Description
Resting-state functional magnetic resonance images are invaluable for evaluating the neurocognitive state of patients, however, rs-fMRIs are highly susceptible to motion. The combination of machine learning and image reconstruction techniques during and after image acquisition holds great promise for harmonizing images and recovering motion-corrupted data. We demonstrate the use of unsupervised machine learning techniques to investigate harmonization characteristics and to identify important features of the motion artifacts in a multi-center neuroimaging study.
Learning Objective: After participating in this session, the learner should be better able to:
- Understand the impact of motion on medical images
- Learn about the current challenges in combining resting-state functional magnetic resonance images (rs-fMRIs) from multiple sites into a single cohesive data set.
- Learn the different measures of quality assessment used to evaluate motion in rs-fMRIs
- Understand the potential impact of applying machine learning to rs-fMRIs for the purpose of classifying subgroups within a population
Authors:
Jenna Schabdach (Presenter)
University of Pittsburgh
Vincent Schmithorst, Children's Hospital of Pittsburgh of UPMC
Vince Lee, Children's Hospital of Pittsburgh of UPMC
Rafael Ceschin, University of Pittsburgh
Ashok Panigrahy, Children's Hospital of Pittsburgh of UPMC
Presentation Materials: