Using the Results Viewer
The Results Viewer is the main place to visually inspect how the selected staining and model behave on individual slices.
Use it to confirm that the summarized values in the Results Table are supported by what you actually see in the images.
Why use the viewer
A table can tell you that one region has a high percentage of ON cells.
The viewer tells you whether that result looks biologically plausible.
This is important because trustworthy interpretation starts with trustworthy visual output.
Select regions and slices
The sidebar lets you work in two directions at once:
- choose the brain regions you want to focus on
- move through animals and slices to inspect different parts of the study
In practice, it is useful to inspect:
- a few strong examples
- a few borderline or difficult examples
- slices from different animals
- slices that look surprising in the Results Table
What the overlay tells you
The viewer combines the image data with the selected anatomical regions and the current classification output.
This helps you judge questions like:
- Are the cells assigned to the expected regions?
- Does the pattern look anatomically plausible?
- Are obvious false positives or false negatives concentrated in one part of the image?
When you hover a region, Neuro QP also shows a short region summary for that slice, including counts and density-related values.
That is useful for local inspection, especially when one slice appears to drive a broader result.
Use the viewer for quality control
The most valuable use of the viewer is quality control.
Look for patterns such as:
- classification output that looks systematically too broad or too sparse
- region boundaries that do not match the local anatomy well enough
- slices that appear different from the rest of the cohort for technical rather than biological reasons
You do not need to inspect every slice exhaustively.
In practice, reviewing a small set of representative and suspicious slices is usually much more useful than trying to review everything equally.
What to do if something looks wrong
If the output looks unreliable, stop short of interpreting the corresponding summaries.
Instead, use what you see to decide what needs attention next.
Depending on the issue, that might mean:
- reviewing model choice
- checking whether the classification output is acceptable for this staining
- reviewing the relevant anatomical placement more carefully
- excluding interpretation until the questionable output is understood
A practical rule
Use the Results Viewer whenever you want to answer the question:
Do I trust what the summary is telling me?
If the answer is not clearly yes yet, the viewer is the right place to continue reviewing before moving on to broader comparisons.