The Prevalence of Creak Across Breath Groups in Adductor Laryngeal Dystonia

Creak is an acoustic feature found to discriminate speakers with adductor laryngeal dystonia (AdLD) from typical speakers with outstanding diagnostic accuracy. Yet creak is also used by typical speakers as a phrase-boundary marker. This study aims to compare the prevalence of creak across estimated breath groups in speakers with AdLD and controls to delineate physiological mechanisms underlying creak in AdLD. Thirty-four speakers read aloud the first paragraph of the Rainbow Passage (17 diagnosed with AdLD and 17 with no history of voice complaints). “Breath-like” pauses were defined as any in which technicians audibly heard a pause and all pauses >500 ms. For each phoneme, the time preceding the next breath-like pause was calculated, and the probability of creak occurrence was calculated. A generalized linear mixed-effects model was performed to determine the relationship between creak and time preceding a breath-like pause. Inter-rater and intrarater reliability of technicians were excellent. There was a statistically greater probability of creak in the AdLD group compared to controls (22% vs. 5%) and a statistically greater probability of creak as speakers approached a breath-like pause in both groups. The interaction between the time preceding a breath-like pause and group was significant, with a stronger relationship between the time preceding a breath-like pause and creak for control speakers (P