Power Spectrum Estimation of the Welch Method Based on Imagery EEG

Abstract

Welch method is a direct evolution from the periodogram method. Periodogram method is also commonly used in the power spectrum estimation, there are some inherent shortcomings in periodogram method, such as the variance and resolution of the spectrum estimation is not good, it does not satisfy the consistency estimation conditions and so on, so this paper uses improved periodogram method (welch method) to estimate the motor imagery EEG power spectrum, with Matlab for tools, through the simulation on the experimental data, comparativing and analysising the welch method’s different spectrum estimation properties with different window functions, which provides the theoretical guidance for selecting a suitable window function and makes the welch method play a good role in the EEG feature extraction.