 |
 |
Dear Anderson, Many thanks. The two cognitive variables are correlated and your response is in-line with a colleagues original feeling - unfortunately a reviewer doesn't seem to agree that we don't have to do anything further. Thanks so much for your helpful comments. Best Rebecca Dr. Rebecca Charlton Lecturer in Psychology Department of Psychology Goldsmiths, University of London New Cross London, SE14 6NW UK Tel: + 44 (0)20 7919 7222 Email: [log in to unmask] ________________________________ From: FSL - FMRIB's Software Library [[log in to unmask]] on behalf of Anderson M. Winkler [[log in to unmask]] Sent: 11 February 2013 20:03 To: [log in to unmask] Subject: Re: [FSL] Direct comparison of 2 regression results - TBSS Dear Rebecca, If the two cognitive variables are uncorrelated (and you can check that easily), then the results of the two randomise runs are also uncorrelated. If, however, these variables have some degree of correlation, then the results will have some degree of similarity. In either case, there is no need for further comparisons between the randomise results, but only of these two variables themselves. Does this help? All the best, Anderson 2013/2/11 Rebecca Charlton <[log in to unmask]<mailto:[log in to unmask]>> Dear Anderson, I've previous run two regressions, separately on two cognitive variables (using two column designs). There are some areas where the significant regions overlap and I've been asked to perform a direct comparison to see if the results are "independent". And I don't how to do this ... Does that make sense? Thanks for your help. It's really appreciated. Best Rebecca ________________________________ From: FSL - FMRIB's Software Library [[log in to unmask]<mailto:[log in to unmask]>] on behalf of Anderson M. Winkler [[log in to unmask]<mailto:[log in to unmask]>] Sent: 11 February 2013 19:08 To: [log in to unmask]<mailto:[log in to unmask]> Subject: Re: [FSL] Direct comparison of 2 regression results - TBSS Dear Rebecca, Could you explain in more detail what do you need? If you'd like to compare two randomise runs on the same imaging data, but using different models, then perhaps you can skip randomise altogether and just see which of the two models explains most of the variance. All the best, Anderson 2013/2/11 Rebecca Charlton <[log in to unmask]<mailto:[log in to unmask]>> I've been asked to perform a direct comparison of two regressions performed in randomise. If I prepare my design matrix with three columns, as follows: column one = column of ones, column two = 1st variable, column three = 2nd variable - is this the correct design? And how should I set my contrast? Many thanks for your advice.