A wavelet transform coupled with a fuzzy neural network for prediction of significant st segmental changes in the ecg.

The leading cause of death in the United States for people 65 and over is heart disease. A significant factor contributing to this disease process is the damage caused by an infarction, which can manifest as an abnormality in the ST segment of an Electrocardiogram (ECG). This research will develop a pattern recognition model that will be capable of detecting these critical changes. This model will be developed using a feature extraction scheme based upon Wavelet analysis and a classification scheme based upon a Fuzzy Neural Network design. These schemes will be implemented using software tools available from MatLab. Evaluation of the model will be accomplished by simulation (MatLab) with representative ECG samples obtained from a database (e.g. MIT-BIH) that have been universally accepted for such a purpose. This model could be available for implementation into a device used in the pre-hospital setting that would provide the capability of early detection of critical ST changes. Accurate detection of these abnormalities can provide the means for establishing guidelines to determine a treatment protocol that may save lives.