This study is focused on two aspects: one is the cognitive psychological experiment design; the other is the basic algorithm research in the field of information processing. The main work of this study is as follows: (1) To propose an fMRI experiment scheme to find the gender differences when human exposing to the standard emotion visual stimuli. This study adopts the recent often-used stimuli (International Affective Picture System, IAPS) as the stimuli set and using the BOLD-fMRI to find the gender differences when subjects exposed to external stimuli. This study is an event-related study with healthy subjects. The results may help explain the fact the women have a higher depressing ratio in the society. (2)To propose and implement an fMRI brain signal detection algorithms framework. This method is a functional feature embedding space mapping based on the spectral analysis which is mainly focused on the block design experiment. We have adopted a non-linear feature extraction algorithm during the embedding space computation which has not been reported in the related areas. At the end of this chapter, artificial data and real human motor data are applied to validate this method of which the non-linear algorithm and traditional method are compared to see which one is better. The experimental results shows that the non-linear method shows a better performance than the traditional one and successfully find the geometry structure of fMRI data in the spatial domain. (3) To propose an ERP experiment scheme to find how the brain works when exposing to the standard emotion stimuli. This study adopts the recent often-used stimuli (International Affective Picture System, IAPS) as the stimuli set and using the ERP (made in EGI) to find the brain cortex spatio-temporal activation patterns when subjects exposed to positive, negative and neutral stimuli. (4) To propose and implement a multimodality image fusion framework and algorithms based on ICA. Different image modality has different characters: fMRI has a higher spatial resolution, while ERP has a higher temporal resolution. How to combine these two modalities to exert their virtues is a meaningful topic. We have proposed a multimodality imaging fusion framework based on ICA which is called FC-ECD. First, we use ICA to exclude the noise and find the interested information. Then for these information, we can localize these information using fMRI constraint which have saved much computational cost. Finally, artificial and real human experiment data are applied to validate this method. This method has successfully localized P200, P300 and P400 which are in accordance with the results reported in the previous study.
修改评论