 |
 |
Hi Experts, I am at the PEB stage of my DCM analysis and am encountering a matlab error (attached). The covariates I am using are at the bottom of this email. If I remove the covariates modelling for each individual subject (M.X(:,4) to M.X(:,14)), the PEB runs. My concern with running this separately is we are using a repeated measures design and with the modified covariates SPM will treat it as cross-sectional. Does it matter to SPM if it is a longitudinal or cross-sectional design or is the maths and stats the same either way? The analysis to date has been done on an older release of SPM12 (6906). To investigate the issue further, all 14 covariates have been run on the latest version of SPM12 (7487) which did work. However, the results are now very different to those obtained when removing covariates 4-14 using the old release – I have attached screenshots of both for the same covariate. Furthermore, when I load the new results using the old release version, I get a different set of result altogether. Any help or advice would be greatly appreciated Best, Sean Covariates: M.X = ones(N,14); % where 14 is the number of M.X variables. Edit this to accomodate for the actual number in your design M.X(:,2)=repmat([-1; 1],N/2,1); % pre < post M.X(:,3)=(repmat([1;1;-1;-1],N/4,1)).*(repmat([-1; 1],N/2,1)); % anodal > cathodal x pre < post NB. this assumes the following order anodal pre - anodal post - cathodal pre - cathodal post. Edit if this is not the order % these are your participants M.X(:,4) = [repmat(1,4,1);zeros(N-4,1)]; %c02 M.X(:,5) = [zeros(4*1,1);repmat(1,4,1);zeros(N-4*2,1)]; %c03 M.X(:,6) = [zeros(4*2,1);repmat(1,4,1);zeros(N-4*3,1)]; %c04 M.X(:,7) = [zeros(4*3,1);repmat(1,4,1);zeros(N-4*4,1)]; %c05 M.X(:,8) = [zeros(4*4,1);repmat(1,4,1);zeros(N-4*5,1)]; %c07 M.X(:,9) = [zeros(4*5,1);repmat(1,4,1);zeros(N-4*6,1)]; %c11 M.X(:,10) = [zeros(4*6,1);repmat(1,4,1);zeros(N-4*7,1)]; %c12 M.X(:,11) = [zeros(4*7,1);repmat(1,4,1);zeros(N-4*8,1)]; %c14 M.X(:,12) = [zeros(4*8,1);repmat(1,4,1);zeros(N-4*9,1)]; %c15 M.X(:,13) = [zeros(4*9,1);repmat(1,4,1);zeros(N-4*10,1)]; %c16 M.X(:,14) = [zeros(4*10,1);repmat(1,4,1);zeros(N-4*11,1)]; %c17 Sean Coulborn PhD Candidate 2.30 Hills Building School of Psychology University of Birmingham Birmingham, B15 2TT United Kingdom