Application of artificial neural network approach to the interpretation of the biosignal of diabetic skin microcirculation

The biosignal of skin blood flow can be measured using a laser Doppler flowmeter (LDF) with a PF3 probe. This paper documents the assessment of these biosignals using an artificial neural network (ANN) to diagnose subjects as diabetic or normal. The interpretation of results is based on receiver operating characteristics (ROC) curves and shows that neural networks have advantages which may be regarded as a potential tool in medical decision making.