I hope you all are fine and in good health. So , I want preprocessing of the ABIDE-02 fMRI dataset using fMRIPrep tool. I downloaded the ABIDE-02 dataset using [ABIDE-02] website but now the issue is that the fMRIPrep needs dataset in BIDS format and for the BIDS to work , we need to have DICOM formatted dataset , whereas, the dataset that I downloaded from the mentioned link is in NIFTI format .
The other way is to download the RAW dataset from amazon s3 bucket , As I guess RAW dataset means DICOM dataset but somehow that is also not working as I tried to use Cyberduck and configured parameters correctly but connection always fails to the S3 bucket.
Can somebody please guide me how can I accomplish my task of preprocessing ABIDE-02 dataset using fMRIPrep tool based on my scenario ? Moreover, I am in Windows environment and will use fMRI docket for all the preprocessing.
As I am new to this preprocessing stuff so lot of my questions will sound I guess pretty stupids.
Neuroimaging files in BIDS are niftis (.nii.gz). BIDS more refers to how they are named and organized.
Is this a computing cluster (usually linux-based) or a personal computer? There are 1000+ subjects in ABIDE II, so I doubt any personal computer will have enough storage to hold all of the data or enough RAM/CPUs to process the data efficiently.
Thanks for you reply . According to the BIDS specifications, BIDS not only needs .nii files but it also needs a bunch of other csv and tsv files which are I guess metadata files. For example in one of the folder that contains the subject scanning data , There are two more folders named “anat” and “func” and in each of those folders is .nii zipped file. So how to get the metadata csv files inside those folders?
Regarding your second suggestion, I have a personal computer but it has good resources but even if it did not work due to memroy and computational resources then I have option of moving the dataset and scripts to the worthy cluster.
Although BIDS has a lot of possible csv/tsv files, not that many are required. Usually it’s just the participants.tsv file in the BIDS root directory, and events.tsv for any non-resting state fMRI. You may have to make these yourself based on the names of the participants (for participants.tsv) and any task timing information (for the events.tsv).
Thanks for your reply, So ABIDE 2 dataset is resting state fMRI dataset. The subject level metadata file seems easy to make but the for resting state and anotomical related tsv or json files , do you have any idea how to make such json files?
In that case, you do not need an events.tsv for the functional files, since there are no events.
What files are currently available? Can you provide the directory structure of a single subject (e.g. with the tree command)? If there are not any available, it isn’t too hard to make minimally sufficient BIDS json files. I believe for fMRIPrep you just need the following fields (though I could be wrong):
RepititionTime
TaskName
PhaseEncodingDirection
SliceTiming (if you want to do slice timing correction)
If you have the raw dicoms, you can also use a tool like dcm2bids, which will convert from DCM to NII+json and take care of the file naming and organization.