Developing univariate neurodegeneration biomarkers with low-rank and sparse subspace decomposition

•A subspace decomposition method to generate univariate neurodegeneration biomarkers.•AD pathology-confirmed univariate morphometry index on hippocampal surface features.•Matrix factorization facilitates subspace decomposition of large-scale matrices.•A novel sparse shrinkage strategy to ensure the original 3D spatial connectivity.•UMI outperforms hippocampal volumes in a series of experiments on the ADNI cohort.