Authors: Benjamin D Kearsley, Jacob H Norby, Micah R Shepherd, R Ryley Parrish
Mentors: Micah R Shepherd
Insitution: Brigham Young University
Status epilepticus (SE) is a seizure which lasts more than five minutes and requires time sensitive treatment to prevent brain damage and even death. Thus, it is crucial to understand and predict the brain signal patterns preempting SE. Previous research into seizure monitoring techniques in humans suggests that seizure occurrence follows non-random patterns, and that big data and machine learning may be key in discovering detection and prediction models for SE events. This provides a promising foundation for our investigation into a prediction algorithm for SE using data-driven methods.
Using a state-of-the-art Multiple Electrode Array (MEA) recording device, high-resolution signals have been recorded that demonstrate the voltage that occurs within mice brain cells during SE. In this study, voltage traces from these brain cells are being examined to identify predictors and indicators of SE events in the brain. Signal processing techniques, such as digital filtering, spectral analysis via spectrograms, and wavelet transforms, are being investigated to analyze these signals and compare them to brain signals exhibiting normal activity. Statistical methods surrounding the first through fourth moments, as well as clustering algorithms may also be used to classify and study the signals.