Dear FreeSurfer experts,
I have figured out what the problem is, however I do not know how to solve it.
Apparently the homogeneity function within the lme_mass_RgGrow function does not recognize the voxels as homogenous if they have zeros in all columns because the correlation coefficient is NaN.
How do I deal with zero columns in my input file (lh.thickness) which will be carried over into the Theta0 and Re?
Thank you for any suggestions! Clara
----- Weitergeleitete Mail ----- Von: "ckuehn" ckuehn@cbs.mpg.de An: "Freesurfer support list" freesurfer@nmr.mgh.harvard.edu Gesendet: Montag, 3. Juli 2017 18:27:46 Betreff: [Freesurfer] Matlab Problem LME
Dear FreeSurfer experts,
I am following the LME tutorial for the spatiotemporal mass-univariate model on Matlab 8.2. When I get to the point where the covariance estimates are segmented into homogeneous regions I run into a problem. Nothing really happens in Matlab anymore.
This is my command:
[lhRgs,lhRgMeans] = lme_mass_RgGrow (lhsphere, lhRe, lhTh0, lhcortex, 2, 95);
And this is what the Matlab output looks like:
connected to 12 workers. Starting region growing algorithm Computing seeds ...
I've tried this already last week and nothing happened for a few days.
When I terminate the command this is the Matlab output:
Operation terminated by user during kron (after line 36)
In lme_mass_RgGrow>homogenity (line 299) if (sum(sum(DistRgParams <= kron(ones(1,nv),thr),1) == np) >= prc*nv/100) &&...
In lme_mass_RgGrow>merge (line 320) if homogenity(Re,Params,auxRgs,Rkn,nst,prc)
In lme_mass_RgGrow>split (line 269) [Rgs,mrg] = merge(Rgs,Re,Params,mrgRgs(i),mrgRgs(j),nst,prc);
In lme_mass_RgGrow>seeds (line 208) [Rgs,spltf(i),nnRgs] = split(Rgs,Re,Params,coord,splRgs(i),nst,prc);
In lme_mass_RgGrow>vtxwGrowing (line 77) [Rgs1,Rgseed,Rgstd]= seeds(Re(:,maskvtx),Params(:,maskvtx),coord(:,maskvtx),nst,prc);
In lme_mass_RgGrow (line 62) Regions = vtxwGrowing(Re,Theta,sphcoord,AdjM,maskvtx,nst,prc);
I have then copied the function and added a few "fprintf" (see attachment) to get an idea as to where matlab has problems and the output looks like this: Starting region growing algorithm Computing seeds ... point5 point6 point7 point8 point7 point8 point7 point8 point7 point8 point7 point8 point7 point8 point7 point6 point7 point8 point7 point8 point7 point8 point7 point8 point7 point8 point7 [...]
If I understand that correctly, matlab gets hung up on functions 'function tf = homogenity(Re,Params,Rgs,Rn,nst,prc)' and 'function [Rgs,tf] = merge(Rgs,Re,Params,Rn1,Rn2,nst,prc)'. Does anybody have an idea what that means and how it can be fixed?
How long is this step supposed to take? I have 81 subjects with 1 - 3 time points.
Any help or suggestions would be much appreciated! Cheers Clara