Open AccessCO-OCCURRENCE MATRIX IN LONGITUDINAL DIFFERENTIATION OF MRI IMAGES OF PATIENTS WITH AD, MCI, AND HEALTHY INDIVIDUALS
Author(s) -
Daniella de Oliveira,
Amanda S. Nascimento,
Rafael da Silveira,
Isadora Cristina Ribeiro,
Thamires Naela Cardoso Magalhães,
Fernando Cendes,
Márcio Luiz Figueredo Balthazar,
Gabriela Castellano
Publication year - 2021
Publication title -
dementia and neuropsychologia
Language(s) - English
Resource type - Conference proceedings
SCImago Journal Rank - 0.54
H-Index - 21
ISSN - 1980-5764
DOI - 10.5327/1980-5764.rpda005
Subject(s) - magnetic resonance imaging , precentral gyrus , voxel , longitudinal study , audiology , medicine , nuclear medicine , psychology , radiology , pathology
Background: Texture analysis based on the gray level co-occurrence matrix (GLCM) has been applied to brain magnetic resonance images (MRI) to detect subtle differences among healthy and lesioned tissue occurring in several neurological conditions, but it has rarely been used in longitudinal AD studies. Objectives: To perform a longitudinal study by applying GLCM to brain MRI of AD and MCI patients and of healthy controls (HC), to verify the suitability of this technique to help detect the evolution of these conditions. Methods: Participants were 14 AD, 21 MCI and 17 HC, who had 2 T1-MRI obtained ~12 months apart. MRI were segmented using the AAL atlas. 3D GLCMs were computed for five voxel distances, for 16 regions per subject. A total of 55 texture parameters were extracted per region per subject. Statistical differences were evaluated using a t test. Results: Significant differences were found only for the MCI group, for the left precentral gyrus and left supplementary motor area, for which the correlation parameter decreased over time at different distances. Conclusions: This result could be due to a subtle motor loss in the MCI group before the onset of AD symptoms, or even, patients in the MCI group could progress to neurodegenerative diseases other than AD. The next step is to compare the obtained texture parameters between groups using analysis of covariance.