Michigan State University - Research Groups

Dr. Oweiss spoke on July 21^st, 2011 at EPFL (Ecole Polytechnique Federale de Lausanne) in Switzerland. His visit was hosted by Prof Martin Vetterli, Dean of computer and communication sciences. An abstract of his talk entitled  "Sparsity in the Brain" can be found below.

"Fundamental to understanding how our world is represented in our brain is the ability to observe the collective activity of ensembles of neurons acting in concert while we associate sensory stimuli with subsequent motor actions. Some recent technological advances have greatly accelerated our ability to simultaneously record and stimulate these ensembles, thereby opening up the possibility to advance our understanding of the brain and its inner workings, and to provide real time neural control of assistive devices to people with severe disabilitiesthrough sophisticated brain-machine interfaces (BMIs). These remarkable advances, however, have outstripped progress in statistical signal processing theory and techniques specifically tailored to: 1) analyze the massive amounts of neural and behavioral data collected; 2) explain many aspects of the natural signal and information processing mechanisms in the brain; and 3) perform real-time neural signal processing within the resource-constrained environment of an implantable system for clinically viable BMI applications.

In this talk, I will discuss a seemingly prominent role of sparsity in the brain to fulfill two objectives: 1) explain brain-evolved mechanisms of sparse coding to optimize information processing in the face of uncertainty; 2) engineer a wireless, fully implantable BMI system that extracts relevant information from neural signals in freely behaving subjects in real time. I will discuss how some guiding principles from compressive sensing, machine learning and systems neuroscience literature can help accomplish these objectives. I will conclude with a brief discussion on the potential of this framework to increase the reliability and effectiveness of BMIs, and to develop better algorithms for statistical inference in sparse, complex networks in the face of uncertainty and nonstationarity."