Hi,

Yep, they look fine to me. Perhaps matlab can’t find randomise. You can fix this as explained on the very bottom of this link: https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FslInstallation/MacOsX#Using_FSL_MATLAB_libraries, or potentially by opening matlab from the command line.

Cheers,

Janine

On 18 Jan 2018, at 17:32, Daniela Valério <[log in to unmask]COM> wrote:

Hi Janine, 

I think that these 2 files are okay. However, I'm sending them attached to you can take a look. 

Thanks, 

Daniela

2018-01-18 17:25 GMT+00:00 Janine Bijsterbosch <[log in to unmask]ac.uk>:

Hi Daniela,

It looks like randomise didn’t run correctly. Do your design_10.mat and design_10.con files look appropriate (i.e. do they contain EVs and contrasts)? The ncon variable is set based on the number of randomise outputs (which is the same as the number of contrasts), and should therefore never be empty.

Cheers,

Janine

On 18 Jan 2018, at 17:00, Daniela Valério <[log in to unmask]COM> wrote:

Dear FSL experts, 

I started recently to do functional connectivity with FSL Nets. And, I stopped in an error that I don't have any idea how to solve it.

 When I ran my_nets code, I have this problem:

"Error using figure

Value must be a 4 element vector

Error in nets_glm (line 48)

figure('Position', [100 100 600*ncon 500]);

Error in my_nets (line 72)

[p_uncorrected,p_corrected]=nets_glm(netmats1,'design_10.mat','design_10.con',1);  % returns matrices of

1-p"

In my nets_glm code, I have the "ncon variable" empty. I think that the problem occurs in line 38 in the function randomize.

save_avw(reshape(netmats',XXX,1,1,TTT),fname,'f',[1 1 1 1]);

 system(sprintf('randomise -i %s -o %s -d %s -t %s -x --uncorrp -n %d',fname,fname,des,con,nperms));

% how many contrasts were run?

[grot,ncon]=system(sprintf('imglob %s_vox_corrp_tstat*.* | wc -w',fname));

ncon=str2num(ncon);

if view==1

figure('Position', [100 100 600*ncon 500]); 

end

I tried to diagnose the error and I identified that the temporary image that is necessary to randomize has the size 169 1 1 10, where 169 is my time series dimension and 10 is my number of subjects. 

I don't know if this helps, but I have the FSL version 5.09. 

Thanks in advance for any help!

Best, 

Daniela Valério

-----
Dr Janine Bijsterbosch
Postdoctoral Researcher
FMRIB Centre, University of Oxford
John Radcliffe Hospital
Oxford, United Kingdom
[log in to unmask].uk

<design_10.con><design_10.mat>

-----
Dr Janine Bijsterbosch
Postdoctoral Researcher
FMRIB Centre, University of Oxford
John Radcliffe Hospital
Oxford, United Kingdom
[log in to unmask]ac.uk