Dear, If we use the 'Extract DW Information' tool from the diffusion toolbox, we find a discrepancy in the number of unique gradient directions. We feed 12 image-files to the script containing 4 unique directions and 3 different b-values, using following options: B-value tolerance: 10 Gradient direction tolerance: 1 Distinguish antiparallel gradient directions: No The result shows however 5 unique gradient directions. SPM5: dti_extract_dirs (v1.7) 15:44:46 - 09/02/2010 ======================================================================== Reading diffusion information # unique b values: 3 # unique gradient directions: 5 extracted information saved into workspace variable 'res' allb: [12x1 double] allg: [12x3 double] bfile: [1x96 char] gfile: [1x96 char] ub: [3x1 double] ubi: [3x1 double] ubj: [12x1 double] ugi: [5x1 double] ugj: [12x1 double] ug: [5x3 double] dirsel: [12x12 logical] b: [12x1 double] g: [12x3 double] If we look into the 'res' variable, we see that the same gradient direction is present twice in the 'res.ug' variable. -0,901242247736658 -0,419113115207296 -0,110030030245444 0,109966984346586 0,663801905510129 0,739779894694929 0,109966984346586 0,663801905510129 0,739779894694929 -0,166021035584538 0,986122211362992 0 -1 0 0 Can you clarify this and what are the consequences of this discrepancy? Kind regards, Geert De Groof