Are Deep Learning (DL) Techniques able to improve the quality of diffusion MRI data to be used in clinical studies? 

Are the subtle details and differences between groups kept or lost when using DL?

Which specific DL method is more suitable to be used in dMRI clinical studies?

Overview

In this challenge, we ask the participant to upgrade diffusion MRI data acquired with only 21 gradient directions to 61 gradient directions via DL. As a final result, we will ask them to provide only three scalar metrics: FA, MD and AD. In order to compare the different methods, we will use a real clinical study in which we statistically compare episodic migraine to chronic migraine. We have detected that differences between groups disappear when using 21 directions instead of 61. Will any DL method be able to recover those differences?

Important Dates

Release of training data (healthy controls):  
Release of challenge data (patients):  
Team registration deadline:       
Final submissions:     

April 1, 2022
June 1, 2022
 June 1, 2022
 July 15, 2022

Organizers

Prof. Santiago Aja-Fernandez

Carmen Martín-Martín, MSc

Tomasz Pieciak, PhD

Álvaro Planchuelo-Gómez, PhD

Rodrigo de Luis-García, PhD

Antonio Tristán-Vega, PhD

Laboratorio de Procesado de Imagen
E.T.S.I. Telecomunicación
Campus Miguel Delibes
Universidad de Valladolid
47011, Valladolid, Spain
WWW.LPI.TEL.UVA.ES