 |
 |
Dear SPM experts: I need to know how to interpret the ROIs obtained using CAT12 and SPM12 doing VBM. I use CAT12 and SPM12 to obtain with results function the spmF with the different ROIs so then I use the same spmF with a program called xjview to obtain other ROIs different p-values manually selected different than the options given by SPM Results button. And with all those ROIs I use “atlasquery” terminal output function of FSL and different neuroanatomical atlases(Juelich Histological Atlas, MNI Atlas, Harvard-Oxford Cortical Structural Atlas, Harvard-Oxford Subcortical Structural Atlas, Cerebellar Atlas in MNI152 space after normalization with FLIRT, Cerebellar Atlas in MNI152 space after normalization with FNIRT, Talairach Daemon Labels, ...) to obtain with varying degrees of probability different names of ROIs. For example analysing bipolar plus shizophrenia subjects compared to controls (19 and 19) I obtain* corticospinal tracts, superior longitudinal fasciculus, anterior thalamic radiation, callosal body, cingulum, cerebellum(8, 9,10 lobes), brainstem, amygdala(lateroabasal, superficial), hippocampus(subiculum, dentate gyrus), insula, cuneus, probably some more. Around 25 clusters with xjview. My question goes to how can I interpret which are significant regions in a cluster? Is the maximum peak given by any program (ie. xjview or spm) good for saying that region is associated to that function? How come with one MNI or Talairach coordinate does some author in literature say only one region is associated to one coordinate and not many regions associated probabilistically to one coordinate? Isn’t it better to point many regions to one coordinate to find out all possible regions associated to a trait or disease? Thanks as always for pointing out the answer to my many questions in advance!!! You could answer with references or direct response, as you wish. SIncerely, Andrade. *These are not published results yet but used obtained by databases published on the web NITRC so data is not new but results could be new in some cases. Right or left depends on the case but it doesn’t matter much for the question I am laying here.