Shared AI2D Neuroimaging Data Resources

This resource is a pilot meant to jump-start a broader effort for datasharing at AI2D.

Many labs at Penn use the same large, publicly available neuroimaging datasets. Historically, each lab downloaded, processed, and maintained their own version of the data – resulting in duplicated efforts, wasted computational resources, and inefficient use of storage. The shared neuroimaging data resources provided by AI2D addresses these challenges by providing centralized access to many of the largest and most commonly used neuroimaging datasets. All datasets are curated, processed, and shared using a standardized workflow that includes widely-used tools developed by AI2D labs. The result? Larger scale, lower cost, and faster progress.

Currently Available Shared Neuroimaging Data Resources

Note: The Sample Size and Number of Sessions shown in the table below represent the full dataset. The number of completed sessions for each processing pipeline can be found on each study-specific page.

Dataset Age Range (y) Sample Size Number of Sessions Raw Processed
T1w fMRI dMRI fmap FreeSurfer fMRIPrep XCP‑D QSIPrep QSIRecon
28andHe 26 1 40 NA NA NA NA
28andMe 23-24 1 60 NA NA NA NA
Developmental Chinese Color Nest Project 6.5-17.9 195 195 NA NA NA NA
Healthy Brain Network 5-22 3887 3887
Human Connectome Project - Young Adult 22-37 1206 1206 NA NA NA NA NA NA NA
Midnight Scan Club 24-34 10 123 NA NA NA
NKI Rockland Sample 6.2-85.6 1329 2306 NA
Neural Modelling variable variable variable NA NA NA NA
Philadelphia Neurodevelopmental Cohort 8-23 1601 1601
Single-echo/multi-echo comparison pilot 21-30 8 17 NA NA NA NA

NA = Not Available
✅ = Available