Autoregressive coefficient changes of EEG signal model

Clinical records of intraoperative ischemic EEG are used to optimize an automatic detecting algorithm which is capable of signaling the onset of the insult accurately at a small delay. The algorithm derives autoregressive (AR) coefficients on all channels, and searches for simultaneous changes, which has earlier been found to be characteristic of the ischemic change. A set of 32 clinical recordings were used to determine optimal parameter thresholds. An intraoperative clinical monitor based on the present method may be of great help in reducing intervention related risk during carotid and cardiac operations.