Abstract #1262

Multilevel classification of Alzheimer’s and Mild Cognitive Impairment patients by using Diffusion Tensor Imaging data

Ranganatha Sitaram¹, Josué Luiz Dalboni da Rocha^2,3, Ivanei Bramati⁴, Gabriel Coutinho⁴, and Fernanda Tovar Moll⁴

¹Institute for Medical and Biological Engineering and Department of Psychiatry, Pontificia Universidad Católica de Chile, Santiago, Chile, ²Biomedical Engineering, University of Florida, Gainesville, FL, United States, ³University of Florida, Gainesville, FL, United States, ⁴Instituto D'Or de Pesquisa e Ensino, Rio de Janeiro, Brazil, Rio de Janeiro, Brazil

The proposed novel approach is based on three levels of analyses of DTI data: 1) voxel level analysis of Fractional Anisotropy, 2) connection level based on fiber tracks between brain regions, and 3) network level based connections among multiple brain regions. This novel approach was applied to differentiate between AD, MCI and controls. We achieved accuracy of 93% between AD and controls, 90% between AD and MCI. Main discriminative areas were Hippocampal Cingulum and Parahippocampal Gyrus. The results suggest that our multilevel DTI analysis not only informs difference between brain conditions, but also shows strong potential as diagnostic tool.

This abstract and the presentation materials are available to members only; a login is required.

Join Here