Abhirup Datta

SPLASH improves task-based fMRI analysis by providing a spatially adaptive and scalable framework that enhances statistical inference and neuroscientific interpretability.

The paper introduces nnSVG, a scalable method for identifying spatially variable genes using nearest-neighbor Gaussian processes, and demonstrates its performance on experimental data and simulations.

This paper introduces PredLMM, a method for rapidly estimating SNP heritability in large cohorts by approximating linear mixed models using predictive process techniques.