ARTIFICIAL NEURAL NETWORK BASED CLASSIFICATION OF CEREBRAL BLOOD FLOW SIGNALS

Abstract

Oscillation of the cerebral blood flow (CBF) is a common feature in several physiological or pathophysiological states of the brain. It is promising to identify the disorders of the cerebral circulation based on the classification of CBF signals. In order to distinguish between different physiological states, an artificial neural network classification model has been developed using spectral matrix based feature vectors describing the temporal blood flow patterns. The efficiency of the classification is evaluated and compared to the results obtained by wavelet subband analysis.