Description
Speaker: Matías Siñartu, Professor at Universidad Técnica Federico Santa María (UTFSM), Chile
Title: Multimodal Sensing and Physiological Modeling for Biomedical Signal Analysis: Research Capabilities and Opportunities at AC3E
Teams link: https://teams.microsoft.com/meet/3719760675855719?p=OOYonRdTDrEf0DGNc1
Abstract: Advances in biomedical sensing, signal processing, machine learning, and computational modeling are creating new opportunities to characterize complex physiological systems through multimodal measurements. This presentation will provide an overview of research capabilities and ongoing developments at the Advanced Center for Electrical and Electronic Engineering (AC3E), with particular emphasis on biomedical signal analysis, physiological modeling, and the integration of heterogeneous data sources. Selected examples from the Biomedical Technologies Impact Area will illustrate interdisciplinary research in wearable sensing, physiological monitoring, neural engineering, and intelligent healthcare technologies. A central focus will be placed on research in human voice production, where multimodal experimental platforms combine high speed videoendoscopy, electromyography, aerodynamic measurements, acoustic recordings, and ambulatory neck surface accelerometry. These complementary signals provide different perspectives on the biomechanics, neural control, and aerodynamic processes underlying phonation. Their integration with computational models offers opportunities to investigate physiological mechanisms, characterize individual variability, and develop objective measures of vocal function. The presentation will discuss how physics based modeling, signal processing, and data driven inference can be combined to extract interpretable physiological information from these measurements. Particular challenges include multimodal data fusion, temporal synchronization, dimensionality reduction, identification of latent physiological variables, and the analysis of intersubject variability. These problems also motivate potential applications of structured tensor decompositions, coupled representations, and other approaches to multidimensional signal analysis. Through selected research examples, the presentation will illustrate how multimodal physiological measurements and computational modeling can support the development of digital biomarkers, ambulatory monitoring technologies, and personalized assessment tools. Opportunities for methodological collaboration in multidimensional biomedical signal processing will also be discussed.
