I don’t know much about ALFF, but probably you would want intensity bias normalized data, rather than variance normalized data.
Peace, Matt. From: erik lee <[email protected]<mailto:[email protected]>> Date: Thursday, April 12, 2018 at 4:31 PM To: "Harms, Michael" <[email protected]<mailto:[email protected]>> Cc: Matt Glasser <[email protected]<mailto:[email protected]>>, "[email protected]<mailto:[email protected]>" <[email protected]<mailto:[email protected]>> Subject: Re: [HCP-Users] Grand Mean Intensity Normalization Thank you both again for your reply. Right now, I am calculating ALFF and comparing it to functional connectivity. Global intensity normalization is probably a good processing step for ALFF, but I had hoped to also look at ALFF without any sort of intensity normalization. >From what you are saying though, it sounds like this might be more effort than >its worth… Best, Erik On Apr 11, 2018, at 6:45 PM, Harms, Michael <[email protected]<mailto:[email protected]>> wrote: We consider those “intermediate” files and thus they aren’t part of our released packages or “CinaBox” disks. So, this could be rather challenging. You would need to use “REST” calls into the ConnectomDB database directly to pull down the files for each run. Perhaps you could describe why you need this scaling factor, and we can see if we have any further insight. Cheers, -MH -- Michael Harms, Ph.D. ----------------------------------------------------------- Associate Professor of Psychiatry Washington University School of Medicine Department of Psychiatry, Box 8134 660 South Euclid Ave. Tel: 314-747-6173 St. Louis, MO 63110 Email: [email protected]<mailto:[email protected]> From: "Glasser, Matthew" <[email protected]<mailto:[email protected]>> Date: Wednesday, April 11, 2018 at 5:35 PM To: erik lee <[email protected]<mailto:[email protected]>>, "Harms, Michael" <[email protected]<mailto:[email protected]>> Cc: "[email protected]<mailto:[email protected]>" <[email protected]<mailto:[email protected]>> Subject: Re: [HCP-Users] Grand Mean Intensity Normalization Jacobian wasn’t used for the 3T fMRI data. Also the bias field correction was non-optimal. These are the files that you would want: 1) ${StudyFolder}/${Subject}/${fMRIName}/BiasField.2.nii.gz 2) ${StudyFolder}/${Subject}/${fMRIName}/rfMRI_REST1_LR_nonlin.nii.gz 3) ${StudyFolder}/${Subject}/${fMRIName}/rfMRI_REST1_LR_nonlin_norm.nii.gz I expect you will find the normalization factor will be correlated with head size and be related to the amount of coil loading each subject has. Keep in mind the same normalization factor was applied to all voxels. Matt. From: erik lee <[email protected]<mailto:[email protected]>> Date: Wednesday, April 11, 2018 at 5:05 PM To: "Harms, Michael" <[email protected]<mailto:[email protected]>> Cc: Matt Glasser <[email protected]<mailto:[email protected]>>, "[email protected]<mailto:[email protected]>" <[email protected]<mailto:[email protected]>> Subject: Re: [HCP-Users] Grand Mean Intensity Normalization Hi Matt and Michael, Thank you both for your helpful replies. If I wanted to back-calculate the original mean of the rfMRI image after bias field and Jacobian correction, do you know what the appropriate files would be to do this? Looking at the HCP data/intensity normalization script, I wasn’t exactly sure what the file names would be for the following: 1. Input rfMRI (pre-Jacobian/Bias Field correction) 2. Bias Image 3. Jacobian Image 4. Output rfMRI (post-Jacobian/Bias Field correction and global intensity normalization, but before any temporal processing) Presumably once I have these, I could do some simple algebra at any given voxel to find the global mean used in the intensity normalization. Thanks, Erik On Apr 11, 2018, at 3:43 PM, Harms, Michael <[email protected]<mailto:[email protected]>> wrote: Just to expand on this, since I think I might know why you are asking. The grand mean is computed on the brain masked volume timeseries, after bias field correction and jacobian modulation is first applied – see the end of IntensityNormalization.sh, which is called as the final step in GenericfMRIVolumeProcessingPipeline.sh There is NOT another grand mean normalization applied specifically to the CIFTI data, so don’t expect the CIFTI data to have a grand mean of 10000. IIRC, the grand mean of the CIFTI timeseries tends to end up in the 8000-9000 range. Cheers, -MH -- Michael Harms, Ph.D. ----------------------------------------------------------- Associate Professor of Psychiatry Washington University School of Medicine Department of Psychiatry, Box 8134 660 South Euclid Ave. Tel: 314-747-6173 St. Louis, MO 63110 Email: [email protected]<mailto:[email protected]> On 4/11/18, 2:31 PM, "[email protected]<mailto:[email protected]> on behalf of Glasser, Matthew" <[email protected]<mailto:[email protected]> on behalf of [email protected]<mailto:[email protected]>> wrote: 1) The overall mean of each scan is 10000, this is not done voxelwise (e.g. like a bias correction would be). 2) Unfortunately this information is not saved. I don¹t think fslmaths outputs it, perhaps it could be back computed from some intermediate files. Peace, Matt. On 4/11/18, 2:21 PM, "[email protected]<mailto:[email protected]> on behalf of erik lee" <[email protected]<mailto:[email protected]> on behalf of [email protected]<mailto:[email protected]>> wrote: >Dear HCP Experts, > >I am currently using the temporally preprocessed rfMRI data in the S900 >release (aka rfMRI_REST?_??_Atlas_hp2000_clean.dtseries.nii). > >According to the Smith 2013 NeuroImage paper, it sounds like the images I >am using have all received global intensity normalization prior to the >temporal preprocessing. > >I have two sets of questions relating to this: > >(1) Does this mean that the global mean of all voxels (averaged across >time points) is used to normalize each voxel? If this is the case, is >this the mean of every voxel in the image, or exclusively those in the >brain? > >(2) Looking through ConnectomeDB, I couldn¹t find a file with the scaling >factor used for normalizing. Is this something that is saved anywhere? > >Thanks for the help! > >Best, >Erik Lee >_______________________________________________ >HCP-Users mailing list >[email protected]<mailto:[email protected]> >http://lists.humanconnectome.org/mailman/listinfo/hcp-users _______________________________________________ HCP-Users mailing list [email protected]<mailto:[email protected]> http://lists.humanconnectome.org/mailman/listinfo/hcp-users ________________________________ The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone or return mail. ________________________________ The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone or return mail. _______________________________________________ HCP-Users mailing list [email protected] http://lists.humanconnectome.org/mailman/listinfo/hcp-users
