Technique to Detect the Onset of Epileptic Seizures

Technology #11241

Identifies Pattern Changes in Time Series to Predict Epilepsy, Sleep Disorders, and Even Weather Forecasts

This exciting technology predicts epileptic seizures. Approximately 2.8 million people in the United States suffer from epilepsy, with an estimated 300,000 new cases diagnosed each year. If untreated, an individual with epilepsy is likely to experience repeated seizures, which typically involve loss of consciousness. The ability to detect an epileptic seizure allows doctors to more effectively treat epilepsy patients by providing preventive treatment. Our researchers have developed a technique to identify brain patterns that predict epileptic seizures before they occur.


  • Predicting the onset of epileptic seizures
  • Analyzing data in multidimensional systems such as stock markets, weather forecasts, and manufacturing processes


  • Provides effective seizure warning and prediction not available through traditional signal processing techniques, offering a competitive advantage
  • Allows doctors to deliver treatment prior to onset to prevent seizures
  • Non-invasive therapy which does not require the removal of hippocampus or cerebral cortex in the brain significantly increasing potential patient population
  • Provides broad potential market application since the technology can be used in various industries, from financial market analysis to geological predictions


Epileptic seizures can be predicted based on spatiotemporal dynamics in brain patterns of patients with epilepsy. There is a temporal transition before a seizure occurs at certain critical sites in the brain. Our researchers have developed a technique to analyze brain behavior at these critical sites in order to detect an impending epileptic seizure. The technique identifies spatiotemporal patterns in EEG time series that characterize the onset of a seizure. This is the first time that spatiotemporal analysis has been used for epileptic seizure detection. These analysis and prediction techniques can be applied to other systems which are based on dynamic time series.