The Study Of Decision-Making Mechanism Of Human Based On Spatial Selective Attention

Humans make decisions according to the information in the environment,and rewards related to event influence human decision-making behavior as well as probabilities,but the combined impact of rewards and probabilities on human decision-making behavior is still unclear.This thesis applied method of brain cognitive research in the study of decision-making mechanism to reveal the decision-making mechanism of human under the influence of rewards and probabilities from the perspective of brain cognition.Due to the high similarity and the intrinsic relationship between spatial selective attention and decision-making,it is possible to study the decision-making mechanism by spatial selective attention.Which includes the study of detection method of spatial selection attention(neurological foundation),modeling method of decision-making mechanism(theoretical foundation)and the experimental method of decision-making(data basis)based on spatial selection attention were carried out in this paper.Furthermore,the experimental data was modeled by theoretical method so that allocation strategy of spatial selective attention and the decision-making mechanism of subject were obtained.The main work and contributions of the thesis are as follows:(1)The neurophysiological foundation of the thesis was established in this part by studying the detection method of spatial selection attention.Event-Related Potentials related to attention and Steady State Visual Evoked Potential(SSVEP)were selected to measure spatial selective attention.Thesis firstly clarified the effects of different stimulator in the study of spatial selective attention.For the detection of steady-state visual evoked potentials,the grey correlation theory was applied to the feature extraction of SSVEP for the first time.Furthermore,a correlation analysis method based on supreme resolution spectral estimation was proposed and the recognition rate of SSVEP signal was increased to 94%.Compared with canonical correlation analysis,the proposed method has a 2%-3% increase over time.(2)The theoretical foundation of the thesis was established by studying the relevant modeling methods of decision-making mechanism.Focusing on modeling the decision mechanism under the influence of reward and probability,a decision model based on Exponentially Weighted Moving-Average model was constructed firstly.Based on EWMA modeling,a Q-learning based model was constructed and applied for the first time in the research of decision-making mechanism based on spatial selective attention.(3)Decision-making experiment based on spatial selection attention was carried out and data foundation of the thesis was acquired by collecting the data of the experiment.A binary visual discrimination experiment under the dual influence of stimulus reward value and stimulus probability was carried out in the thesis.Moreover,a research of decision-making mechanism based on SSVEP signal was proposed in the thesis.The experimental results showed that the SSVEP signal can be used to measure the spatial selective attention due to the high correlation between them,therefore SSVEP could be used in the study of decision-making mechanism.(4)Experimental data was modeled by the model method in the previous study according to computational neuroscience.Results showed subjects prefer the equalization strategy to allocate spatial attention and the decision-making mechanism conformed to the Q-learning model mechanism.Furthermore,a hybrid decision-making model based on psychological factors was proposed and the analysis of subjects in different gender subjects based on this model showed the similarities and differences of the decision-making mechanisms between different gender subjects.In summary,this thesis studied the human decision-making mechanism based on spatial selective attention from the perspective of brain cognition.The results showed that under the dual influence of probability and reward,the subject chose the alternative option with the match strategy and in line with the hybrid decision-making model based on psychological factors.The differences in decision characteristics of different genders were reflected by the model parameters.