Statistics Overview

Statistics is where Neuro QP turns slice-level results into comparisons that are easier to interpret across animals, groups, conditions, and stainings.

Use Statistics after the underlying Results look trustworthy.

The goal here is not to inspect individual cells, but to compare summary measurements across the study in a structured way.

What Statistics is for

Statistics helps answer study-level questions such as:

• Which groups differ in a selected region?
• Does one staining show a stronger regional pattern than another?
• Are differences driven by counts, proportions, or density?
• How do patterns look across the anatomy as a whole?

Before calculating statistics

Statistics is only as meaningful as the results underneath it.

Before relying on the plots, make sure that:

• the staining and model selections are correct
• the slice-level output looks reasonable in Results
• the brain regions included in the study are appropriate for the question you want to answer

Select a model for each staining

Statistics works per staining, and each staining uses the selected model for that staining.

This matters because the biological interpretation depends directly on which model output is being summarized.

If you switch to a different model, treat the resulting statistics as a different analysis state that should be reviewed again.

Why statistics can become outdated

If the underlying analysis changes, previously calculated summaries may no longer reflect the current state of the project.

In practice, that means you should recalculate after meaningful changes to the selected analysis inputs.

Do not mix an old statistical summary with a newer interpretation of the underlying results.

Start with filters, then plots

A good workflow is:

1. Apply filters so the data matches the exact comparison you want to make.
2. Choose the metric that matches the biological question.
3. Use the plot type that best supports that comparison.

This usually leads to cleaner interpretation than starting with a visually attractive plot before deciding what comparison actually matters.

Three plot families

Statistics currently supports three main visual summaries:

• Heatmaps for broad pattern comparison across animals and regions
• Mirrored plots for direct side-by-side comparison of grouped summaries
• Anatomical representations for mapping the selected metric back onto anatomy

Each plot answers a slightly different question.

The best choice depends on whether you care most about ranking, direct comparison, or anatomical localization.

Co-expression is a separate analysis mode

Statistics also includes co-expression analysis for pairs of stainings.

That view is useful when the scientific question is about overlap or separation between two markers rather than one staining at a time.

The separate Co-expression Analysis page explains how to think about those paired-marker summaries.