Acoustic and Physiological Outcome Measures for Laryngeal Dystonia - 1 Project Summary/Abstract 2 Adductor laryngeal dystonia (AdLD) is a neurological voice disorder causing involuntary laryngeal spasms of 3 the larynx, leading to communication impairments. AdLD is diagnosed based on perceptual features, which 4 lack reliability and can present similarly to primary muscle tension dysphonia (pMTD). However, treatment for 5 the two disorders is vastly different. Misdiagnoses of AdLD delay treatment by 5–6 years, imposing financial 6 and quality-of-life burdens. No automated, clinically feasible objective diagnostic measures exist for 7 distinguishing the two disorders. Automated estimates of creak, defined as irregularly spaced vocal pulses 8 often perceived as vocal fry, show promise as a diagnostic marker for AdLD. Preliminary studies demonstrated 9 automated creak detection effectively differentiates AdLD from MTD with high diagnostic accuracy (AUC = 10 0.86). Understanding creak's physiological origins is critical. In typical speakers, creak occurs at the end of 11 breath phrases due to low subglottal pressure. In AdLD, creak occurs throughout the breath phrase, 12 suggesting distinct underlying mechanisms. Recent advances in accelerometer-based subglottal pressure 13 (ACC-Ps) estimation allow detailed physiological analysis during connected speech, making it feasible to study 14 in AdLD, pMTD, and controls. In the proposed study, a single experiment will record simultaneous speech, 15 ACC-Ps, and respiratory kinematics in speakers with AdLD, MTD, and controls reading aloud. An automated 16 creak detector will analyze speech data. Our first aim is to investigate the underlying physiology of creak in 17 speakers with AdLD by measuring subglottal pressure in connected speech. Our second aim will evaluate 18 discriminative accuracy of creak that occurs early in the breath group phrase, called early-phrase creak, 19 between speakers with AdLD, speakers with MTD, and controls. We will benchmark the discriminative 20 performance of early-phrase creak against standard clinical measures used to measure voice. The dataset 21 collected in this project will serve as pilot data for an R01 application to develop creak-based diagnostic tools, 22 which may improve differential diagnosis of AdLD and MTD, thus mitigating the current delay in diagnosis and 23 treatment of this debilitating voice disorder.
