Abstract
Sleep-related phenomena or disorders, including snoring and tooth grinding, can be investigated using polysomnography. This method, however, generates large amounts of synchronically recorded data that are often analyzed visually with subjective interpretation. The purpose of this study was to minimize the need for subjective evaluation by developing a computer program for analysis of EMG data linked with polysomnographic recordings in a standardized and semi-automatic way. The selected algorithm differs from the Root Mean Square (RMS) method by being based on the theory of "differentiated EMG" (DIFEMG), which relies on two principles. The first says that the activation of a larger number of motor units results in a greater force production. The second principle says that the force production will continue for some time after the muscle is no longer stimulated. After a visual check for artifacts in the basic EMG recordings, the computer program is used to analyze the corrected basic EMG signal. The results were that both methods yield identical results as far as the detected number of events is concerned. There is, however, a significant difference when the duration of the events is considered, because the start and end of an event can be more accurately determined with the new method presented here. The computer program described will make comparison of data from different studies easier.