Decoding Action Selectivity of Observed Images Using fMRI Pattern Analysis
Previous research indicates that mirror neuron system (MNS), comprising posterior parietal (PPC) and ventral premotor (PMv) cortices, has a role in action understanding. However, action selectivity within these regions has notbeen elucidated. We investigated whether the MNS contains information about observed actions. Participants observed objectdirected hand actions. We independently manipulated five dimensions of observed images: action (grasp/touch), object (cup/bottle), perspective (egocentric/allocentric), hand (right/left), and image size (large/small), and investigated whether this information could be decoded from multi-voxel pattern activity. The early visual area showed significant above-chance classification accuracy consistent with dissimilarity of input images. In contrast, significant decoding accuracy was observed for action, object, and mirror-image of hand in the PPC and for action and object in the PMv. Our study indicates that the MNS processes observed actions in a hierarchical manner, where the PPC represents action together with hand, and the PMv encodes more abstracted representations of transitive action.
Action observation Posterior parietal cortex Premotor cortex fMRI
Kenji Ogawa Toshio Inui
ERATO Asada project, Japan Science and Technology Agency, Kyoto 606-8501, Japan
国际会议
The Second International Conference on Cognitive Neurodynamics--2009(第二届国际认知神经动力学会议)
杭州
英文
129-134
2009-11-15（万方平台首次上网日期，不代表论文的发表时间）