Computational Methods for Physiological Data

Large volumes of physiological data are now collected as part of routine patient care and in clinical trials. These datasets present an exciting opportunity to advance patient care by capturing prognostic phenomena associated with specific medical conditions, and by providing fresh insights into disease dynamics over long time scales.

In this talk, I will describe how progress in medicine can be accelerated through the use of sophisticated machine learning and signal processing methods for the structured analysis of large multi-patient, multi-signal datasets. In particular, I will describe two new approaches physiological symbolic analysis and morphologic variability that we have recently proposed. When evaluated on cardiovascular data from nearly 5,000 patients, these approaches were better predictors of cardiovascular and sudden cardiac death than other automated ECG-based metrics. Furthermore, these results were independent of information in echocardiography, clinical characteristics, and biomarkers.