A sleep scoring algorithm for the rat EEG based on AR modeling

An approach using the second-order autoregressive (AR) modeling technique is presented for distinguishing between the electroencephalographic patterns associated with vigilance states of quiet waking, slow-wave sleep, and REM sleep in the rat hippocampus and frontal cortex. The specific measures of energy, dominant frequency and its associated intensity, are computed. These parameters are used as the decision variables to discriminate between the various vigilance states.<
>