Gopal Nataraj, a doctoral student in Electrical and Computer Engineering, received a U-M Rackham Predoctoral Fellowship to support his research that promises to lead to improved techniqes in magnetic resonance imaging (MRI).
The Rackham Predoctoral Fellowship is awarded to outstanding doctoral candidates in the final stages of their program whose research is unusually creative, ambitious and impactful.
Magnetic resonance imaging (MRI) is a safe, non-invasive, and flexible tool that is used to help diagnose neurological disorders and autoimmune diseases such as multiple sclerosis and Alzheimer’s disease. It can also detect biomarkers that indicate elevated risk for a variety of additional disorders, including strokes.
Gopal is working to make MRIs even better by improving their sensitivity to specific disorders as well as improving doctors’ ability to distinguish closely-related disorders. He intends to do this while acquiring the data as quickly as other fast, yet less reliable, methods of MRI.
His specific field of research is quantitative MRI (qMRI). Nataraj stated:
QMRI has potential to be more informative than conventional MRI because qMRI produces images that are spatially localized measurements of so-called ‘biomarkers’, or physical parameters of direct clinical significance. One example biomarker that we study is myelin water fraction (MWF). MWF is an indirect measurement of myelin content. Myelin is characteristic of healthy white matter in the brain, and its degradation is indicative of several disorders, notably multiple sclerosis (MS). Therefore, measuring MWF using MRI would provide doctors a route to specific, non-invasive information about the onset and progression of diseases such as MS.
In addition to MS, quantitative MRI can be used to improve our understanding of the progression of diseases like Parkinson’s, Alzheimer’s, and cancer.
However, qMRI needs to be improved before it is feasible for wide adoption. Nataraj has created an approach using big data techniques that may completely transform the field of quantitative MRI. The success of his technique is illustrated below.
The three images shown above reflect two existing methods of medical imaging, and one developed by Nataraj, to diagnose multiple sclerosis (MS). The images are attempting to track the myelin water fraction (MWF), a key in the diagnosis and tracking of MS.
The image on the left used a gold-standard method, called GRASE, that is unfortunately very slow. The middle image used a method called mcDESPOT that is fast, yet unreliable (see the review study of GRASE and mcDESPOT for a reference to these two techniques and images). The image on the right used Nataraj’s method which processed the data as quickly as the mcDESPOT technique, yet yielded results as reliable as the GRASE technique. His method makes use of a novel machine-learning algorithm to acquire precise biomarkers.
MWF is expected to be about 15% in healthy white matter (meaning, 15% of the content in each pixel that images white matter is due to myelin) and near 0% in other regions. The GRASE and proposed MWF images reflect this, whereas mcDESPOT clearly overestimates.
Through his improvements to qMRI, Nataraj hopes that this relatively safe diagnostic tool will be used even more broadly by the medical community.
Gopal Nataraj received his bachelor’s degree from Cornell University, and his master’s degree in Electrical Engineering:Systems from the University of Michigan in 2014. He is a member of Prof. Jeffrey Fessler’s research group, and is co-advised by Prof. Fessler and Dr. Jon-Fredrik Nielsen, a research scientist in the Department of Biomedical Engineering.
In addition to his research, Gopal served as president of the ECE Graduate Student Council (GSC) for over two years. The GSC strives to improve the climate for all graduate students in ECE.
Nataraj previously received a fellowship from Innovative Signal Analysis, Inc., and was recently awarded the Towner Prize for Distinguished Academic Achievement from the College of Engineering.