Hi Paola, in the following I assume that you have exactly four groups - and not four patient groups plus one group of healthy controls. Please correct me if I am wrong.
The mass-univariate tutorial also has four groups (HC, sMCI, cMCI, AD) but you will always only see three regressors - three for the group effects, three for group-by-time effects, and three for group-by-time-squared effects: all of these are only stated for the sMCI, cMCI, AD). However, the HC group is not omitted from this model, but chosen as a reference group, and hence implicitly modeled. Assuming exactly four groups in your case, you'd need to do the same, i.e. choose one of your groups as a reference group. This means that you should only have three group regressors, and three group-by-time interaction regressors. I hope that I've understood your issues correctly - otherwise let us know. Best regards, Kersten -----Original Message----- From: Valsasina Paola <valsasina.pa...@hsr.it<mailto:valsasina%20paola%20%3cvalsasina.pa...@hsr.it%3e>> Reply-to: Freesurfer support list <freesurfer@nmr.mgh.harvard.edu> To: Freesurfer support list <freesurfer@nmr.mgh.harvard.edu<mailto:freesurfer%20support%20list%20%3cfreesur...@nmr.mgh.harvard.edu%3e>> Subject: [Freesurfer] strange output from lme analysis - Max r value=inf Date: Thu, 25 Jan 2018 11:05:42 +0100 Dear all I am using the lme toolbox to perform a longitudinal comparison of cortical thickness among 4 patient groups. Following the lme tutorial, I concatenated all files and I built a design matrix with following columns: -one column for intercept (all ones) -one column for time -four columns for each group (1 for subjects belonging to the group, zero otherwise) -four columns for group by time interaction I then built appropriate contrasts to test the effects of time, group and time x group effect; e.g., to test time x group interaction between group 1 and group 2, I built the following contrast: CM.C=[0 0 0 0 0 0 1 -1 0 0] However, I obtain quite strange results, because I get a lot of very small, very significant clusters of difference (having max t value = inf or –inf). Do you have any idea of the reason why I am getting this strange output? In your opinion, is there anything wrong in my design matrix or in the way I defined the contrasts? This is the first time I am trying to compare four groups in the same LME models, I previously ran LME statistics including just one or two groups, and everything ran smoothly. Please see a snapshot of my output statistic below: # ClusterNo Max VtxMax Size(mm^2) TalX TalY TalZ NVtxs Annot 1 -inf 112586 0.96 -19.0 -95.2 8.3 1 lateraloccipital 2 -inf 93693 0.89 -22.0 54.4 16.4 1 rostralmiddlefrontal 3 inf 29207 0.86 -35.9 46.6 -9.9 1 parsorbitalis 4 inf 158564 0.78 -13.7 -98.8 2.3 1 lateraloccipital 5 -inf 67614 0.69 -8.3 35.1 25.4 1 superiorfrontal 6 -inf 123144 0.69 -40.3 19.5 29.8 1 caudalmiddlefrontal 7 -inf 2273 0.68 -55.6 -30.9 24.2 1 supramarginal 8 -inf 140663 0.67 -6.2 39.1 -4.4 1 rostralanteriorcingulate 9 -inf 5857 0.64 -35.5 8.6 50.8 1 caudalmiddlefrontal 10 inf 87756 0.61 -14.1 -98.2 1.7 1 lateraloccipital 11 inf 95576 0.56 -43.9 24.5 27.5 1 rostralmiddlefrontal 12 -inf 112122 0.56 -29.0 -62.8 35.9 1 inferiorparietal 13 inf 54800 0.54 -18.7 -59.3 2.6 1 lingual 14 -inf 140378 0.53 -19.0 -42.5 62.6 1 superiorparietal 15 inf 99044 0.49 -43.7 -20.0 52.6 1 postcentral 16 inf 119372 0.47 -52.1 -2.5 25.5 1 precentral 17 -inf 125951 0.47 -44.3 -55.1 12.7 1 inferiorparietal 18 inf 8776 0.46 -57.0 -2.0 11.7 1 precentral 19 inf 10137 0.44 -27.8 -49.8 -1.9 1 lingual 20 inf 20153 0.42 -6.6 13.8 29.5 1 caudalanteriorcingulate 21 -inf 19380 0.40 -61.6 -8.6 13.7 1 postcentral 22 -inf 38802 0.39 -23.9 -56.1 49.0 1 superiorparietal 23 inf 53936 0.38 -45.3 -25.9 46.7 1 postcentral 24 inf 45406 0.35 -38.8 -36.8 36.7 1 supramarginal 25 -inf 46213 0.32 -6.4 -47.6 15.7 1 isthmuscingulate 26 inf 41829 0.26 -61.1 -8.7 10.5 1 postcentral 27 -13.032 76609 0.40 -14.4 -15.8 64.7 1 precentral 28 12.272 78924 0.27 -6.0 -6.5 30.3 1 posteriorcingulate 29 -10.960 822 0.63 -51.8 -30.1 -17.2 1 inferiortemporal 30 10.326 96572 0.48 -31.7 11.6 0.5 1 insula 31 -8.509 43473 0.54 -62.2 -26.3 6.7 1 superiortemporal .. continued with a lot of significant clusters, all 1 vertex in size .. Thank you for any suggestion Best regards Paola Rispetta l’ambiente: non stampare questa mail se non è necessario. Respect the environment: if it's not necessary, don't print this mail. _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
