Position:  PhD Candidate

Current Institution:  University of Washington

Abstract:  Physiological Computing at Scale

Today’s smartphones have multiple wireless sensors, such as Wi-Fi and acoustic sensors, and processing power equal to that of a computer. These wireless sensors can be leveraged to enable novel applications, specifically in the field of medical diagnosis. Further, the ubiquity of these devices enables deployment at a larger scale, creating an impact in the quality of life. To this end, I developed and deployed two applications:  1) ApneaApp to diagnose sleep apnea, and 2) a system tht detects opioid overdose. Sleep apnea is a common medical disorder that occurs when breathing is disrupted during sleep. Today, diagnosing sleep apnea requires the polysomnography test which is an expensive and labor-intensive process. We designed ApneaApp which is the first contactless system that detects sleep apnea using just an off-the-shelf smartphone. It enables this by tracking the fine-grained breathing movements from multiple subjects without requiring any contact with the phone. We deployed the system with 37 real patients at Harborview sleep lab and demonstrated that ApneaApp can detect apnea events with a correlation >0.95 compared to Polysomnography. Our work was the Best Paper Nominee at Mobisys 2015 and was exclusively licensed by ResMed Inc. Fatal opioid overdose remains a public health epidemic in the United States. Each day, 115 Americans die from opioid overdose. Unlike many life-threatening medical emergencies, opioid toxicity is readily reversible. Thus, a fundamental challenge of fatal opioid overdose events is that victims die alone with no or insufficiently timely diagnosis and treatment. To help connect potential overdose victims with life-saving interventions, we developed algorithms for smartphones that unobtrusively recognize opioid overdose by its physiologic precursors. We deployed our system in an approved supervised injection facility (SIF) where users self-inject illicit opioids. Our system had 96 percent sensitivity and 97 percent specificity (n=206) for identifying opioid-induced central apnea, a key event commonly preceding fatal opioid overdose.

Bio:
Rajalakshmi Nandakumar is a PhD candidate in the Paul G. Allen school of Computer Science and Engineering at University of Washington.  She works in the Networks and Mobile Systems lab, advised by Professor Shyamnath Gollakota.  Her research focuses on leveraging the ubiquity of smart devices with wireless sensors to enable and deploy novel applications that can improve quality of life.  her research ‘ApneaApp’ was licensed by ResMed, a major sleep health company. She received the UW CoMotion Graduate student innovator for her work on ApneaApp.  Previously, she was a research assistant in the Mobility Networks and Systems lab at Microsoft Research India.