Premium
[IC‐P‐144]: AUTOMATED FEATURE EXTRACTION ALGORITHM FOR T1‐MRI DATA SELECTS AD‐VULNERABLE REGIONS TO SEPARATE MCI INDIVIDUALS FROM COGNITIVELY NORMAL CONTROLS
Author(s) -
Rane Swati,
Levendovszky Tihamer
Publication year - 2017
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2017.06.2418
Subject(s) - random forest , entorhinal cortex , pattern recognition (psychology) , feature selection , voxel , artificial intelligence , estimator , computer science , psychology , mathematics , neuroscience , statistics , hippocampus
segmentation. While large (>3000 mm) structures with distinct gray matter-white matter contrast such as the thalamus are well-delineated and consistent amongst software, segmentation of smaller structures such as the amygdala shows considerable variability and bias depending on the software used. The plots for Caudate, Putamen, Hippocampus, Pallidum were similar to Thalamus. However, the putamen showed a proportional error between FreeSurfer and Multi-atlas measurements. All volumes normalized by ICV.