Home

We discuss software infrastructure of the Imagen project.

Recruitment and acquisition centres

The Imagen project is a longitudinal study with 4 time points so far:

ID	Time point	Age
BL	baseline	14
FU1	follow up 1	16
FU2	follow up 2	19
FU3	follow up 3	22

The databank team collect, curate, and publish data from the following recruitment and acquisition centres:

ID	CENTRE
01	LONDON
02	NOTTINGHAM
03	DUBLIN
04	BERLIN
05	HAMBURG
06	MANNHEIM
07	PARIS
08	DRESDEN

From acquisition centres to databank

Acquisition centres send pseudonymized data to the database team:

• Acquisition centres collect clinical and environmental data using Psytools, Dawba and Cantab:
  • Psytools data are exported daily and automatically from the Delosis server into CSV files pseudonymized with PSC1.
  • The database team manually download Dawba data from the Dawba server into CSV files pseudonymized with specific Dawba codes, different for each time point.
  • Cantab data are sent alongside neuroimaging data.
• Acquisition centres pseudonymize biological samples at the source using PSC1 codes. The biobank team collect samples and collate genetic data before sending them to the databank.
• Acquisition centres anonymize DICOM files before sending them to the databank. Again subjects are identified by their PSC1 code only.

The database team, acting as a trusted third party, pseudonymize data a second time, by converting dates to age and PSC1 identifiers to PSC2. We provide a list of valid identifiers to help end-users detect and investigate possible identifier errors.

Databank operations

Databank operations are described in specific pages:

• Questionnaires: download, anonymize and preprocess questionnaires.