Removing outliers from the normative database improves regional atrophy detection in single-subject voxel-based morphometry

Vivian Schultz,Dennis M. Hedderich,Benita Schmitz-Koep,David Schinz,Claus Zimmer,Igor Yakushev,Ivayla Apostolova,Cansu Özden,Roland Opfer,Ralph Buchert,Schultz, Vivian
DOI: https://doi.org/10.1007/s00234-024-03304-3
2024-02-22
Neuroradiology
Abstract:Single-subject voxel-based morphometry (VBM) compares an individual T1-weighted MRI to a sample of normal MRI in a normative database (NDB) to detect regional atrophy. Outliers in the NDB might result in reduced sensitivity of VBM. The primary aim of the current study was to propose a method for outlier removal ("NDB cleaning") and to test its impact on the performance of VBM for detection of Alzheimer's disease (AD) and frontotemporal lobar degeneration (FTLD).
radiology, nuclear medicine & medical imaging,clinical neurology,neuroimaging
What problem does this paper attempt to address?