An adaptive backpropagation neural network for real-time ischemia episodes detection: development and performance analysis using the European ST-T database

Abstract

A supervised neural network (NN)-based algorithm was used for automated detection of ischemic episodes resulting from ST segment elevation or depression. The performance of the method was measured using the European ST-T database. In particular, the performance was measured in terms of beat-by-beat ischemia detection and in terms of the detection of ischemic episodes. The algorithm used to train the NN was an adaptive backpropagation (BP) algorithm. This algorithm drastically reduces training time (tenfold decrease in our case) when compared to the classical BP algorithm. The recall phase of the NN is then extremely fast, a fact that makes it appropriate for real-time detection of ischemic episodes. The resulting NN is capable of detecting ischemia independent of the lead used. It was found that the average ischemia episode detection sensitivity is 88.62% while the ischemia duration sensitivity is 72.22%. The results show that NN can be used in electrocardiogram (ECG) processing in cases where fast and reliable detection of ischemic episodes is desired as in the case of critical care units (CCU's).

Ischemia detection using supervised learning for hierarchical neural networks based on kohonen-maps

“…As it is expected from the beat classification results, the sNet-SOM with SVM as supervised expert yields a better average episode detection performance. Generally, the results we have obtained are close to the results reported by other authors [10], [11], [12]. At the SVM case we can claim that we have a small improvement of the detection ability.…”

“…Also, we obtained better results using the Manhattan distance measure [2] in comparison to those obtained with the alternative Euclidean measure. Although the SOM is trained with the usual SOM unsupervised training algorithm [2], it has the potential to obtain a classification accuracy close to those reported in [10], [11], [12] with supervised neural models.…”

“…We defined the ischemic episodes in terms of ischemic beats according to the same set of criteria as those used in [10]. Correctly detected episodes are termed True Positive (TP) episodes.…”

Dimensionality Reduction Oriented Toward the Feature Visualization for Ischemia Detection

“…On the other hand, NNs have been widely used as pattern and statistical classifiers in many application areas, including medicine [22]. A supervised NN-based algorithm was used in [23] for the automated detection of ischemic episodes resulting from ST segment elevation or depression. The NN training was carried out by an adaptive backpropagation (BP) algorithm, which drastically reduced training time.…”

“…The only vestige of adaptation thatsome systems present is threshold adjustment, which can be performed manually or automatically [26]. Other systems use classifiers based on NNs [23], for which the adaptation process is time-consuming and hard to implement. Hidden Markov models (HMMs) are also employed, which have been shown to be suitable for adaptation, due to the low complexity and the suitability of their structure in implementing online adaptation.…”