In the summer of 2015, we were admitted to RMP held at the University of California, Santa Barbara. It was here where we got our first taste of undergraduate level research and the blood, sweat, and tears that comes with it. Our mentor, Abhejit Rajagopal, provided quite the computational programming challenge. After six weeks of research, we had finally constructed a working algorithm that was a step in the direction of anticipating epileptic seizures minutes before their onset. A new computational approach to signal development made multivariable brain signals into simple 3-dimensional points. Our algorithm then implemented multivariable calculus that analyzed the points and predicted the location of the next brain signal based on patterns. The accuracy of the prediction determined the likelihood of a seizure. During our stay, we attended GRIT talks, which are lectures organized by RMP that invites modern researchers to discuss their current projects and thoughts toward their endeavors. These talks presented research that is grossly interesting and opened up our eyes to the opportunities available to us and the problems that need to be solved. Knowing that we had helped future epileptic patients gave us a wonderful appreciative feeling. Our current research strives to continue aiding the medical world, but in a different domain.
Anticipating epileptic seizures is a conundrum that has eluded researchers because of their characteristic unpredictability and severely hindered patients' quality of life. While existing technologies can recognize seizures seconds before they occur, we present an algorithm that has the potential to complete the same prediction task with minutes to spare. We have delved into implementing a new computational approach to signal processing that utilizes discrete mathematics to identify the seizure transition period. A sensitivity analysis demonstrates that during a seizure, brain signals evolve to create a discernable pattern. Drawing inspiration from least squares regression, we apply a parameterization technique to contour fit the electroencephalogram (EEG) data. Quantifying the anomalous change that occurs within the transition period, we subsequently calculate the margin of error in order to illustrate how patterned the brain signals are.