Quality Control
QC Overview
Quality control is applied in five stages as data moves from collection at sites, through the Data Coordinating Center, to curated releases. Select a stage to see when it runs and what it checks.
The QC pipeline#
At the site, before data leaves the collection instrument.
Example checks
- Survey prompts/design to prevent invalid response submission
- EEG session log forms to confirm protocol compliance
As data enters the Centralized Data Repository.
Example checks
- Naming, file size, and metadata checks for ingested file-based data
- De-identification checks
After conversion into standard formats.
Example checks
- BIDS validation to confirm successful file/formatting conversion
- EEG capping quality review
Ongoing review of scored and summarized data.
Example checks
- Monthly SME data review via QC Dashboards
- Review of summary statistics and distributions to check for outliers
Final review before data ships in a release.
Example checks
- Final QC performed by SMEs of data in release format prior to approval for public release
Neuroimaging#
1
Source QC
- Session Log Form - session metadata
- Upload Data - uploaded in last step of form
2
Ingestion QC
- Ingestion Checks - metadata checks for ingestion
- Notifications - sent to Core
3
Pre-Processed Data QC
- Raw Values - BIDS conversion and QC
4
Processed Data QC
- Pipeline Processing - derivatives QC
5
Pre-Release QC
- SME Review - SMEs review and sign off on release-formatted data
Data Release
REDCap#
1
Source QC
- EDC Validation - range checks, required fields, branching logic
2
Ingestion QC
- Transfer QC - checksums
- Scoring & CRF - scoring and case report form checks
3
Pre-Processed Data QC
- Pre-processing - BIDS conversion & shadow matrices
4
Processed Data QC
- Not applicable for this modality
5
Pre-Release QC
- SME Review - SMEs review & sign off on data in release format
Data Release