The goal of our area is to be able to predict that a seizure is going to occur, and then stimulate the brain in such a way as to break up the seizure before it ever happens.  We continuously record and monitor high frequency electrical activity from the brains of epileptic rats. We analyze this data in multi-dimension space using custom machine learning algorithms in an attempt to find patters of activity that predict a seizure will occur in the near future.  We use the result of these analyses to trigger stimulation of the brain at appropriate times. To stimulate we inject small amounts of current in specific patterns and locations. The goal of this research is to find the right pattern of stimulation that will break up the brain activity leading to seizure; and to find the right algorithms to detect the patterns of brain activity the lead to seizures.

Want to listen to a seizure?  The following windows medial file is a recording of a spontaneous seizure as recorded in the hippocampus of a rat.  The signal has been filtered so that what we are hearing are the firing of individual cells.  The very loud cracklings starting about half way through are many cells firing at the same time and are clipped (at 1/10 their original amplitude).   It was recorded from electrodes adjacent to each other and is in stereo:  se1121d.wma