Computer automated versus visually determined electroencephalographic and electromyographic seizure duration

Seizure durations were determined during electroconvulsive therapy (ECT) by a computer-automated procedure for interpreting electroencephalographic (EEG) and electromyographic (EMG) signals. These seizure durations were compared with durations determined by two experienced clinicians from simultaneous paper recordings of EEG and EMG and with independently recorded observations of cuffed-limb motor movements. The computer EEG seizure end point was programmed as the point where the moving average EEG voltage first fell below the pre-ECT value and remained there for at least 2 s. The computer EMG seizure end point was programmed as the point where the moving average EMG voltage first fell below 200 mV and remained there for at least 2 s. Close correspondence between seizure durations rated from EEG recordings and computer-automated EEG measurements demonstrated validity and high reliability of the computer procedure. Likewise, validity and reliability were demonstrated for motor seizure durations both computer-derived and interpreted from the paper EMG recording by their close correspondence with cuffed-limb motor seizure durations.