range we would expect in normal sounding voices. Both mean jitter and shimmer were higher in the voices rated G1 [j1=0.65(0.54); s1=0.39(0.17)] and G2 [j2=0.92(0.78); s2=0.42(0.25)]. However, as indicated by Figure 1 [Fig. 1], this was due to several patients with extremely high voice perturbation. In the present study the majority of pathological sounding voices had jitter and shimmer within the range of the normal voices (G0). Also dysphonic voices rated G1 and G2 could not be distinguished on the basis of jitter and shimmer. Consequently, for all voices together a moderate correlation between shimmer and G was observed (r=0.25; p=0.05). This was similar for jitter and G in all voices (r=0.27; p=0.04). However in dysphonic voices only (G≥1) there was no clear correlation with increasing shimmer (r=0.03; p=0.9) or jitter (r=0.17; p=0.3).

Among pathologic voices, as measured by G of the GRBAS scale, there was no clear association of jitter or shimmer with dysphonia severity. Further it was not possible to reliably distinguish normal from pathologic sounding voices by jitter or shimmer. Therefore we conclude that neither jitter nor shimmer are an objective index of perceptual hoarseness. This might be explained by the low reliability of instrumental acoustic measurements in more irregular voice signals [2]. Nevertheless jitter and shimmer might provide clinically valuable information in patients with normal sounding voices. Future research should establish what jitter and shimmer might indicate in these patients and describe how this could be applied in voice clinics.