Meeting Banner
Abstract #0236

Deep convolution neural network-based DWI tractography connectome analysis to predict language improvement after pediatric epilepsy surgery

Jeong-Won Jeong1,2,3,4, Min-Hee Lee1,2, Nolan O'Hara2,4, Eishi Asano1,3,4, and Csaba Juhasz1,2,3,4
1Pediatrics, Wayne State University, Detroit, MI, United States, 2Translational Imaging Lab, Children's Hospital of Michigan, Detroit, MI, United States, 3Neurology, Wayne State University, Detroit, MI, United States, 4Translational Neuroscience Program, Wayne State University, Detroit, MI, United States

Early surgery helps improve language function in pediatric epilepsy. We investigate if an advanced DWI approach combining deep convolution network-based tract classification with DWI connectome can help early surgery by providing preoperative imaging markers which indicate a high likelihood of postoperative language improvement. Our approach revealed two nodes in preoperative DWI data, including left middle temporal gyrus and left angular gyrus, of which preoperative local efficiency values are not significantly different in patients having postoperative improvement of receptive language, compared with age-matched healthy controls, which can be as effective imaging markers for prediction of the postoperative language improvement.

This abstract and the presentation materials are available to 2020 meeting attendees and eLibrary customers only; a login is required.

Join Here