| Spatial ability is one of the basic cognitive ability of individuals,especially in some special occupations such as pilots,astronauts and the recently popular field of unmanned aerial vehicles.Spatial visualization,as one dimension of spatial ability,plays an important role in the selection of these special occupations.Mental rotation task,which is a classical paradigm of spatial visualization,can effectively predict the individuals' visual-spatial ability.From the beginning of the 1970 s,Shepard et al.began to study about mental rotation,and with the deepening research,researchers are increasingly concerned about the advanced cognitive function of mental rotation.The development of eye tracking and neuroimaging technology provides more means for the study of cognitive processes included mental rotation.Therefore,more researchers began to combine these technologies to study the mental rotation or other cognitive tasks for multi-dimensional comprehensive analysis.First,in our study,we employed the synchronization methods of eye tracking and EEG,and evaluated their own temporal accuracies,and proposed some reasonable suggestions to correct the clock drifts.We investigated the influencing factors of the 3-D mental rotation processing,and analyzed the correlation between eye movement parameters and the behavioral performance as well as the influencing factors,and then revealed the cognitive process of mental rotation.The synchronization method of eye movement and EEG was employed to further explore the cognitive processes and strategies of mental rotation and brain mechanisms.We have adopted the data-driven model to predict the different mental rotation strategies and revealed the processing states of 3-D mental rotation.The main research contents,results and conclusions of this thesis are as follows:(1)The study on combining methods of eye tracking and EEG and their temporal accuracy measurementsWe constructed four synchronization methods according to the introductions and the characteristics of the two devices,and evaluated the temporal accuracy of each method.It was found that the synchronization method with SDK had the highest temporal accuracy,in which the cumulative temporal error and periodic temporal error can be reduced to the order of 10 ms.This method is suitable for many other research with the synchronization of eye movement and EEG,and we proposed a calibration method for clock drift of the synchronization platform for other researchers.(2)The study on the behavioral and eye-movement features of mental rotationWe examined the influence of the pixel levels on the mental rotation.In addition to the statistical analysis of behavioral performance such as reaction time and accuracy,we also carried out the correlation analysis between the eye movement parameters and different influencing factors,as well as the relationship between eye movement parameters and behavioral performance,and then we revealed the processing stages of mental rotation from the one-dimensional eye movement parameters.It was found that the reaction time of mental rotation was significantly positively correlated with the angle disparity,which was consistent with the previous research conclusion.In addition,the reaction time to the Z-axis-rotated-stimuli was shorter than that of the X-axisrotated-stimuli,and the reaction time to identical stimuli was shorter than that of the mirrored stimuli,indicating that the subjects cognitively processed the Z-axis-rotatedstimuli and identical stimuli faster and easier.Additionally,the higher the stimulus contained the pixel level,the shorter reaction time the subjects had,indicating that the large projected area and visual angle are easier and faster to process.The eye movement characteristics such as fixation count,total fixation duration,total saccade duration was correlated with the reaction time significantly.Through the analysis of the distribution of fixations,we also found that 3-D stimulus had some critical regions,that is,the corner of the upper and lower arms,which determined the critical topology of a 3-D stimulus.(3)The study on the brain mechanism of 3-D mental rotation processesWe applied the synchronization of eye movement and EEG,that is,eye-fixation related potentials,to further examine the processing states and strategies of mental rotation,and determine the importance of different areas of interests in mental rotation.In addition,we combined the eye movement and EFRP parameters to compare the results under the same factors,and analyzed the inherent processes,and then revealed the processing states and strategies of mental rotation based on electrophysiology.It was found that there was no significantly difference in the average fixation durations in the key AOIs and the general AOIs,whereas in EFRP analysis,the average voltages of the P1 component in the right parietal cortex were found to be significant in the two different types of AOIs,that is,the average voltage in the key AOIs is higher than that in the general AOIs,which is probably due to the transfer and fluctuation of attention during cognitive tasks,and the cognitive load of the subjects in the processing in the key AOIs increased to some extent.In addition,it was found that the average voltage of the P1 component in the parietal cortex showed a significant difference between the different types of functional fixations,that is,in the process of “comparing” in the mental rotation tasks,the intensity of visual information processing and the cognitive load increased to some extent.Finally,the results have suggested that the EFRP technology can provide the convergence and differences in evidence to validate the research hypotheses for researchers,and provide an effective solution to study the visual perception and cognitive processing in a more natural experimental environment.(4)The study on eye-movement patterns and processing states of mental rotationThe data-driven models were established to predict two different mental rotation strategies,which can provide experimental validation and data support for the mental rotation.The mental rotation strategies were:(I)left-side fixed and right-side rotated,that is,the subjects fixed the imagery of the left-side figure and mentally rotated the imagery of the right-side figure;(II)right-side fixed and left-side rotated,that is,the subjects fixed the imagery of the right-side figure and mentally rotated the imagery of the left-side figure.We used the data-driven modeling on logistic regression,support vector machine and discriminative hidden Markov models,to carry out experimental analysis to determine the best classification models and parameters.In addition,we further revealed the processing states of 3-D mental rotation through the hidden Markov models.The results showed that the accuracy of hidden Markov models to predict the mental rotation strategies can achieve an accuracy of 76.8%,which is higher than these of other models.What's more,our modeling results revealed three processing states contained in mental rotation tasks: encoding and searching,comparing,searching on one-side pair.These results further confirmed that the mental rotation of 3-D figures is a discrete process in a piecemeal strategy.This thesis aims to reveal the processing states of 3-D mental rotation by combining behavioral performance,eye movement and EEG.The main contributions to this field include:(1)technical basis: the solutions to synchronization of eye tracking and EEG,and their temporal accuracy measurement.(2)experimental factors: the area of the plane projection of the 3-D figures was included,and its influence on mental rotation processing was determined.(3)research methods: the EFRP technology was applied to study the cognitive process of mental rotation,which contains the benefits of eye tracking and EEG.(4)analytical methods: the data-driven models,which predicted the processing strategies and revealed the processing states,provides a new research method,experimental validation and analytical methods for this field. |
PDF Full Text Request