Abstract #2979
A machine learning approach to identify structural connections affected in diffuse axonal injury
J. Mitra 1 , S. Ghose 1 , K-K. Shen 1 , K. Pannek 2 , P. Bourgeat 1 , J. Fripp 1 , O. Salvado 1 , J. L. Mathias 3 , D. J. Taylor 4 , and S. Rose 1
1
Australian e-Health & Research Centre, CSIRO
Digital Productivity Flagship, Herston, QLD, Australia,
2
Imperial
College London, London, United Kingdom,
3
School
of Psychology, University of Adelaide, Adelaide, SA,
Australia,
4
Dept.
of Radiology, The Royal Adelaide Hospital, Adelaide, SA,
Australia
Patients with mild TBI sustain diffuse axonal injury
(DAI) which is microscopic in nature and difficult to
detect using conventional MRI. Diffusion MRI, along with
probabilistic tractography, is ideally suited to detect
DAI within specific white matter (WM) pathways. These
approaches, based on measures of structural
connectivity, can be used to identify damaged neural
pathways in group-wise analyses of TBI and healthy
control cohorts. We present a new method to identify
significantly different and discriminative structural
connections between the TBI and healthy control groups
by integrating a statistical GLM-based (generalized
linear model) network clustering and random forest
classifier.
This abstract and the presentation materials are available to members only;
a login is required.
Join Here
