虚幻面孔加工有效连通网络的非线性DCM分析

Investigation of effective connectivity of the illusory face detection network based on nonlinear dynamic causal models analysis

采用让被试从纯噪声图片中产生虚幻面孔的实验范式来实现面孔top-down加工激活模式的提纯,并使用神经理论基础更为完善的非线性动态因果模型(DCM)来分析top-down方式下虚幻面孔加工有效连通脑网络.所得到的最优脑网络模型表明,枕部面孔区(OFA)在虚幻面孔加工中是关键的面孔信息生成器,它能够被顶下小叶(IPL)施加在其上的top-down注意力所调节,实现在纯噪声图片中检测出类似于面孔的特征信息,然后提供给梭状回面孔区(FFA)作进一步的面孔整体信息加工.

In order to extract the activation patterns of top-down face processing, the present study uses an experimental paradigm in which participants detect illusory faces in pure noise images. The nonlinear dynamic causal models （DCM） analysis, which has a perfect neural theory foundation, is used to investigate the effective connectivity of the illusory face detection network under the top-down processing mechanism. The optimal network model indicates that the occipital face area （OFA） serves as a key generator of illusory face detection. Under directing top-down visual attention exerted by the inferior parietal lobule （ IPL）, OFA searches for the pure noise images for face-like features, and then provides those face-like feature information to the fusiform face area（FFA） for further holistic face processing.