 |
 |
Dear SPM list, I have four different PET conditions, say A, B, C, D. I wanted to evidence potential differences between the contast (A-B) and the contrast (C-D) I have used two alternative methods: 1. Differences of contrast e.g (A-B)-(C-D) 2. Mask exclusively the contrast (A-B) by the contrast (C-D)thresholded at 0.05 The results are different because very few voxels survived to the (A-B)-(C-D) difference. On the other hand the exclusive masking exhibited more subtle differences between the contrast. My feeling is that the first method is dramaticlly more conservative and evidence areas which are activated in (A-B) and deactivated in (C-D) The masking method allows to evidence areas that are less activated in (C-D) than in (A-B) or not activated (or deactivated) in (C-D) but activated in (A-B). My question is: what is the correct method ? Note that the difference between (A-B) and (C-D) is likely to be subtle. Thank you very much for your answer Emmanuel ************************************************** Dr Emmanuel Mellet Groupe d'Imagerie Neurofonctionnelle UMR 6095 CNRS, CEA, Univ. Caen & Univ. Paris V GIP Cyceron Bd Becquerel BP 5229 14074 CAEN Cedex phone: +33 (0)2 31 47 02 14 Fax: +33 (0)2 31 47 02 22 **************************************************