Research On Spatial Cognitive Training And Evaluation System Integrating Brain-computer Interface And Virtual Reality

Spatial cognition is a research hotspot in the field of cognitive science at home and abroad,and has achieved a series of results,but there are few studies on training and evaluation of spatial cognition.The main reason is that its scientific training evaluation system is difficult to construct,the training effect is quantitatively simple,and the evaluation of the training effect is greatly influenced by the subjective factors of the observer.Therefore,this paper has carried out in-depth research from the scientific and effective spatial cognitive training evaluation system design and the quantitative analysis of the training evaluation effect based on the system.Firstly,this study builds a spatial cognitive training and evaluation system that integrates brain-computer interface with virtual reality technology.The system trains and tests subjects with spatial cognitive tasks in a virtual environment,and simultaneously collects EEG signals and records the behavior data of the subjects.The off-line analysis method is used to analyze the data before and after training,the behavioral data uses statistical methods to analyze multiple indicators such as time and distance,and the EEG signals are extracted and classified by means of granger causality analysis and permutation conditional mutual information method,so as to objectively and quantitatively evaluate the training effect.Secondly,in order to further test the performance of the system and solve the problem of the traditional EEG signal analysis method easily ignores the relative position information between channels,this paper proposes a spatial cognitive evaluation EEG signal analysis method based on Multivariate Permutation Conditional Mutual InformationMultispectral Image.The method fully considers the spatial properties of the coupling characteristics of EEG signals,transforms the multi-band coupling features into multispectral images,and classify these image data by means of convolutional neural networks.Thus effectively assessing changes in spatial cognitive ability before and after training.Finally,this study conducted an experimental analysis of the above content.This experiment recruited 7 subjects to use the above system for 20 days of training and pre-and post-testing,the effectiveness of the system in spatial cognition training and effect quantification was verified from different angles.In addition,the superiority of the Multivariate Permutation Conditional Mutual Information-Multispectral Image method in the application of spatial cognitive ability evaluation is verified by comparing with other methods.