The files look fine.

On 13 July 2015 at 09:23, ali javanray <[log in to unmask]> wrote:

Hi Anderson,

Thanks for your guidance, I will put them in a txt file instead.

Were the design matrix fine, isn't there any problem with the .mat & .con files? 

Did you check the files I have attached to the previous mail,weren't they the problem?

Thanks a lot for your clarification.

Javanray. 

On Mon, Jul 13, 2015 at 12:26 PM, Anderson M. Winkler <[log in to unmask]> wrote:

Hi Javanray,

For some reason, of the 70 inputs, fewer are reaching the step in which randomise is performed. If you're copying/pasting from a different application, maybe make sure that the file names contain no spaces, and that there are no hidden "newline" characted that could cause the command to begin execution before all 70 have been entered.

Or perhaps better, put all 70 names in a text file, then instead of listing all 70 files when calling dual_regression, use $(cat filelist.txt).

All the best,

Anderson

On 12 July 2015 at 11:29, ali javanray <[log in to unmask]> wrote:

2)  dual_regression /home/fsl/MELODIC_output/melodicoutput.gica/groupmelodic.ica/melodic_IC.nii.gz 1 /home/fsl/GLM/GLM.mat /home/fsl/GLM/GLM.con 5000 grot (in this part I addressed the filtered_Func data  that was too many so I didn't bring it here.)

after creating common mask and doing the dual regression it goes to sorting maps and running randomise but in this part the error came up.

3)   fslinfo /home/fsl/Documents/DataSet/UM_1/0050272/session_1/rest_1/rest_melodicoutput.ica/reg_standard/filtered_func_data.nii.gz

data_type      FLOAT32

dim1           45

dim2           54

dim3           45

dim4           297

datatype       16

pixdim1        4.000000

pixdim2        4.000000

pixdim3        4.000000

pixdim4        2.000000

cal_max        0.0000

cal_min        0.0000

file_type      NIFTI-1+

There were 70 address of data that I gave to the dual regression code that I just did fslinfo for this one ofcourse all the files are in the same size.

Thanks a lot for your guidance and reply,I would kindly appreciate that.

On Sun, Jul 12, 2015 at 12:03 PM, Anderson M. Winkler <[log in to unmask]> wrote:

Hi Javanray,

Could you please do the following:

1) Attach to the email the design.mat and design.con files.

2) Paste the actual command line that you are using to invoke randomise.

3) Run the command fslinfo on the same file used as input to randomise (that is, the file used with the option -i), then paste here the output.

Thanks.

All the best,

Anderson

On 12 July 2015 at 05:26, Mohammadali Javanray <[log in to unmask]> wrote:

Dear FSL experts,
Hi,

I have designed a design matrix by GLM within these specifications: Higher level/non timeseries design, inputs:70, number of main EVs:2 (1.Controls &        2.Autism),contrast & F-tests: (contras:2, F-tests:0, C1:Controls>Autism  & C2:Autism>Controls) and of course I filled the 0,1,-1 where they needed to be filled.
The problem is: when I use the ".con" & ".mat" GLM outputs in dual regression, during randomise at "drD" step   I get this error: (" number of rows in design matrix doesn't match number of "time points" in input data!")
I've got 70 data: 35 Autism, 35 Controls.
what should I do to make the number of rows match the number of input time points?

Any clarification would be fully appreciated.
Javanray.