Research On EEG Measuring System And Compression Algorithm

Brain science is one of the highlights of life science and technology research of 21 st century.EEG is the totality of the electrophysiological activity of nerve cells,and it is a key tool to understand the brain neural physiological features.Thus,EEG is one of the cornerstones of brain research.EEG monitoring is not only a widely used tool of significance for diagnosis of brain diseases in clinics,it has been widely used in the braincomputer interface applications.Effective utilization of EEG involves EEG acquisition,transmission,and many other problems.This paper designs a portable EEG acquisition system for the convenient acquisition of EEG,and puts forward a novel and effective multi-channel EEG compression algorithm for EEG transmission and storage issues.The traditional EEG measuring devices usually have too large volume and high power consumption,making them not suitable for portable use.This paper designs an EEG measuring circuit based on the analog front-end chip of TI Company ADS1299,with whose high level of integration and excellent performance the size,power consumption and overall cost of the circuit is dramatically reduced.The low cost,open source Arduino platform is used as the controller of the EEG acquisition system.The measuring circuit,a Bluetooth module and an Android mobile phone constitute a complete portable EEG measuring system.And the experiments like measuring real EEG signals prove the validity of this system.The proposed multi-channel EEG compression algorithm combines principal component analysis(PCA),independent component analysis(ICA)and SPIHT algorithm.FastICA algorithm is a kind of ICA algorithm,and it has the feature of low computational complexity and fast convergence.The proposed algorithm adopts the FastICA algorithm because it has the property that independent components can be extracted one by one.Using the PCA algorithm as a preprocessing step improves the performance of FastICA algorithm.By using SPIHT compression algorithm after rearranging one-dimensional data into a two-dimensional matrix form,the correlation between the samples is efficiently utilized.The paper also discusses the selection of parameters in the proposed algorithm.Experiment of comparison with two other cutting-edge algorithms proves the value of the proposed algorithm.