Hi Elin, it could be several things. First of all, when you add age you are changing the number of regressors from 16 to 32, and so your DOF drops from 84 to 68. 68 is still pretty good, but you are certainly loosing some power. Try running it as a DOSSThis changes the number of regressors from 16 to 17. (change your contrast accordingly).

doug

On 08/09/2013 11:41 AM, Rees, Elin wrote:

Hi,

I'm trying to run a cortical thickness group comparison (HD v Controls), controlling for site, gender (1 or 2) and age. Without age as a covariate the results are as expected but once age is added on all significant results disappear. I think this is a problem with the method rather than because all results are dependent on age. My fsgd and contrast files for the glm are below. I would really appreciate your input as to what might be going wrong to effect the results so strongly. Am I doing something wrong?

GroupDescriptorFile 1 Title HDvControl

Class HD-london-1 Class HD-leiden-1 Class HD-paris-1 Class HD-ulm-1 Class HD-london-2 Class HD-leiden-2 Class HD-paris-2 Class HD-ulm-2 Class C-london-1 Class C-leiden-1 Class C-paris-1 Class C-ulm-1 Class C-london-2 Class C-leiden-2 Class C-paris-2 Class C-ulm-2

Variables Age Input 30009 HD-paris-2 26 Input 30014 C-ulm-2 51 Input 30016 C-london-1 48 Input 30024 C-leiden-1 38 Input 30029 HD-london-2 66 Input 30030 HD-london-1 54 Input 30037 HD-london-2 42 Input 30041 C-london-1 43 Input 30042 HD-london-2 40 Input 30043 HD-london-2 41 Input 30047 HD-london-2 66 Input 30051 C-london-2 45 Input 30054 HD-london-2 55 Input 30056 C-london-2 66 Input 30060 C-london-1 50 Input 30068 C-london-1 51 Input 30071 HD-london-1 43 Input 30072 HD-london-2 44 Input 30074 C-london-2 58 Input 30075 HD-london-2 48 Input 30078 C-london-2 53 Input 30079 HD-london-1 58 Input 30082 C-london-2 62 Input 30089 HD-london-2 54 Input 30093 HD-leiden-1 44 Input 30098 HD-leiden-2 45 Input 30100 HD-leiden-2 42 Input 30102 C-leiden-2 48 Input 30107 HD-leiden-2 61 Input 30112 HD-leiden-2 50 Input 30114 C-leiden-1 48 Input 30115 HD-leiden-2 47 Input 30117 HD-leiden-2 49 Input 30118 C-leiden-1 51 Input 30119 HD-leiden-2 59 Input 30124 C-leiden-1 57 Input 30125 HD-leiden-2 39 Input 30136 C-leiden-2 51 Input 30138 C-leiden-2 56 Input 30139 HD-leiden-2 33 Input 30142 HD-leiden-2 32 Input 30148 HD-leiden-1 48 Input 30154 C-leiden-1 47 Input 30155 HD-leiden-2 46 Input 30160 HD-leiden-2 38 Input 30162 HD-leiden-2 52 Input 30163 C-leiden-2 42 Input 30168 C-ulm-2 58 Input 30177 C-london-2 49 Input 30185 C-ulm-2 34 Input 30189 HD-ulm-1 50 Input 30191 C-ulm-1 44 Input 30193 HD-ulm-1 46 Input 30195 C-ulm-2 50 Input 30198 C-ulm-1 56 Input 30199 HD-ulm-1 53 Input 30208 HD-ulm-1 59 Input 30209 C-ulm-2 44 Input 30210 HD-ulm-1 30 Input 30214 HD-ulm-2 57 Input 30215 HD-ulm-1 34 Input 30220 HD-ulm-2 54 Input 30222 HD-ulm-1 64 Input 30223 C-ulm-1 44 Input 30229 HD-ulm-1 40 Input 30235 HD-ulm-1 48 Input 30240 HD-ulm-1 36 Input 30241 C-paris-2 65 Input 30247 HD-paris-1 24 Input 30248 HD-paris-1 44 Input 30250 HD-paris-2 62 Input 30254 HD-paris-1 37 Input 30256 HD-paris-1 23 Input 30259 HD-paris-1 64 Input 30265 C-paris-2 50 Input 30269 C-paris-2 58 Input 30273 C-paris-2 49 Input 30275 C-paris-1 59 Input 30280 HD-london-2 42 Input 30281 HD-london-2 65 Input 30288 HD-london-2 51 Input 30311 HD-paris-1 40 Input 30312 HD-paris-2 51 Input 30314 HD-paris-2 46 Input 30315 HD-leiden-2 67 Input 30318 C-paris-2 64 Input 30320 C-paris-1 49 Input 30321 HD-paris-1 61 Input 30326 C-paris-2 45 Input 30334 HD-london-1 49 Input 30347 C-paris-1 62 Input 30349 HD-paris-2 48 Input 30352 C-paris-2 66 Input 30357 C-ulm-1 28 Input 30361 HD-paris-2 63 Input 30363 HD-ulm-1 51 Input 30365 C-ulm-2 46 Input 30369 HD-ulm-2 45 Input 30378 C-leiden-1 53 Input 30385 HD-leiden-2 61 Input 30392 HD-ulm-2 51

Contrast: 1 1 1 1 1 1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Thanks, Elin

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer