Multi-channel Electric Field Data Processing And Monitoring Interface Design Based On Transcranial Magnetic Stimulation

Transcranial magnetic stimulation(TMS)is a kind of magnetic therapy technique which makes patients noninvasive and painless.It uses time-varying magnetic field to induce biological current in brain tissue to control the electrical activity and metabolism of cells to treat diseases.Due to the brain cells are sensitive to the current,any errors may have impacts on the treatment.Thus,the stimulus signal must be accurate.In general,the signal is inevitably contaminated by various noises in the process of transmission.In order to remove these noises effectively,an improved signal subspace denoising algorithm is proposed in this thesis.The noise variance is estimated by the contribution rate of eigenvalues obtained from eigenvalue decomposition of noisy covariance matrix.In addition,an adaptive Lagrange factor formula is proposed via simulating the relationship between noise and Lagrange factor under the constraint of maximum signal-to-noise ratio.Then different noises are added into the original signal for testing,respectively.The results show that the proposed subspace method has significant improvement in performance indicators.Compared with original subspace method,the proposed subspace approach based on spectrum domain constraints increases by 14% in signalto-noise ratio,and decreases by 25% and 29% in root mean square error and mean absolute error separately.Correspondingly,the proposed subspace approach based on time domain constraints increases by 23% in signal-to-noise ratio,and decreases by 36% and 16% in root mean square error and mean absolute error,respectively.And the proposed subspace method can effectively remove both the random noise and pulse interference in stimulus signal.Moreover,in order to monitor the stimulation signals in real time during the treatment,a graphical user interface is designed for TMS system using Qt Creator5.12 as the integrated development environment.It contains eight threads to get the real-time performance.Monitoring interface of 88 channels is implemented successfully.The nephogram and rotated model are developed via Open GL to display the distribution of magnetic field.Oracle database is utilized to back up all the data.The test results indicate that the monitoring interface has a good real-time monitoring and complete visualization.It not only makes the signal more certain and the distribution of magnetic field clearer,but also promotes the development of TMS.