This is the first analysis plan.
It is provisional.
- Read about Nipy diagnostics
- Consider Principal Component Analysis (good tutorial?)
- Consider Independent Component Analysis (e.g. FSL)
- Your ideas here.
- Build simplest possible draft using e.g. standard deviation
- Review identified scans to see whether they are plausible outliers.
- Extend with new tools:
- Motion correction
- Statistical modeling