Pediatric Speech Biomarker Detection

Research into digital biomarkers is an area of rapid growth in digital medicine.  Speech has emerged as a powerful biomarker that has been used to identify diseases where some of the early signs involve a decrease in speech function.  Speech biomarkers allow for more ecologically valid measurement beyond current clinical tools. Current methods for speech measurement rely heavily on human speech perception which is subject to problems of interrater reliability and can be time as well as labor-intensive. The use of speech biomarkers offers a promising solution for improving access to early detection of speech-related disorders.  However, there is a need for rigorous evaluation of the potential, especially in the pediatric population.   In our lab, we are investigating how to leverage the acoustic signal to identify features that can be used as speech biomarkers that aid in the early detection of speech disorders in children.

Suthar, K., Yousefi Zowj, F., Speights Atkins, M., He, Q. P. (in press). Feature engineering and machine learning for computer-assisted screening of children with speech disorders. PLOS Digital Health.

Speights Atkins, M., Boyce, S, MacAuslan, J., Silbert, N. (2019). Computer-assisted syllable complexity analysis of continuous speech as a measure of child speech disorder. In Sasha Calhoun, Paola Escudero, Marija Tabain & Paul Warren (eds.) Proceedings of the 19th International Congress of Phonetic Sciences, Melbourne, Australia 2019 (pp.1103-1107 ) Canberra, Australia: Australian Speech Science and Technology Association. https://assta.org/proceedings/ICPhS2019Microsite/pdf/full-paper_465.pdf