Biomedical Monitoring, Signal Processing and Control

点击翻译按钮获取中文摘要

Although automation technology has become ubiquitous in all of modern engineered systems and devices, see for instance current developments in driverless cars, it has yet to play an essential role in healthcare. Among the many reasons for the current lack of impact of control technology in that sector are the challenges posed by i) the significant variability among patients, e.g. in the way they respond to a given drug; ii) the difficulty in developing non-invasive, reliable and easy-to-use sensors; iii) and the safety concerns when using an automated, somewhat autonomous that e.g. administers potentially dangerous drugs. However, the recent advent of closed-loop implantable cardiac defibrillators and closed-loop control of blood glucose for type-1 diabetes, such as the FDA-approved Medtronic MiniMED 670G system that adjusts basal insulin every 5 min, an era of automated, personalized healthcare is about to usher. For the last 17 years, I have been at the forefront of such technology, as my work has focused on providing clinicians tools based on signal processing and control to help improve patient outcomes and safety not only in critical care settings but also in global health settings. This proposal is a continuation of this work with a strong focus on sensing, signal processing and control techniques for safe automatic drug delivery for anesthesia and sedation. Within this program I will: 1. Develop robust and safe control algorithms for safe automated drug delivery for anesthesia and sedation using a number of innovative control methodologies 2. Develop noninvasive sensing techniques, e.g. photoplethysmography (PPG) for the cardiovascular system; more performant electroencephalography (EEG) technology capable of detecting very low frequency, large amplitude activity associated with migraines with aura and certain types of epilepsy; near-infrared spectroscopy (NIRS) for brain and tissue monitoring 3. Develop efficient signal processing techniques to extract clinically-relevant features in physiological parameters, e.g. the PPG and EEG As a result of this work, automated systems for anesthesia and sedation should become standard of care, resulting in reduced adverse events in operating rooms and intensive care units and better long term outcomes for patients such as reduction in post-operative delirium and cognitive impairment. Novel EEG technology could result in i) better understanding of migraines, potentially leading to better prevention and treatment ii) better way to predict and localize the focus of epileptic seizures. Combining EEG with NIRS will lead to a better understanding of pain perception and better ways to avoid and manage it in preterm infants during their stay in the neonatal intensive care unit. Finally, work on signal processing should lead to innovative ways to extract clinically relevant information from data obtained from wearables.