Program Overview
In the United States, epileptic seizures affect about 1% of the entire population. Seizure prediction, early recognition, and blockage of seizures are considered by the membership of the American Epilepsy Society as the first research priority listed among fifteen. The disease's abnormal brain activity can be monitored by an electroencephalogram (EEG) recorded via electrodes attached to the scalp, or by an electrocorticogram (ECoG) from electrodes in direct contact with the cortex. At the present, EEG and/or ECoG are the only windows through which the origin and dynamics of epilepsy can be revealed. Analysis of EEG and ECoG is thus of fundamental importance for reaching the grand goal of clinical seizure prediction.

In collaboration with Flint Hills Scientific, L.L.C., The University of Kansas Medical Center, and Arizona State University, the Scientific Research Group is investigating techniques derived from nonlinear dynamics for the detection, prediction, and tracking of epileptic seizures. Techniques are being developed based on analysis of signal recurrences in ECoG, including detection of unstable periodic orbits, in order to characterize spatiotemporal evolution of seizures arising from focal epilepsy. Statistics from recurrences are also being developed to quantify signal rhythmicity, a property known to be present in both seizure and seizure precursor brain wave activity. All statistics will be tested on a database of 25 multichannel ECoG recordings taken from five patients who underwent invasive evaluation prior to surgical treatment.

• Developed and tested rhythmicity statistic; and
• Paper contributed to peer-reviewed journal Chaos.