A seizure warning system for intracerebral EEG is proposed. It is designed for clinical use with the intention of identifying sections of EEG containing seizure activity and alerting medical staff as a seizure occurs. The system is based on data filtering, spectral feature extraction, probability analysis using Bayes' theorem, spatial and temporal context analysis, and user tuneability. / The system was designed using 407 hours of EEG from 19 patients having 152 seizures. Once developed, the system was tested with a different set of EEGs from 19 patients having a total of 100 seizures during 389 hours. / Average results for the testing data were promising, with 86% sensitivity, a false detection rate of 0.47/hour, and a delay time of 16 seconds. Compared to current clinical systems, this shows a 9% enhancement in sensitivity, and a false detection improvement by a factor of 9.6. The aim of tuneability was also reached.
| Identifer | oai:union.ndltd.org:LACETR/oai:collectionscanada.gc.ca:QMM.81537 |
| Date | January 2004 |
| Creators | Grewal, Sukhjit |
| Publisher | McGill University |
| Source Sets | Library and Archives Canada ETDs Repository / Centre d'archives des thèses électroniques de Bibliothèque et Archives Canada |
| Language | English |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Format | application/pdf |
| Coverage | Master of Engineering (Department of Biomedical Engineering) |
| Rights | All items in eScholarship@McGill are protected by copyright with all rights reserved unless otherwise indicated. |
| Relation | alephsysno: 002166463, proquestno: AAIMR06554, Theses scanned by UMI/ProQuest. |
Page generated in 0.0057 seconds