Error in batch_fitting with GridSearchRician

70 views
Skip to first unread message

Blake Neyland

unread,
Oct 15, 2019, 6:26:47 AM10/15/19
to NODDI
Hi Gary and NODDI group,

I am running into a new issue with batch_fitting. I am using matlab 2015a and have the parallel computing toolbox installed. My binary masks are created using BET in FSL. I have attached my mask and bvec/bval files here as a reference if needed. Would this error have anything to do with my data type? I can convert with fslmaths if needed. Currently, the NODDI is FLOAT32.

>> protocol = FSL2Protocol('subject1_dwi.bval', 'subject1_dwi.bvec');
>> noddi = MakeModel('WatsonSHStickTortIsoV_B0');
>> CreateROI('subject1_dwi.nii','subject1_bet_mask_final.nii','noddi_roi.mat');
loading the DWI volume : sub-imove1011_ses-1_acq-apNoddi_run-1_dwi.nii
loading the brain mask : noddi_bet_mask_final.nii
creating the output ROI ...
saving the output ROI as noddi_roi.mat
done
>> batch_fitting('noddi_roi.mat', protocol, noddi, 'FittedParams_imove1011.mat', 8);
143759 of voxels to fit
Error using GridSearchRician (line 60)
Input to EIG must not contain NaN or Inf.

Error in ThreeStageFittingVoxel (line 66)
[x0 liks] = GridSearchRician(Epn, modelname, grid, protocol, constantsGD, sig);

Error in batch_fitting>(parfor body) (line 96)
            [gsps(i,:), fobj_gs(i), mlps(i,:), fobj_ml(i), error_code(i)] = ThreeStageFittingVoxel(voxel, protocol,
            model);

Error in batch_fitting (line 89)
    parfor i=split_start:split_end
noddi_bet_mask_final.nii
subject1_dwi.bval
subject1_dwi.bvec

Hui Zhang

unread,
Oct 15, 2019, 1:55:53 PM10/15/19
to no...@googlegroups.com
Hi Blake,

The issue is that in your bval file, b=0 values are non-zero.  If you take a look at the protocol structure, you will see that it says the number of b=0 volumes is zero.  With NODDI toolbox 1.01, you can resolve this issue by calling FSL2Protocol with an extra parameter like this:

protocol = FSL2Protocol('NODDI.bval', 'NODDI.bvec', 10);

This way, any b-values up to 10 will be rounded off to 0.

Gary


--
You received this message because you are subscribed to the Google Groups "NODDI" group.
To unsubscribe from this group and stop receiving emails from it, send an email to noddi+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/noddi/38f91bf9-37dc-4dab-95e4-7659a7d2e0f2%40googlegroups.com.
Reply all
Reply to author
Forward
0 new messages