A Data-driven fMRI Analysis Method Using Temporal Clustering Technique and an Adaptive Voxel Selection Criterion

摘要: A data-driven fMRI (functional magnetic resonance imaging) analysis method is proposed for studying brain responses to stimulation when the in- formation predicting their onset or duration un- available. The suitable experiments involving a single event non-repetitive multiple events. built upon pre-existing temporal clustering techniques with addi- tional features that make use of signal changes neighbouring voxels ensure selected response detection are those most likely have been activated by stimuli. For val- idation, eight sets data from three dierent kinds sensory applied. results demonstrated our able detect time bins during which stimuli were administered and identified corresponding areas, typically in this kind experiments. Moreover, these data, accuracy 75% compared 58.33% without selection criterion.