Classification of Parkinson rating-scale-data using a selforganising neural net

An application of a self-organizing neural net of Kohonen type to the data of 666 de-novo Parkinsonian patients of a multicenter study is presented. The data to be learned are the ten items of the Webster rating scale and one additional item with four stages, following the classification by Hoehn and Yahr. Multivariate linear statistical methods are applied to the data, yielding linear models, which are able to derive the Hoehn and Yahr staging from the staging of the Webster rating scale. The methods succeed with a quote of correct classification of about 50%. In contrast to these unsatisfying results, a Kohonen net with 40×40 neurons achieves a surprisingly high classification rate of approximately 90% for the four stages of Hoehn and Yahr