A biological-based neural network model of leech reflexive behaviors

An artificial neural network is built and trained to model the local bending reflex of the leech Hirudo Medicinalis. The model is validated by the data provided from physiological experiments. The training data and topology of the neural network is based on data gathered from physiological experiments. A graphical representation of the leech is constructed as a high level interface to the artificial neural network. This interface is used to determine the appropriateness of responses to different stimulus patterns. The artificial neural network is tested to see if it can generalize its repertoire of behaviors to untrained stimulus patterns. The leech provides the authors with a simple neurostructure that can be understood with today´s technologies