---

Revision with unchanged content. The perception of dynamic spatio-temporal patterns is a fundamental part of visual cognition. Psychophysical experiments have shown that a spatiotemporal memory for early vision is evident and that the processing is realized in different representational layers of abstraction. The properties of this memory structure are reflected in a multi-layered neural network model which was developed in this work. Major architectural features of the model are recurrent derivatives of Kohonen´s self-organizing maps. The model has the advantage of a self-teaching learning algorithm and stores temporal information by local feedback in each computational layer. Prediction capability was integrated by adding neural associative memories to each processing layer.

---

studied Informatics at the University Ulm wherehe already focused on neural informationprocessing and statistical learning theory. Afterreceiving his diploma he joined the TU München tocontinue working on artificial neural networks andstatistical learning methods with the focus ondynamic pattern processing.

---