Representational similarity analysis (RSA) is a rapidly developing multivariate platform to investigate the structure of neural activities. Similarity/dissimilarity is the core concept of RSA, realized by the construction of a representational dissimilarity matrix, that addresses the closeness/distance for each pair of research elements (e.g., one minus the correlation between the brain responses to 2 different stimuli) and in turn, constitutes a multivariate pattern as its analytic foundation. This approach is also welcome for its sensitivity in detecting subtle differences of distributed experimental effects in the brain. Importantly, RSA is not only an experimental tool but a promising data-analytical framework that can integrate cross-modal imaging signals, explore brain-behavior link, and verify computational models according to measured neural activities. RSA substantiates its integrative power by relating similarity structure in one domain (e.g., stimulus features) to that in another domain (e.g., neural activities). This review summarizes dissimilarity/similarity definition of RSA, introduces how to derive the dissimilarity structure in neural response pattern, and carry out connectivity analysis based on RSA platform. Several recent advances are highlighted, such as the extraction of across-subjects regularity, cross-validation of brain reactivity in human beings and monkeys, the incorporation of computational models and behavioral profiles into RSA. Voxel receptor field modeling, another promising multivariate tool of pattern elucidation, is presented and compared. The application of RSA is expected to surge and extend in many fields of neuro-science, computation, psychology and medicine. We also discuss the limitations of RSA and some critical questions that need to be addressed in future research.
Multivariate pattern analysis（MVPA） is a recently-developed approach for functional magnetic resonance imaging（fMRI） data analyses.Compared with the traditional univariate methods,MVPA is more sensitive to subtle changes in multivariate patterns in fMRI data.In this review,we introduce several significant advances in MVPA applications and summarize various combinations of algorithms and parameters in different problem settings.The limitations of MVPA and some critical questions that need to be addressed in future research are also discussed.
Objective The left-lateralized N170, an event-related potential component consistently shown in response to alphabetic words, is a robust electrophysiological marker for reading expertise in an alphabetic language. In contrast, such a marker is lacking for expertise in reading Chinese, because the existing results about the lateralization of N170 for Chinese characters are mixed, reflecting complicated factors such as top-down modulation that contribute to the relative magnitudes of N170 in the left and right hemispheres. The present study aimed to explore a potential electrophysiological marker for reading expertise in Chinese with minimal top-down influence. Methods We recorded N170 responses to Chinese characters and three kinds of control stimuli in a content-irrelevant task, minimizing potential top-down effects. Results Direct comparison of the N170 amplitude in response to Chinese characters between the hemispheres showed a marginally significant left-lateralization effect. However, detailed analyses of N170 in each hemisphere revealed a more robust pattern of left-lateralization - the N170 in the left but not the right hemisphere differentiated Chinese characters from control stimuli. Conclusion These results suggest that the selectivity of N170 （a greater N170 in response to Chinese characters than to control stimuli） within the left hemisphere rather than the hemispheric difference of N170 with regard to Chinese characters is an electrophysiological marker for expertise in reading Chinese.
