Neural network analysis of internal carotid arterial Doppler signals: predictions of stenosis and occlusion

Abstract

Abstract Doppler ultrasound is a noninvasive technique that allows the examination of the direction, velocity, and volume of blood flow. Doppler ultrasound has proven to be a valuable technique for investigation of artery conditions. Therefore, Doppler ultrasonography is known as reliable technique, which demonstrates the flow characteristics and resistance of internal carotid arteries in stenosis and occlusion conditions. In this study, internal carotid arterial Doppler signals were obtained from 130 subjects, 45 of them had suffered from internal carotid artery stenosis, 44 of them had suffered from internal carotid artery occlusion and the rest of them had been healthy subjects. Multilayer perceptron neural network employing backpropagation training algorithm was used to predict the presence or absence of internal carotid artery stenosis and occlusion. Spectral analysis of internal carotid arterial Doppler signals was done by Burg autoregressive method for determining the neural network inputs. The network was trained, cross validated and tested with subject's internal carotid arterial Doppler signals. Performance indicators and statistical measures were used for evaluating the neural network. By using the network, the classifications of healthy subjects, subjects having internal carotid artery stenosis, and subjects having internal carotid artery occlusion were done with the accuracy of 95.2, 91.3, and 91.7%, respectively.