New perspectives for surface EMG in clinical neurophysiology: from biophysics to applications.

An electromyogram (EMG) is a registration of the electrical activity of a muscle. Although often made using a small intramuscular needle, the EMG can also be recorded non-invasively, using electrodes on the skin over the muscle. The technique is then called surface EMG (sEMG). By contrast with needle EMG, sEMG plays only a minor role in the diagnosis of neuromuscular disorders. The major reason for this is that a single sEMG signal contains too little information regarding motor units. A motor unit (MU) is the functional unit of a muscle, consisting of a neuron together with the muscle fibers which this neuron innervates. This thesis shows that a biophysical approach can bring new perspectives to sEMG. The main goal of the presented research was the realisation of a system that would allow exploration of the clinical and scientific possibilities of topographical sEMG. The term 'topographical sEMG' indicates that the method uses not only temporal but also spatial information, in this case acquired with 128 surface electrodes. To provide a better understanding of the topographical sEMG signal characteristics and to guide the design process, a biophysical model was developed. Its results determined, amongst other things, the distance between the electrodes in the 128-electrode grid. To enable the comparison of the new, multichannel technique with conventional needle EMG, the design focus was on its possibilities in MU analysis. For initial application and validation of the method, data-acquisition and -analysis tools were developed to determine two clinically important characteristics of MUs: their size and number. In addition, based on the topographical nature of the multichannel data, it proved possible to develop software that allows (at least for low contraction levels) accurate determination of the firing moments of individual MUs in the sEMG interference pattern. These could be used to determine the MU's spatiotemporal profile, which in turn allowed for extensive MU characterization