Pitch Strength & Percent Vocal Creak to Assess Treatment Response in Adductor Laryngeal Dystonia.

Abstract

Quantifying treatment-related voice change in adductor laryngeal dystonia (AdLD) is essential for optimizing botulinum neurotoxin (BoNT) therapy, yet clinical practice lacks objective, automated metrics to track response. Counting voice disruptions remains the standard but is labor-intensive and prone to error. This study evaluated whether pitch strength (PS) and percent vocal creak (%VC) (automated indices of pitch salience and irregular phonation) are sensitive to BoNT-related changes. We examined whether these metrics capture treatment effects at the group level, and through individual correlations with manual/perceptual severity. Standardized speech from 16 patients was analyzed for PS, %VC, and two multi-parametric indices before and 1 month after BoNT. Mixed-effects models assessed BoNT and task effects. Spearman correlations examined relationships among acoustic, manual, perceptual, and patient-reported outcomes. Discontinuity counts, patient-reported outcomes, and perceptual ratings improved post-BoNT. Changes in PS (r = -0.83) and %VC (r = 0.82) strongly correlated with change in disruption frequency. No main BoNT effects or BoNT-by-task interactions were observed for PS or %VC. Acoustic Voice Quality Index increased with a small-medium effect (d = 0.48, p = 0.03). While PS and %VC correlate with changes in disruption frequency and perceptual severity, they did not demonstrate robust group-level treatment sensitivity. Because diverse acoustic manifestations of AdLD disruptions (e.g., F0 shifts vs. aperiodicity) can obscure improvement in summary metrics, these measures should be interpreted cautiously at the group level. At the individual level, PS and %VC may offer supplemental value for longitudinal monitoring but are not substitutes for manual disruption counts.