Toward improved differential diagnosis of adductor spasmodic dysphonia and muscle tension dysphonia.
ABSTRACT Adductor spasmodic dysphonia (ADSD) is characterized by a strained-strangled voice quality, whose diagnosis relies exclusively on auditory-perceptual features. However, muscle tension dysphonia (MTD) -- a functional voice disorder -- can mimic ADSD thereby contributing to diagnostic confusion. Unlike MTD, ADSD has been described as 'task-specific', implying that certain vocal tasks such as sentences loaded with predominantly voiced consonants will provoke greater sign expression. This investigation examined the diagnostic value of variable sign expression based upon phonetic loading as a means to disambiguate ADSD and MTD.
Five listeners, who were blinded to the purpose of the study, used a 10-cm visual analog scale to rate the dysphonia severity of two sentences (one all-voiced and one containing primarily voiceless consonants) produced by participants with ADSD (n = 29) or MTD (n = 33).
A mixed-design ANOVA, with Group (ADSD vs. MTD) as the between-subjects variable and Sentence Type (all-voiced vs. voiceless) as the within-subjects variable, confirmed a significant Group-by-Sentence Type interaction effect (p = 0.0002). In ADSD, ratings of dysphonia severity for the all-voiced sentence were significantly more severe than for the voiceless sentence (p < 0.0001), whereas in MTD no significant difference was observed (p = 0.9981). The ROC curve confirmed that differences in dysphonia severity between voiced and voiceless sentences represented a highly specific (88-100%), but only 48% sensitive diagnostic marker.
Phonetic loading influences sign expression in ADSD, and assists in discriminating ADSD from MTD.
[show abstract] [hide abstract]
ABSTRACT: In this article, we investigate whether (1) the IINFVo (Impression, Intelligibility, Noise, Fluency and Voicing) perceptual rating scale and (2) the AMPEX (Auditory Model Based Pitch Extractor) acoustical analysis are suitable for evaluating adductor spasmodic dysphonia (AdSD). Voice recordings of 12 patients were analysed. The inter-rater and intra-rater consistency showed highly significant correlations for the IINFVo rating scale, with the exception of the parameter Noise. AMPEX reliably analyses vowels (correlation between PUVF (percentage of frames with unreliable F0/voicing 0.748), running speech (correlation between PVF (percentage of voiced frames)/voicing 0.699) and syllables. Correlations between IINFVo and AMPEX range from 0.608 to 0.818, except for noise. This study indicates that IINFVo and AMPEX could be robust and complementary assessment tools for the evaluation of AdSD. Both the tools provide us with the valuable information about voice quality, stability of F0 (fundamental frequency) and specific dimensions controlling the transitions between voiced and unvoiced segments.Archives of Oto-Rhino-Laryngology 12/2009; 266(12):1915-22. · 1.29 Impact Factor