 |
 |
1, Why do you need to combined the two Session? Is that necessary? It seems you conjuncted the session as if there were acquired in one session. 2, My opinion is that the first procedure is more feasible. Preproccess in SPM is always done seesion/run by seesion/run. And so I don't think it a good idea to realign(/coregister/normlize...) run B to A. On 11/5/07, aaron qin <[log in to unmask]> wrote: > Our group is focusing on the study and research on attention using > fMRI. In the study, we encounter some puzzled problems with data > analysis. And, unfortunately, no unanimity was achieved. Would you > please provide us your own understanding for reference and your help > will be greatly appreciated. Thanks. > We used acquired a dataset of a Session consisting of TWO runs, A, B, > in which, we follow the same paradigm but different conditions. We > employed two different processing sequences to analyze them. > 1. Realignment- Coregister ¨C Segment - Normalize - Smooth were done, > and the processed images of Run A and B were combined together before > the GLM. > 2. the manipulation that the images of Run A and B were combined > together which is followed by Realignment- Coregister ¨C Segment - > Normalize ¨C Smooth ¨CGLM. > After these , we find huge differences in the activated area, which > were supposed to be SIMILAR or not significantly different. > Could you give us some reasonable explanation mathematically and logically? > -- ºÎºê½¨, He Hongjian BioX laboratory of physics department, Zhejiang University. Zheda Road 38, Hang Zhou, China. 310027 Phone: 86-571-87952811