Estimation Of Single Trial Evoked Potentials Using Parametric Modeling

This paper analyses the bias of Batch Least Squares estimation of single trial Evoked Brain Potentials using ARX modeling. Then, a new parametric model is proposed. Making full use of pre-stimulus EEG signal and Average EP (AEP)signal, this method can get the unbiased, minimum variance estimation of EP waveform. Experiment result is presented to demonstrate the avaiablity of the model in practice.