Complex symptomatology simulated by unstructured neural nets.

RESUME Un modele de reseau neural, avec d'interconnections au hasard ("neural net model"), est presente, pouvant demontrer les manifestations essentielles de la maladie de Parkinson, inchiant rigidite de forme roue-dentee, tremblement au repos el dysdiadohokinesie. Ces derniers sont simules settlement par rapport a la maladie extra-pyramidale; on prevoil spasticite et hyper-reflexie, avec d'atttres lesions de neurones mot rices supe rie ure s. INTRODUCTION The pathophysiology of Parkinson's Disease is presumed to involve an imbalance of complex servomechanisms incorporating the alpha motoneurons of the spinal cord, Ward (1968). The nature of such neuronal circuitry and the precise manner in which it generates the characteristic features of the disease have, however, remained unclear. In an attempt to elucidate the organizational features essent ia l to the d e v e l o p m e n t of such s t r i k ing phenomena as cogwheel rigidity and tremor, a neural net model developed by Anninos et al. (1970), Anninos (1970), Anninos and Harth (1971), Anninos (1972a), and Anninos (1972b), is proposed as representative of the alpha motoneuron pool. As will be shown, even though it contains only random interconnections, such a net is capable of exhibiting the characteristic features found in Parkinson's disease.