Construção de exercício para verificação do aprendizado utilizando redes neurais artificiais - fase ii medula espinhal

Abstract

The disciplines that need human body parts to teach their programs have great problems dealing with obtaining and maintaining those organs. To avoid these difficulties this paper suggests a prototype constructed with Artificial Intelligence, through Artificial Neural Nets, used to develop a system with which the student can interact during his learning. The program has computer treated real images to substitute body parts and the user can point, given signals and symptoms, which areas of the organs are their cause. The constructed prototype is directed to the spinal cord and an Artificial Neural Net is developed with patterns obtained in normal Neuroanatomy books with an expert help. The user is presented, at each interaction, to a random set of signals and symptoms and then he is urged to indicate in a displayed image, which parts of the organ could be damaged. Tests with medicine students using this dynamically conducted software had good results, showing that adequate systems could replace the usual methods or be used in cooperation with them, diminishing the related problems. © Springer-Verlag Berlin Heidelberg 2007.