Release Notes

NeuroQP now has a more consistent interface across Study, Registration, Training, and Results.

This update improves layouts, viewer controls, table controls, brain-region selection, and Results co-expression review so the main workflow feels more unified across the app.

NeuroQP now supports the Kim Unified Mouse Brain Atlas in addition to the Allen Mouse Brain Atlas! The Kim atlas includes more detailed mouse brain subregions.

You can change the used atlas in the Study tab where you select your brain regions.

NeuroQP now supports legacy Detail only projects without a whole-slice reference image.

You can upload 10x or 20x detail images, assign them directly to brain regions, and provide pixel size for area-based measurements.

When a whole-slice reference is available, we still recommend using a registered project mode.

Statistics now has more plot options and better controls for publication-ready figures.

• New plot types: bars, vertical bars, dot/strip, box/violin, effect size, and scatter plots
• Manual ranges for heatmaps, anatomical plots, bar-style plots, and mirrored plot axes
• Shared-range controls for comparing plots on the same scale
• Selectable first color for grouped plot palettes
• Improved heatmap, mirrored, and anatomical plot layouts, legends, and SVG exports

Fix: Training, cell detection, and atlas registration now update queued, processing, and error states more quickly.

Registration, Training, and Results viewers now include compact brightness controls for adjusting image display.

NeuroQP now supports more flexible staining workflows beyond the original DAPI-centered setup.

• Project creation now allows you to specify different project configurations
• Cell detection can use a normal cell nuclei marker such as DAPI or NeuN in cell-counting projects
• Marker stainings such as cFos or TRAP can run independent cell detection when there is no shared DAPI/NeuN nuclei staining for cell counting
• Full-brain high-resolution slice projects are supported at 10x and 20x without a separate 4x anatomical reference image
• Full-brain high-resolution projects skip detail-to-reference alignment because the analyzed image is already the anatomical reference
• Training, Results, Statistics, and co-expression now understand independent-detection projects and their marker-positive cell semantics

Existing 4x reference plus 10x or 20x detail projects continue to work as before.

NeuroQP now supports projects built around 10x detail images in addition to 20x.

With this release:

• You choose the detail magnification when creating a project
• Upload validation now supports 4x, 10x, and 20x image labels
• Alignment, cell detection, and downstream analysis follow the project's selected detail magnification
• Existing 20x projects continue to work as before

You can now export a project as a single reproducible archive. This is part of our ongoing effort to make your experiments reproducible and support you with writing your own code on top of NeuroQP.

This export can include:

• Original uploaded images
• Registration point pairs and AP values
• 4x detail alignment points
• Cell detection masks
• Training samples
• One selected classifier per staining

• Automated 20x DAPI cell detection
• Whole-slice brain atlas registration from 4x DAPI images
• 20x overlay placement and manual registration refinement
• Custom cell-classifier training for staining-specific analysis
• Results review in tables and image viewers
• Brain-region statistics and publication-ready plot exports
• Co-expression analysis in the statistics workflow
• Public docs covering project setup, uploads, registration, training, results, and statistics