The Effect And Evaluation Method Of Virtual Reality Games On Working Memory

Since the game elements were developed initially,the application scenes have gradually expanded from the commercial fields to the academic research fields.It is unquestionable that people are attracted by game elements.However,there is also a lack of empirical studies examining the effects of game elements on behavior and performance of users.One of the most important cognitive abilities is working memory.It is directly related to people's reading ability,reasoning ability and fluid intelligence.Therefore,as the scope of gamification is gradually expanding,it is necessary to study whether gamification or game elements have a beneficial effect on working memory.Moreover,with the development of virtual reality technology,the game has a richer form of expression than before.It is also a question that whether the VR games with highly immersive will have a beneficial effect on working memory,and whether ordinary 3D games and VR games have the same effect on working memory.The use of VR games to qualitatively or quantitatively study working memory is still missing,this thesis uses VR games as the material of working memory experiments.The thesis explores the impact of virtual reality games on working memory through the results which includes EEG signal data analysis and behavioral data analysis based on the contrast experiments.The main contents of this thesis are as follows:(1)Using the classical evoked paradigm of working memory,the N-back paradigm,as the material of test experiment.The EEG signals of the participants in the paradigm task are collected,and the horizontal classification standard and classifier of the working memory are obtained.The classifier will be used to predict the working memory level classification label of the participants in different game modes in the control experiment.(2)Using the 3D version and the VR version of the same game as the material to design comparison experiment.The experimental results were used to explore the differences in working memory performance of the participants in the 3D game environment and in the VR game environment,and verify whether the hypothetical theory proposed by the research is correct.(3)For the data processing,we will perform denoising,feature extraction and classification on the P300 EEG signals collected by the experiment,and use the SVM classifier based on machine learning and the two-way LSTM based on deep learning to build the classifier and compare the classification accuracy.Among them,the SVM classifier achieved average classification accuracy of 88.28%,and the bidirectional LSTM classifier achieved average classification accuracy of 92.66%.(4)Combining behavioral data with EEG classification results data,this paper analyzes the performance of participants in different game modes,and verify the influence of VR games on the working memory level of participants through statistical analysis methods such as paired T test,and the results show that VR games have a beneficial effect on working memory.This study obtained the following conclusions by processing and analyzing the EEG signals of the participants.Compared to the 3D environment,participants have better game performance in VR mode and can maintain a high working memory level for a longer period of time.This conclusion can provide some theoretical support for the application of gamification to working memory training.