Design And Implementation Of EEG Visualization Platform Based On Microservice

With the advancement of brain science,reliable and user-friendly Electroencephalography(EEG)signal processing platforms have become an important way to help researchers better understand the nature of EEG signals.However,EEG processing platform related products are still relatively scarce in China,and most EEG research work depends heavily on foreign software.The existing EEG processing platforms usually need to be deployed on the local machine,which leads to the bottleneck of hardware resources and difficult data sharing.In addition,they also have many other problems,such as inconvenient data management,poor software stability,lack of customization,non-open source system,high fees,and limited regulation.In order to solve the above problems and fill the gaps of existing EEG processing platforms,this thesis aims to develop an EEG visualization platform,based on microservice architecture,which includes functions of researcher information management,EEG data management,raw data processing,EEG signal visualization and real-time data sharing.The main research work and results of this paper are as follows:(1)In view of the problem that the traditional EEG signal processing platform cannot be flexibly expanded and can only be deployed on a single machine,the EEG visualization platform designed in this thesis adopts a distributed architecture composed of multiple microservices with independent responsibilities,and is deployed on the cloud platform through Docker containerization technology.This architecture scheme supports the independent deployment of different services.Compared with the traditional monomer architecture,this architecture solution has higher fault tolerance and scalability,and can maximize the utilization of elastic resources of the cloud platform to solve the problem of hardware resource limitation caused by the single-node deployment of existing software.(2)In view of the problems that the user interface is difficult to understand and the visualization is not friendly,the front-end system of the EEG visualization platform implemented in this thesis is developed using React ecological technology to provide users with a friendly operation interface.At the same time,this thesis proposes an EEG visualization scheme based on Kriging interpolation algorithm.By spatially modeling the brain research area and using the semivariogram to calculate the spatial correlation between the unknown point and the sample point,the unknown point can be obtained.attribute value.Through the comparison of the effect of estimating unknown points,it proves that the Kriging interpolation algorithm has a 24% accuracy improvement in the rendering of the brain topographic map compared with other commonly used interpolation algorithms,providing users with clear and intuitive EEG signal processing results,which is convenient for data observation and analysis.(3)In view of the performance bottleneck of the data sharing function in high-concurrency scenarios,the back-end system employs Redis to cache frequently accessed hot data.The stress test results show that Redis cache can significantly improve system performance.The performance improvement rate in high-concurrency scenarios is over 34%,and the higher the concurrency,the more obvious the performance improvement.The research results of this thesis offer an efficient and user-friendly data processing and analysis tool for EEG researchers.The subsequent promotion and application of this platform are expected to promote the development and progress of EEG research,which is of great significance to the EEG signal processing research and clinical treatment.