Improving the Analysis of Object (or Cell) Counts with Lots of Zeros

An approach using hurdle and zero-inflated models with brms

NeuN and GFAP staining in a mouse brain following cerebral ischemia. Manrique-Castano et al. (2024) (CC-BY).

Counting is a fundamental task in biomedical research, particularly when analyzing cell populations. Imagine staring at countless cells within a tiny brain region — sometimes numbering in the hundreds or thousands. Yet, in other areas, these numbers may dwindle to few or even none.

The challenge arises in how we analyze these counts. For …

data analysis data science data visualization statistical modeling thoughts-and-theory