Reliability of brain volume measurements: A test-retest dataset

Citations Over TimeTop 10% of 2014 papers

Abstract

Evaluation of neurodegenerative disease progression may be assisted by quantification of the volume of structures in the human brain using magnetic resonance imaging (MRI). Automated segmentation software has improved the feasibility of this approach, but often the reliability of measurements is uncertain. We have established a unique dataset to assess the repeatability of brain segmentation and analysis methods. We acquired 120 T1-weighted volumes from 3 subjects (40 volumes/subject) in 20 sessions spanning 31 days, using the protocol recommended by the Alzheimer's Disease Neuroimaging Initiative (ADNI). Each subject was scanned twice within each session, with repositioning between the two scans, allowing determination of test-retest reliability both within a single session (intra-session) and from day to day (inter-session). To demonstrate the application of the dataset, all 3D volumes were processed using FreeSurfer v5.1. The coefficient of variation of volumetric measurements was between 1.6% (caudate) and 6.1% (thalamus). Inter-session variability exceeded intra-session variability for lateral ventricle volume (P<0.0001), indicating that ventricle volume in the subjects varied between days.

Related Papers

• → Distribution System Reliability Assessment for Improved Feeder Configurations(2019)16 cited
• → Power Systems Reliability Assessment based on Load Shaping Consideration(2023)3 cited
• → Separating HBT wearout from defects during early life operation(2013)1 cited
• → TIME REDUCTION TESTS OF RELIABILITY ELECTRICAL EQUIPMENT(2020)
• → Reliability and maintainability design of the micro ammeter(2024)