Department of Computer Science and Engineering

Project Abstract

As we grow and progress the medical sciences, the tools we have access to become better and stronger. We will be able to monitor almost every aspect of the human function, from the slightest twitch in the muscles to the electrical signals in our very brains. However, with this new influx of information, a technology is grossly overlooked: the method in which we view this data. Imagine that you need to correct your taxes and need some information from a large purchase like a car or a home; you would not sift through every single purchase that year, from the smallest grocery purchase to the quarters that you spend in vending machines. You would want to focus on those documents that are relevant to that purchase. Yet, doctors working in epilepsy research must deal with exactly that: rather than being able to focus on important data, they must sift through continuous irrelevant monitoring records to find the few spots that may indicate a seizure event. SmartEEG aims to remedy this issue. SmartEEG can show doctors the exact data that they need in order to more efficiently recognize EEG patterns around seizure events, allowing doctors to focus their time on more important issues. SmartEEG can take an .edf file, a standard file output format for EEG monitoring machines, and prunes the data down to digestible increments. SmartEEG accomplishes this by recognizing alarm tags, annotations that indicate a possible seizure risk or seizure event and categorizing the level of risk present (being low, medium, or high). In the case of a medium or high-risk event, SmartEEG will save fifteen minutes before and after the event and prunes the rest of the irrelevant data. SmartEEG then presents the EEG graph in a scrollable, tabulated format. This will assist doctors by expediting the EEG reviewing process and provide an easily manageable location where doctors can quickly jump between important data periods and safely ignore areas of little or no interest.

Project presentation


CMS Login