Title :
An expert system to aid diagnosis of epilepsy
Author :
Dingle, Alison A. ; Jones, Richard D. ; Carroll, Grant J.
Author_Institution :
Christchurch Hospital, New Zealand
Abstract :
An expert system has been developed to detect epileptiform activity in EEGs. Epileptiform events are reported as definite or probable, which helps overcome the problem of maintaining satisfactory detection rates, while minimizing false detections. The system has been evaluated on EEGs from 21 patients, a total of 380 minutes of recordings. On average the system detected 53% of epileptiform events as definite with no false detections, and 64% of events as definite or probable but at the expense of 3.5 false detections per hour. The latter detection rate compares very favourably with that of other systems. However, the outstanding feature of the system is its ability to detect 53% of events as definite with no false detections
Keywords :
electroencephalography; expert systems; medical diagnostic computing; performance evaluation; EEG evaluation; detection rate; epilepsy diagnosis; epileptiform activity detection; medical expert system; Biomedical engineering; Diagnostic expert systems; Electroencephalography; Epilepsy; Event detection; Expert systems; Hospitals; Medical diagnostic imaging; Medical expert systems; Physics;
Conference_Titel :
Artificial Neural Networks and Expert Systems, 1993. Proceedings., First New Zealand International Two-Stream Conference on
Conference_Location :
Dunedin
Print_ISBN :
0-8186-4260-2
DOI :
10.1109/ANNES.1993.323034