Research On Evaluation Of Spatial Cognitive Training Methods And EEG Signal Analysis Based On Brain-computer Interface And Virtual Reality

Since the 21 st century,with the accumulation of knowledge and the continuous innovation of scientific and technological means,the field of brain science has developed rapidly,and the training and evaluation of spatial cognition has also received extensive attention at home and abroad.Due to the aggravation of the aging of the Chinese population,and with the increase of age and the decline of spatial cognitive ability,the elderly are easily lost in the complex environment.Therefore,improving the spatial cognition ability of the elderly and exploring the spatial cognition training methods of the elderly have become a hot and urgent problem in the field of brain science.This paper evaluates the spatial cognitive training system that integrates brain-computer interface and virtual reality technology from the perspectives of behavioral data and EEG signal data,and proposes a more effective classification method for spatial cognitive EEG signals.,In order to play a more accurate role in judging the effect of this training method.First of all,this article uses behavioral data and EEG signal data to evaluate the effects of the subjects' spatial cognitive training.In terms of behavioral data,the distance between the memory platform position selected by the participants before and after the training,the total time spent and the distance moved are used as indicators for statistical analysis;in terms of EEG signal data,through MPCMI-MSI The method extracts the features of the data and classifies the data through CNN to objectively and quantitatively evaluate the training effect.Then,this paper proposes a multi-scale deep dense convolutional neural network.Since the features obtained by the traditional CNN method are not diverse enough,and will cause problems such as gradient dispersion,this chapter uses the Dens Net structure to design a multi-scale deep dense convolutional neural network on the basis of a multi-scale convolutional neural network.And use the adaptive gradient stochastic descent algorithm to optimize the process of the proposed classification algorithm.Finally,this study analyzes the experimental data through the above-mentioned research methods.The 20-day spatial cognition training effect of the subjects was evaluated through behavioral data and EEG signal data,which verified the effectiveness of the system in improving spatial cognition during the training process.For the proposed multi-scale deep dense convolutional neural network,compare it with ordinary convolutional neural network and multi-scale convolutional neural network.The conclusion shows that the algorithm has better classification effect and is beneficial to enhance the spatial cognitive EEG.Signal identification.