PRIME-Preprocessed

ongoing project: preprocess PRIME-DE data

Goal:

• 1. Quality assessment
• 2. Minimal preprocess
• 3. End-to-end preprocess

Datasets:

• 5 core fMRI sites (n≥10 per site, function data)
  • Newcastle University (N=14)
  • UC-Davis (N=19)
  • Oxford (N=20)
  • University of Wisconsin-Madison (N=592)
  • University of Western Ontario (N=12, 7 MP2RAGE + 5 MPRAGE)
• 5 core T1+T2 sites (n≥5 per site, T1w+T2w)
  • Mount Sinai-Philips (n=8 total, 5 T1+T2, 6 func+T1)
  • Mount Sinai-Siemens (n=5)
  • Newcastle (n=14)
  • UC-Davis (N=19)
  • University of Western Ontario (N=12, MP2RAGE)

Example data: 2 subjects for each of 5 core sites above: aws s3://fcp-indi/data/Projects/INDI/PRIME/ExampleData.tar.gz

Brain masks for example data: link

Pipeline configuration example: https://github.com/HumanBrainED/PRIME-Preprocessed/blob/main/pipeline_config.yml

---

Anatomical brain mask (pre-release, all T1w masks)

https://github.com/HumanBrainED/PRIME-Preprocessed/releases

v0.1_anat_brainmask: brain masks for T1w images from PRIME-DE

Masks are generated using the U-Net model (https://github.com/HumanBrainED/NHP-BrainExtraction)

• Masks for all sites except site-uwmadison are created based on brainmask_T1w_136macaques.tar (Ref: Reference: Wang et al.,2021)
• U-Net model for site-uwmadison can be downloaded here
• Join Quality Control

Note: The masks are not perfect here. Use them as initial masks

---