Colloquium Details
Neurobehavioral Engineering: A new computational frontier
| Author: | Jan P. H. van Santen Professor and Head of the Division of Biomedical Computer Science at OHSU |
|---|---|
| Date: | December 11, 2008 |
| Time: | 15:30 |
| Location: | 220 Deschutes Hall, University of Oregon |
| Host: | Stephen Fickas |
Abstract
Research, diagnosis, and intervention for neurological disorders are growing more interdisciplinary, and involve individuals with backgrounds in, e.g., behavioral science, genetics, neurology, and environmental science. There is a major obstacle, unfortunately: Behavioral observation is the primary means for diagnosis, yet it is intrinsic unreliable. As a result, neurological disorders for which there is no easy blood test such as most neurodevelopmental and neurodegenerative disorders, as well as disorders due to neurotoxic agents, stroke, and head injury, surprisingly often go undetected or are misdiagnosed.
The vision of Neurobehavioral Engineering is to revolutionize behavioral diagnosis and intervention by creating specialized algorithms for processing speech, language use, movement patterns, and facial expression. These algorithms build not only on the availability of powerful and low-cost hardware, but also on recent progress in areas such as machine learning, natural language processing (NLP), and signal processing. However, available algorithms rarely address the specific problems when dealing with disorders. Thus, a guiding motivation of Neurobehavioral Engineering is that will move computational science forward by defining new challenges -- challenges that are plainly very difficult.
In this talk, examples will be given of current Neurobehavioral Engineering projects in OHSU's Biomedical Computer Science Division in which we apply a wide range of algorithms to problems in autism, dysarthria, and mild cognitive impairment.