Specific processing of the spontaneous EMG. Detection and classification of multiplets using self-organizing neural networks.

Abstract

To the authors' knowledge, their study is the first approach to a quantitative study of spontaneous EMG. Using the given algorithm, the spontaneous EMG firing that occurs in spasmophilia has been studied on a quantitative basis, and clinical preliminary results have confirmed it as a useful tool to make the diagnostic process more sensitive, and to help in the quantitative analysis concerning clinical correlation with different pathologies. After a classification of the multiplets is performed, the final maps may be used, together with other clinical and paraclinical data, as tools the clinician may rely on in monitoring the patient's status and eventual effects of therapy. In a study on thyroidian pathology with signs of spontaneous EMG activity where 41 subjects were investigated, the program based on the authors' algorithm allowed the dynamic quantitative monitoring of the impact of the modifications of the plasmatic concentration of the thyroidian hormones on the P-Ca metabolism in order to properly initiate and conduct the therapy. The authors' quantitative approach seems promising for further use in clinical practice; it is a distinct noninvasive quantitative EMG examination that can be used by the physician in conjunction with other laboratory and clinical data.