Computational Modelling

Background

Speech changes with altered brain function. Acoustic analysis of speech provides objective data on these changes. Acoustic features differ between healthy speakers and individuals with disease and evolve over the course of disease. On this basis, sophisticated speech biometrics can act as a proxy for brain integrity and may assist in optimising diagnostic pathways or identifying symptom onset in neurodegenerative diseases.

We use machine learning to identify neurodegenerative diseases across disease stages using acoustic features of speech. The ultimate goal is to create an open-source tool that can monitor disease progression and accurately identify disease onset.

Potential projects

This research project is available to PhD students to join as part of their thesis. Please contact the Theme Leader to discuss your options.

Collaborators

• Anthanasios Tsanas, The University of Edinburgh
• Prof. Howard Bondell, University of Melbourne
• Melbourne Data Analytics Platform
• Prof. Sonja Kotz, Maastricht University
• Prof. Duane Watson, Vanderbilt University
• Dr. Miriam Lense, Vanderbilt University

Seminal work

• Vogel, A.P., Tsanas, A., Scattoni, M.L. Quantifying Ultrasonic Mouse Vocalizations Using Acoustic Analysis in a Supervised Statistical Machine Learning Framework. Sci Rep. 2019;9(1):8100.
• Schultz, B. G., & Palmer, C. The roles of musical expertise and sensory feedback in beat keeping and joint action. Psychological research, 2019;83(3), 419-431.
• Schultz, B. G., & van Vugt, F. T. Tap Arduino: An Arduino microcontroller for low-latency auditory feedback in sensorimotor synchronization experiments. Behavior research methods, 2016; 48(4), 1591-1607.
• Schultz, B. G., O’Brien, I., Phillips, N., McFarland, D. H., Titone, D., & Palmer, C. Speech rates converge in scripted turn-taking conversations. Applied Psycholinguistics, 2016; 37(5), 1201-1220.