System for classifying working memory task magnetoencephalography based on machine learning
A system for classifying working memory task magnetoencephalography based on machine learning, including: the magnetoencephalography data acquisition module configured to acquire magnetoencephalography data of a subject in different working memory task states; the magnetoencephalography data preprocessing module configured to control the quality of magnetoencephalography data in different working memory tasks and separate noises and artifacts; the magnetoencephalography source reconstruction module configured for sensor signal analysis and source reconstruction analysis for the data processed by the magnetoencephalography data preprocessing module; and the machine learning classification module is configured to classify the working memory tasks to which the subjects belong by taking power time series as features. The present disclosure integrates the complete analysis pipeline from preprocessing to source reconstruction of the working memory magnetoencephalography data, classifies the working memory task magnetoencephalography data, and is of great significance to the study of working memory decoding and brain memory related mechanisms.
1. A system for classifying working memory task magnetoencephalography based on machine learning, wherein the system comprises a processor and a memory, and the processor performs following steps:
acquiring magnetoencephalography data of subjects in different working memory task states and inputting the magnetoencephalography data;
preprocessing the magnetoencephalography data in the different working memory task states, comprising:
performing quality check on the magnetoencephalography data in the different working memory task states;
filtering channels and data segments that do not meet requirements; and
separating a noise and artifact for noise removal and artifact identification;
performing sensor signal analysis and source-level reconstruction analysis on the magnetoencephalography data in the different working memory task states to obtain power time series features;
wherein the sensor signal analysis comprises:
time-locking analysis: acquiring a processing process of an event by the brain to obtain activity states before and after the event; and
time-frequency analysis: obtaining a relationship of signal frequency changing with time; and
wherein the source-level reconstruction analysis comprises: reversely calculating a magnetic field change inside the brain by a magnetic field distribution around a head of the subjects, and adopting a beamforming technology comprising a dynamic imaging of coherent sources method or a linearly constrained minimum variance method, wherein the dynamic imaging of coherent sources method is based on frequency domain data, and performs the time-frequency analysis before the source-level reconstruction analysis to obtain a time-frequency distribution of the magnetoencephalography data; and the linearly constrained minimum variance method is based on time domain data, and performs the time-locking analysis on the magnetoencephalography data before the source-level reconstruction analysis to calculate a covariance matrix of the magnetoencephalography data; and
performing dimension reduction through a principal component analysis method based on the power time series features, and classifying working memory tasks of the subjects by using a machine learning model.
2. The system for classifying working memory task magnetoencephalography based on machine learning according to claim 1 , wherein said performing quality check on the magnetoencephalography data in the different working memory task states further comprises: preliminarily verifying the magnetoencephalography data and outputting a document of data quality information.
3. The system for classifying working memory task magnetoencephalography based on machine learning according to claim 2 , wherein the output data quality information comprises: magnetoencephalography sampling frequency, a duration of recorded data, a number of magnetoencephalography (MEG) channels, a number of reference channels, a number of electrocardiogram (ECG) channels, a number of electromyogram (EMG) channels, a number of recorded events and average coil movement.
4. The system for classifying working memory task magnetoencephalography based on machine learning according to claim 1 , wherein said filtering channels and data segments that do not meet requirements comprises: detecting noise channels by checking a signal similarity between each magnetoencephalography channel sensor and sensors adjacent to magnetoencephalography channel sensors, and marking channels and data segments, when adjacent channels exhibit lower than a correlation threshold or higher than a variance ratio threshold, as bad channels and bad data segments and removing the bad channels and the bad data segments from subsequent analyses.
5. The system for classifying working memory task magnetoencephalography based on machine learning according to claim 1 , wherein said separating a noise and artifact further comprises: extracting each independent component by using an independent component analysis method, classifying the independent component into a brain or noise component, and performing noise removal and artifact identification by thresholding three parameters comprising correlation between independent component signals, correlation between power and time series, and correlation between frequency spectra and by multiple iterations of selecting a highest brain component and a lowest artifact pollution.
6. The system for classifying working memory task magnetoencephalography based on machine learning according to claim 1 , wherein the machine learning model consists of a support vector machine model, a logistic regression model or a random forest model.