Entwurf und Validierung eines klinischen Diagnose-Assistenten zur Klassifikation des Greifverhaltens nach realen und virtuellen Objekten

In this work, a concept for the execution and kinematic analysis of motion studies is introduced and realized, adding innovative computer-aided methods to the so far existing procedures. Firstly, a classification assistant for grouping kinematic motion data is developed and evaluated, which is alternatively based on Adaptive Resonance Theory or Self-Organizing Feature Maps. It can be demonstrated that this tool is capable of classifying patients, suffering from sensorimotor disturbances after parietal lesions, according to their grasping behaviour. Secondly, it is shown how computer-generated experimental setups can be used in grasping studies. In this context, the motor behaviour while grasping at computergraphic, virtual objects is investigated and related to existing theories about the sensorimotor mechanisms of the reach to grasp movement. Here, the classifcation assistant can be used to identify different motion strategies while grasping at virtual objects.