How Registration Works
Registration tells NeuroQP where your tissue sits anatomically so detected cells and marker-positive objects can be assigned to brain regions.
The exact registration behavior depends on acquisition mode.
4x reference + detail projects
These projects have two placement steps:
- Brain slice to atlas registration: Aligns the anatomical reference image, usually
4x, to the atlas. - Detail-image placement: Places the
10xor20xdetail image onto the anatomical reference.
Together, these steps let NeuroQP determine which brain regions are covered by the analyzed detail field and assign detections to the right anatomical regions.
High-res whole-slice projects
These projects use a 10x or 20x full-slice image directly as the anatomical reference.
There is no separate detail-image placement step. Atlas registration runs directly on the high-resolution reference image, usually using a downscaled copy for the registration algorithm.
Detail-only projects
Detail-only projects do not use atlas registration.
They are intended for legacy detail-image datasets without a whole-slice reference. Instead of mapping detections through the atlas, NeuroQP uses the brain region assigned to each uploaded whole image.
This means Detail only does not provide atlas-derived region boundaries or the same spatial review benefits as registered projects.
What happens automatically
After upload, NeuroQP creates an initial atlas registration automatically when the anatomical reference image is available.
For 4x reference + detail projects, NeuroQP also places detail images onto the reference automatically.
Detail-only projects skip both atlas registration and detail-to-reference placement.
What to review manually
The most important thing to refine in registered projects is the atlas registration in the anatomy that matters for your analysis.
For 4x reference + detail projects, that usually means the area covered by the detail image and selected brain regions.
For high-res whole-slice projects, that means the relevant full-slice anatomy and selected regions.
For Detail only projects, review the direct brain-region assignment instead.
What a good registration looks like
A good registration means atlas boundaries match visible anatomy well enough for trustworthy region assignment.
In practice, around 10 to 20 well-placed points is often enough for a strong local fit, but the exact number depends on the slice.