Neural network classification of cerebral embolic signals

The presence of circulating cerebral emboli represents an increased risk of stroke. The detection of such emboli is possible with the use of a transcranial Doppler ultrasound (TCD) system. When a gaseous or particulate embolus passes through the TCD sample volume, it produces high intensity transient signals that are normally relatively easily detected. However, because most current TCD systems rely on human experts for the detection and classification of candidate events, this technique is not widely used. The appearance of a reliable automatic system, able to detect these signals and to classify them as originating from either a gaseous or solid source, would encourage the widespread utilization of this technique. This paper reports the application of new signal processing techniques to the analysis and classification of embolic signals. We applied a wavelet neural network algorithm to approximate the embolic signals, with the parameters of the wavelet nodes being used to train a neural network to classify these signals as resulting from normal flow, or from gaseous or solid emboli.