Research On Data Acquisition And Automatic Artifact Removal Technology Of Multi-channel EEG Signals Based On Flexible Needle Dry Electrode

The acquisition and preprocessing of electroencephalography(EEG)is the most important part of the brain-computer interface application acquisition module and plays an important role in neuroscience,medicine,psychology and artificial intelligence.The acquisition of multi-channel EEG signals and the artifact removal have become important research topics in brain-computer interface.The current hairy site multi-channel EEG signals acquisition and preprocessing technology have the following problems: The traditional wet electrode is cumbersome to operate and has a long preparation time.The hair masking causes poor quality of collected EEG signals,and it is susceptible to artifacts such as movement,swallowing,and blinking during the acquisition process.The traditional artifact removal algorithms have the disadvantages of non-full automation,low real-time performance,and poor objectivity.In response to these deficiencies,this paper focuses on three aspects:1)A flexible needle hairy site dry electrode is designed and prepared.The array probes and the bottom of the electrode have a rounded chamfer structure.It is not easily broken during use;the probe shape is a cone and can penetrate the hair through the scalp;the electrode material is rubber,and the conductive silver powder is doped,making the electrode conductive.In order to verify the validity of the electrode,impedance characteristic analysis experiments,eyes open/closed experiments and SSVEP experiments are performed.Impedance analysis experiments show that the impedance values of 10 Hz in the forehead area and the hairy site of the individual electrode are 57 ?6.2k? and 90 ?12.7k? respectively,and can reach the impedance requirements when the equipment collects EEG data.The frequency domain analysis of the EEG signals collected by the eyes open/closed experiments and the SSVEP experiment,the results show obvious characteristics.2)A method for automatic removal of multi-channel EEG artifacts based on low-rank sparse decomposition is proposed.This method analyzes the multi-channel EEG signal and improves the GoDec algorithm.The observed EEG signal affected by artifacts X is decomposed into ' ' 'X(28)L(10)S(10)G,which L' is pure EEG signal,S' and 'G is the artifact part.Remove 'S and 'G to achieve automatic removal of artifacts.The proposed method performs artifact removal experiments on standard EEG data sets,and the results show that this method is effective in automatically removing artifacts.Compared with the existing automatic artifact removal algorithms,the proposed method greatly improves the SNR and reduces MSE of EEG signals,and has the advantages of better stability and short artifact removal time.3)The prepared flexible dry electrode is used to collect multi-channel EEG signals,and the proposed algorithm is used to automatically remove the artifacts.The results show that the amplitude of the artifact signals are significantly reduced,and the processed signals reach the standard of normal EEG signals,and the waveband of signal decomposition accords with the frequency characteristics of EEG signature wave,the prepared flexible dry electrode and the proposed method are suitable for daily wear monitoring.