Skip to content

DeepPrep

Xuan Mai PHAM edited this page Mar 12, 2025 · 7 revisions

DeepPrep is an accelerated, scalable, and robust pipeline for neuroimaging preprocessing empowered by deep learning.

You can now run deepPrep in CBRAIN by following these steps:

Select your input files

Please note the following prerequisites:

  1. The BIDS format

DeepPrep is able to end-to-end preprocess anatomical and functional MRI data for different data size ranging from a single participant to a HUGE dataset. It is also flexible to run the anatomical part or functional part that requires a complete Recon folder to be specified. The DeepPrep workflow takes the directory of the dataset that is to be processed as the input, which is required to be in the valid BIDS format.

  1. BIDS format validation

Any new dataset has to be BIDS-compliant (see specifications at http://bids-specification.readthedocs.io/) and should be validated with tools provided by BIDS, such as the BIDS-validator available in CBRAIN.

  1. The FreeSurfer License

DeepPrep uses FreeSurfer tools and thus requires a valid license. To obtain a FreeSurfer license, simply register for free at https://surfer.nmr.mgh.harvard.edu/registration.html.

Full information can be found HERE

Once your file(s) is uploaded, please select from the Files list:

image

Process the Launch Task steps and select deepPrep tool:

image

4. Task parameters

Set up the task parameters according to your analysis

image

5. Output (Derivatives)

The outputs of DeepPrep have three categories:

  • Anatomical derivatives: The anatomical images are preprocessed through motion correction, tissue segmentation, cortical surface reconstruction, volume and cortical surface registration, and etc.

  • Functional derivatives: The functional images are preprocessed through motion correction, slice timing correction, susceptibility distortion correction, confounds estimation, and registration to both standard and non-standard spaces.

  • Visual reports: For better visualization and understanding of the entire process, DeepPrep generates a visual report for each subject.

You can navigate in the tree directory by expanding each folder and visualize files without the need to download them by simply clicking on the link.

image )

image

image

Clone this wiki locally