An Artificial Neural Network Approach to Estimation of the Regional Glucose Metabolism Using PET

This chapter introduces a method for fast pixel-by-pixel estimation of parameters in kinetic brain models. The method is based on artificial neural network (ANN) models. Especially, it is shown how the ANN model can be used to estimate the glucose utilization on a pixel-by-pixel basis in the brain. The data used are derived from dynamic positron emission tomography scans using the tracer [ 18 F]-fluorodeoxyglucose. Training data for the ANN model are generated by fitting the parameters in Sokoloff's model directly. It is assumed that this method can be used to identify abnormal glucose metabolism in different brain regions for subjects with serious brain disorders. By using the neural estimation procedure, the processing time for a brain scan volume is reduced from 48 hours to 4 minutes. The neural network method is proposed as a general tool for fast estimation of parameters in kinetic models.