Invited Talks

Invited Talks

Dynamical Assays: A New Frontier in Precision Medicine?

Ary L. Goldberger, MD

Madalena Damásio da Costa, PhD

Margret and H.A Rey Institute for Nonlinear Dynamics in Medicine

Beth Israel Deaconess Medical Center/Harvard Medical School

Boston, MA

This presentation, from the perspectives of a physician and physicist, is intended to: 1) give a brief overview of complex (nonlinear) systems in biomedicine; 2) discuss the differences between variability and complexity, and 3) provide practical examples of new ways of visualizing and quantifying the dynamics of living systems in health and disease.

Healthy systems display dynamical fluctuations that are among the most complex in nature. In contrast, advanced aging (frailty syndrome) and disease are associated with the breakdown of system complexity, which can be quantified by an ensemble of multiscale measures, including those related to entropy, fractality and time irreversibility. For example, we recently reported a loss of complexity in glucose concentration dynamics in patients with type 2 diabetes compared with control subjects. The loss of complexity could not be quantified by changes in conventional metrics of glycemic control.

These and related findings raise the possibility of using “hidden” information in glucose time series (“dynamical glucometry”) to help refine personalized therapy, as well as promote drug discovery and prevent pharmacologic toxicity in this highly prevalent disease. More generally, diagnostic assessments and therapeutic interventions across a broad array of biomedical contexts may be enhanced by probing dynamical behaviors.

Future wearables and surprising new ways they can impact our lives

Prof. R W Picard, MIT Media Lab

In the mid-90's, I led the creation of some of the first smart wearable computers (back then we considered under 50 lbs to be "wearable"). Our initial goal was to recognize human emotion from objective data measured from our bodies, and use it to make the technology smarter. We created new algorithms for signal processing and used machine learning to combine information from physiology, faces, voices, posture, and gesture. In this talk I will highlight some of the more surprising discoveries enabled recently as these wearable sensing technologies have evolved. These include how electrical signals on the surface of the wrist can communicate emotional excitement, stress and deep brain activity, and how everyday wearable motion sensors can be mined for information communicating your heart-rate, respiration, and identity. I'll present ways in which wearables can affect our social-emotional and personal connections, with surprising implications drawn from autism, sleep, epilepsy, and learning.

Building a [machine] learning healthcare system

Nigam H. Shah, MBBS, PhD

Associate Professor of Medicine (Biomedical Informatics Research) at Stanford Univeristy

In the era of Electronic Health Records, it is possible to examine the outcomes of decisions made by doctors during clinical practice to identify patterns of care — generating evidence from the collective experience of patients. We will discuss methods that transform unstructured EHR data into a de-identified, temporally ordered, patient-feature matrix. We will review use-cases, which use the resulting de-identified data, to discover hidden trends, build predictive models, and drive comparative effectiveness studies in a learning health system.

Computational Subtyping: Is Lupus one disease or many disease subtypes?

Suchi Saria

Assistant Professor of Computer Science, Health Policy and Statistics (joint) at Johns Hopkins University

Rapid adoption of electronic health records has opened up the possibility for contextualizing decision making in healthcare at a much higher resolution than was possible before. For many chronic diseases, patients show highly variable disease course. Treatment guidelines derived from clinical trials are often population based and not sufficiently tailored to the needs of the individual. We will discuss new approaches for subtyping disease using the framework of disease trajectories. These newly defined subtypes provide clinicians a novel approach for targeting therapy. I will discuss example applications in autoimmune diseases.