Most CNS projects are secondary analyses of existing open datasets. The work at this stage is figuring out which dataset can actually answer your question, and starting the access process early — applications often take weeks, and they run independently of everything else you are setting up.
Datasets used in this program¶
PREVENT-AD — longitudinal pre-symptomatic Alzheimer’s cohort, with imaging, MEG/EEG, PET and CSF. The most thoroughly documented dataset in this book, and the one several scholars build on
ADNI — the long-running Alzheimer’s neuroimaging initiative
UK Biobank — very large population cohort with imaging and genetics
ABCD — adolescent brain and cognitive development
Portals to search¶
Portals index many datasets rather than hosting one:
OpenNeuro — the largest open archive of BIDS-formatted neuroimaging data
CONP — the Canadian Open Neuroscience Platform
BrainCODE — Ontario Brain Institute data, including ONDRI
DANDI — open neurophysiology
brainlife.io — data plus browser-based pipelines
UK Biobank RAP — the analysis platform you work through for UKBB
Choosing¶
Work backwards from the question. Two practical filters:
Does it have the modality you need? Check fMRI, structural MRI, DWI, EEG/MEG, PET
Is it longitudinal? If your question is about progression, you need repeated scans, which rules out a lot of otherwise attractive datasets
Or start from someone else’s answer: the research topics hubs list which scholars work on what, and their pages name the datasets they chose.
Before you can download anything¶
Open data is rarely open unconditionally. Several of the datasets above need an application, a Data Use Agreement, or institutional ethics approval — see ethics and data access, and start that process now rather than after you have chosen.
Next: Get compute — start this in parallel, not after.