Objective Analysis of Pathological Speech

The primary method for diagnosing speech pathologies has traditionally relied on perceptual analysis. However, recent advancements in speech AI provide valuable tools to support clinicians in their decision-making processes. Since my graduate studies, I have focused on analyzing speech features associated with motor speech disorders, particularly dysarthric speech. I am continuing this work at Carnegie Mellon University advised by Prof. Christina Bjorndahl in the Philosophy Department, with a special emphasis on the acoustic properties of fricatives. While these focuses on adult dysarthric speakers, I also work on analysis of child dysarthric speech, with my friends Jiyoung Choi at Communication Sciences and Disorders at Teachers College, Columbia University, and Blake Vente at Audible.

In addition, I am expanding my research into cognitive disorders, particularly mild cognitive impairment, in collaboration with Prof. Sunghye Cho at the University of Pennsylvania’s Linguistics Department and Prof. Min Seok Baek at Yonsei University’s Wonju College of Medicine.

Relevant publications include (Yeo et al., 2021), (Yeo et al., 2021), (Yeo* et al., 2023), and (Yeo* et al., 2023), which focus on automatic intelligibility assessments of Korean dysarthric speech. Furthermore, in (Lee et al., 2023), a protocol originally proposed for dysarthric speech was adapted to explore its applicability in diagnosing autistic children by analyzing their speech characteristics.