适于自然情境脑成像研究的分析方法：被试间相关及其应用

A New Method of Data Analysis for fMRI Studies in Naturalistic Situation:Inter-Subject Correlation and its Application

被试间相关分析是一种基于大脑活动的时间模式的数据分析方法。该方法通过计算接收相同刺激时被试间脑区活动的一致性,探讨认知加工与脑区功能的关系。与传统的基于激活水平的数据分析方法相比,该方法不需要设置严格的实验条件,能更好地应用于自然情境下的脑成像研究。文章介绍了被试间相关分析的基本原理和方法,分析了该方法如何识别认知功能脑区及其可靠性,并结合其在自然情境脑成像以及特定研究领域的应用,阐明被试间相关在自然情境脑成像研究中的优势,以及该方法在多个研究领域的广泛应用扩展了认知神经科学研究的深度和广度。

The functional Magnetic Resonance Imaging （fMRI） is a powerful and fast growing technique to investigate the human brain function. However, the strict experimental conditions have limited its application in the studies of the naturalistic situation. Inter-Subject Correlation （ISC）, a new data-driven analysis method, has been developed to examine the neural mechanism of cognitive processing in the naturalistic situation. ISC is a new data-driven analysis method and a more ecological approach to conducting fMRI studies in a naturalistic situation. In conventional fMRI studies, the common way to recognize functional locations is identifying active regions under experiment conditions with tasks or stimuli contrast. The basic idea underlying data analysis is to construct a general linear model with experimental events as regressors. This approach depends on strict experimental control and its ecological validity, therefore, is relatively low, On the contrary, ISC computes the correlations of the BOLD signals among subjects along the time course of stimuli presentations. The hypothesis of this approach is that the same or similar psychological events would evoke reliable BOLD signals among participants. When individuals are exposed to the same situation, an Inter-subject synchronization of cortical activities would be observed in comparing with the individual difference of activities for those random events. Thus, ISC can reveal the neural correlates of overlapped cognitive processing among subjects. Most importantly, ISC has been successfully used in naturalistic situations since the experimental conditions are not necessary and since it is a robust data-driven approach in fMRI studies. In the implementation of ISC approach, researchers have developed two methods for identifying function regions. One way is searching for brain areas highly correlated to the time course of stimulus presentation individually, and then applying ISC to determine reliable areas among subjects. The other way is reverse： the ISC was computed first to identify highly correlated regions, and then the nature of those regions is examined by calculating the co-occurrence of increasing BOLD signals and the presenting stimulus. Both ways were proved to be reliable in comparing with the conventional GLM approach. Taking the advantages of the ISC approach, researchers have broadly utilized ISC in studying numerous areas under the naturalistic situation. The significant features of ISC are that it is sensitive to large-scale time events; it can reveal multiple processings mixed at the same time, and it can reveal the neural correlates of context dependent events. Those features are common in reality so that ISC is a powerful method in studying the naturalistic situation. Those advantages have advanced the utilities of ISC in numerous areas including developmental neuroscience studies in which children are not comfortable in strict experimental conditions, and cognitive neuropsychological studies in which most of the patients are not capable of performing the experiment tasks. In summary, this review provides the theory and method, validity and the advantages of ISC in the most up-to-date literature of fMRI studies. It shows that ISC method is a suitable method for fMRI studies in the naturalistic situation. It suggests that the ISC can be widely applied to many research areas, effectively expanding the depth and breadth of studies in cognitive neuroscience. ISC is proposed as a growing trend towards cognitive neuroseienee for more realistic and naturalistie fMRI studies.