Hello,

I want to do ROI analysis for each patient, but i have a problem. While creating for each patient's output roi file's (ex: ROI_Table_20160225T200429.tbl) values (size, average, std. dev., etc) are same. My image format is "smwc.nii".

Why this is happening?

Thanks,
A.BeriL

Hello A.BeriL,

I assume you are going through the Generate Table.  Are you supply a file list or add files manually?  Is each tbl done with only one subject?  What version of PA, Matlab, and what OS are you using?

Ben

Benjamin,

Thank you so much for your quick response :)

When i clicked "Generate Table", i selected with only one subject analyze image file so i chosen .smwc file extention and added files manually.

I have one more question.. I used "aal" human anatomy atlas and i want to save mask. If i choose all regions and click save mask, is it any problem about saving mask?

My Matlab version is 7.10.0 (R2010a) and i used SPM12.

Best,
A.BeriL

Hello again,

I am not familiar with the smwc extension.  SPM/PickAtlas requires nifti files in either .nii or .img/.hdr pairs.

Question 2: Nope.

Ben

Hello again Ben,

Thank you very much for your response.

I'm sorry, I think i said wrong to you about file extension. Smwc files are smoothed images and their extensions are .nii

I'm sending prtsc photo from my pc.

I'm trying again extracting ROI files today but again all values are same. i selected with only one subject analyze image file...

I can't solve this problem. I would appreciate if you can help...

Best,
Beril

Hi Beril,
  Thanks for the graphic and clarification.  Any chance you could send me a copy of your output tbl and flist?  If you don't want to send the whole thing you could truncate it to the first 10 or 20 lines.  By way of some clarification, the fist column (size) should be the same for each ROI region.  However, that does not explain why Avg., SD, and T are the same.

Thanks,
Ben

Hi Ben,

I'm sorry for the late reply.

I'm sending to you for two different patients output tbl and flist. As we talked about before, all values are same. I hope that tbl and flist are more clarification about this situation..

Hi Beril,
  I'm not seeing the problem.  See the attached image.  The dark blue are differences.  Yes, the Size is the same.  That is the size of the ROI.  But the Average (0.29 v/s 0.32) and STD (0.15 v/s 0.12) are different.  This happens for the rest of the file.

Ben

Hi Ben,

Really thank you very much for your care and attention :)

The problem has been solved.

Thank you, again...

Best,
BeriL

Hi Ben,
 
Me again :)

I know max and min location's unit are mm. What are units average and std. deviation in the ROI table (mm^3 or cm^3)?

Thank you,

Best,

BeriL

Hi BeriL

mm^3

Ben

Hi Ben,

Thank you for your quick response..

I want to ask you a question finally.. I hope. What is value of "average" and "std.deviation" the ROI table? Volume or voxel?

I apologize for the inconveniences..

Thank you so much Ben,

Best,
BeriL

Hi BeriL,
  I believe I mis-spoke yesterday.  Avg and STD are not mm^3 or cm^3.  They are the mean and standard deviation across the values in the voxels of the mask.  If the image a volumetric then you can get average volume by taking average times the product of the voxel sizes.

Hope this helps,
Ben

Hi Ben,

I am a bid confused, I am working on a single brain image, but I get "average" volume from mask. Shouldn't I get a special volume for each ROI? 

I understand from your reply that to find a ROI's volume, I must product the "size column" and "average column" in the corresponding line, mustn't I ? 

your sincerely,
BeriL

Hello again BeriL,

The average is the average of the VOXELS in the ROI.  We don't do a volume as the decision to make the value volumetric depends on the type of input image.

For an image where creating a volumetric response is appropriate, then you would multiple the value by the size of the voxel.

Using he size*volume approach for quick and dirty numbers shouldn't be a problem, but I wouldn't base a paper on them. To get total I would save the mask and reslice it in matlab/spm.  Then mask the image and sum the remaining values.  Reasoning: taking the size and average column is not advisable as the size is the number of voxels in the ROI image, not your image.  The standard PA Atlas is 2mm iso.  If your image is of a different size then  some sort of interpolation will happen to get  value at the ROI voxel location.  Results would be similar, but not the same.