Prof. Walden Co-Authors Research to Pinpoint Disorder-Specific Voice Markers, Advancing Precision Assessment

Patrick R. Walden, Ph.D., CCC-SLP, associate professor, chair, and program director of the Department of Speech-Language Pathology at Monmouth University, is the co-author of a new study in "Journal of Voice" that identifies disorder-specific acoustic-perceptual patterns to improve diagnosis and treatment planning for voice disorders. The article, "From Screening to Precision: Searching for Voice Disorder-Specific Acoustic and Auditory-Perceptual Metrics," appears online ahead of print.

Along with co-authors Eric J. Hunter, Ph.D., chair of the Department of Communication Sciences and Disorders at the University of Iowa, and Lady Catherine Cantor Cutiva, Ph.D., SLP, MSc, assistant professor in the Department of Audiology and Speech-Language Pathology at East Tennessee State University, the team analyzed 14 acoustic measures and paired them with expert auditory-perceptual ratings from Walden's perceptual voice qualities database to test how well specific patterns distinguish common conditions-including vocal fold paralysis, atrophy, phonotraumatic lesions, and muscle tension dysphonia-across both sustained vowels and connected speech.

The analysis used generalized linear models, principal component analysis, and ROC curves; results showed strong discrimination for vocal fold paralysis (AUC ≥ 0.75) and consistent patterns across speech tasks, supporting flexible, evidence-based clinical protocols and precision-focused assessment.

"This work moves clinicians closer to precision assessment, matching what we hear and measure to the disorder in front of us," Walden said. "By integrating acoustic features with expert ratings, we can design task-flexible evaluations that serve patients better."

The study drew on the perceptual voice qualities database, a standardized collection of CAPE-V vowels and sentences that Walden developed with support from The Voice Foundation and used internationally in education and research.

The findings offer a practical path from broad screening toward disorder-specific diagnosis, provide validated, task-robust markers that clinicians can apply in either vowels or sentences, and supply a foundation for future machine-learning tools that combine what clinicians hear with what algorithms detect-without locking protocols to one speaking task.