A decision table and rule based interpretation system for epileptic discharges.

Abstract

The representation of the various features of waveforms and their correlations, in EEG recording for the diagnosis of different diseases have been carried out by many researchers due to the impact of knowledge based on expert systems development tools and techniques. The realisation of these system requires a specific hardware and software tool for its implementation, which may be a costly affair. The design and development of low cost effective system for the diagnosis of epileptic patients have been reported in this paper. Two different and linked (at certain stage) approaches (i) decision table; (ii) rule based system have been followed to model the reasoning processes of physician in the diagnosis. In the decision table the features of specific waveforms of EEG are represented in the tabular form. The features are obtained from a 8086 microprocessor based data acquisition system. The rule based system is designed with IF and THEN form of rules using Turbo-Prolog as programming language tool and is implemented on low cost PC-AT. The results obtained are at an intermediate stage of data processing by decision tables and at the final stage being carried out by rule based model. The performance of the system is evaluated by recording of EEG of some epileptic patients. The results obtained are comparable and to a certain extend appreciable in the opinion of the physician.