What is Stars2Cells?

Stars2Cells is a neuron tracking pipeline for longitudinal calcium imaging experiments. It identifies and matches individual neurons across sessions using geometric relationships rather than relying on a single global alignment.

The method is inspired by astronomical star matching, where constellations are recognized despite rotation, translation, missing points, or partial overlap.

When should you use it?

Stars2Cells is designed for multi-session or longitudinal experiments where neuron identity must be preserved across days or conditions.

Common use cases include pre and post intervention studies, drifted fields of view, partial overlap between sessions, and recordings with cell turnover.

Quad-based geometric hashing

Local constellations of four neurons create invariant spatial signatures.

Voting-based correspondence

Each quad match votes for neuron pairs, producing robust consensus matches.

RANSAC refinement

Outlier matches are rejected while estimating the global rigid transform.

Quality control outputs

Residuals, inlier rates, and match confidence scores support inspection.

Pipeline overview

Neuron centroids
→ Quad generation
→ Similarity threshold estimation
→ Quad matching and voting
→ RANSAC refinement
→ Final neuron identity assignment
→ Match tables and QC outputs
  

Status and citation

Stars2Cells is currently an early release. Features, APIs, and documentation may change as the pipeline is validated across datasets.

If you use Stars2Cells, please cite the preprint:

Peden-Asarch AM, Honan LE, Bai JZ, Asarch EM, Quinn JA, Coffey KR, Neumaier JF. Stars2Cells: Astrometric Tracking of Neurons Across Imaging Sessions. bioRxiv, 2026. https://doi.org/10.64898/2026.07.03.736144