Intern uses neural networks to advance medical imaging
Emily Corcoran always wanted to prove she was more than just a rural girl living in a small Wisconsin town. Even though she excelled at math and her family and teachers encouraged her studies, school never felt like it was pushing her to be the best that she could be. But Corcoran was determined, and after high school she graduated from Northland College with her bachelor’s degree in mathematical sciences with a minor in physics.
Now, Corcoran is in her second year at Marquette University to receive her doctoral degree in computational mathematical and statistical sciences. She feels excited knowing that she can show how women excel in the science, technology, engineering, and mathematics (STEM) field.
“Women can love and be great at math and computer science, and women can get doctorates. I get such a wonderful feeling knowing that when my little cousins picture a mathematician, they will picture a woman: me,” she said.
Long before she applied to be an Oak Ridge Institute for Science and Education (ORISE) intern, Corcoran was participating in other summer programs. She developed her skills through the Summer Institute for Biostatistics at the University of Minnesota, and then the next summer with the Mathematical Biosciences Research Experiences for Undergraduates at Ohio State University. After two summers of study, she discovered the Energy Efficiency & Renewable Energy High Performance Computing for Manufacturing (HPC4Mfg) Internship Program and applied.
The HPC4Mfg Internship Program provides the opportunity for participants to perform advanced simulation and modeling to strengthen computational techniques to research reducing energy and material use during manufacturing.
For ten weeks Corcoran worked remotely with her mentor Youngsoo Choi and their collaborator Kyung Hyun Sung applying her computational science skills. She began learning about neural networks and how they could be used to advance the medical field. Magnetic Resonance Fingerprinting (MRF), a new form of Magnetic Resonance Imaging (MRI), was her main topic of research.
MRI machines use magnets and radio waves to generate images of the human body, and MRFs are currently being studied to fine-tune and advance the time MRI scans take, as well as their overall quality. Corcoran sees MRF scans like FBI fingerprinting.
“The FBI can take a fingerprint and try to match it to their database. If they find a match, then they all of a sudden have a whole bunch of information about a person like their name, phone number, address, etc.,” she explained. “In MR Fingerprinting, we start with a signal from an MRI machine, and like a fingerprint, the signal doesn’t tell us anything on its own. But if we are able to find a match to that signal in a dictionary, then we have a whole bunch of information about different parameters of our choosing, and that information can be used to build 3D MR images.”
With the usage of neural networks, this ever expanding MRF technology can be made more efficient and timely. The future usage of MRF is certain to help treat the sick in a faster, much more tailored way.
During her time as an intern, Corcoran was glad to add new knowledge to her tool belt. Having not worked with neural networks before, she found the ability to extend her basic knowledge outward to be an enriching experience. She also credited her mentor, Youngsoo Choi, for helping her develop invaluable researching skills.
Corcoran hopes of one day becoming a mathematics professor, though she will be keeping a career with a national lab in the back of her mind. For now, she is beginning to study another type of medical imaging called electrical impedance tomography, and will be using what she learned about neural networks in her research.
“I would certainly recommend the ORISE program,” Corcoran finished. “I think any STEM-focused internship appointment at a large federal research facility sounds intimidating to a lot of students, or at least it did to me, so my advice would be to not let imposter syndrome hold you back. If my experience is any indication, everyone is there to help you learn and they want to see you do well.”
The EERE High Performance Computing for Manufacturing (HPC4Mfg) Internship Program is funded by the U.S. Department of Energy (DOE) and administered through the Oak Ridge Institute for Science and Education (ORISE). ORISE is managed for DOE by Oak Ridge Associated Universities (ORAU).