My well in the exams, but I know

My first contact with MRI was a voluntary experiment in a neuroscience lab. I was put
into the bore of a scanner and showed a succession of pictures. When I got out, I saw, for the
first time, what my brain looked like. The researcher told me that the highlighted area was the
active part of my cortex. I was astonished because I had no idea how MRI could detect the
activity of my brain.
When I started taking academic courses, I knew more about the principles of different
imaging modalities. With plenty of toolboxes and tutorials available online, I found it more
meaningful to understand the intrinsic principles rather than memorize every point on the slides.
For example, when I studied signals and systems, I did not try to memorize every single theorem
of the Fourier transform. What I wondered was why we need to do the Fourier transform, why is
the Fourier transform so special, and what is the Fourier transform essentially. Thinking about
these questions would not get me a higher score compared to doing more exercises, but not
thinking about them made me uncomfortable, because I felt like I had deviated from the truth.
This habit did not serve me well in the exams, but I know in the long run I will benefit from
exploring the essence, because great innovations start from challenging the basics.
Of all the imaging modalities, MRI is very promising with so many possibilities. To learn
more about MRI, I started to read the textbook by Professor Liang and Lauterbur, trying to
derive every equation in the textbook. The measurement of diffusion seemed natural in the
textbook, but in fact it took nearly 20 years to come true. We often take things for granted and
seldom ask why. However, if we do not look at technological development from a historical
perspective, there is very little chance we can understand why it has become the way it is.
Neither can we foresee its future. And I started to wonder, after reading so much about other
people’s accomplishments, what difference can I make to MRI?
I had my first opportunity to explore this question in my junior year when I joined the lab
of Professor Kui Ying. She wanted me to take over the project of the MRI tabletop built for
educational purposes, which many students had quitted because it was too difficult for
undergraduates and it would not produce a publication. But it certainly was rewarding. I spent
much time reading the datasheets and testing the system chip by chip. Finally, we could
successfully generate a designed waveform for the coils and run the whole system with a manual
setup. Further work with the interface is still in progress. From this project, I strengthened my
skill of debugging an electronic system and gained a deeper understanding of MRI physics. Besides hardware, I also attempted to investigate the “software” of MRI. I proposed a
correction term of susceptibility to the hybrid model of MR Thermometry and validated it on a
specially designed phantom. The algorithm of the hybrid model by Professor Grissom was
beyond my mathematical knowledge then, thus I learnt non-linear optimization through online
courses and wrote the code all over again. In the summer of 2017, I worked as an intern in MGH.
In order to correct the eddy-current effect on a multi-echo UTE sequence, I developed a
sequence to measure its k-space trajectory under Siemens IDEA environment. In this project, I
experienced the full process of pulse development, from pulse programming, phantom and invivo
experiments to reconstruction and data analysis. After I returned to my school, I started to
learn about deep learning. Since our lab had worked on MR Fingerprinting (MRF) before, I was
wondering if deep learning could be applied to MRF. I trained a CNN model to predict T1 and T2
in place of the conventional dictionary matching. The accuracy of CNN was surprisingly high
even with a shorter acquisition time. Through these projects, I had chances to study different
aspects of MRI and improve my programming skills and mathematical skills. More importantly,
after writing my own abstracts, I am able to look at others’ research with more insight.
From the perspective of social development, a great many of our needs have been
satisfied by electronic devices and appliances. Now people begin to study VR and AR to add
more excitement to people’s lives. But is any need more fundamental than the need for health?
Medical imaging is critical for medicine as a powerful tool. Therefore, I decided I am going to be
a researcher in medical imaging. I am going to be the one to provide more advanced tools for
doctors and make a difference to people’s health.
Yale University is a prestigious school where considerable influential work in MRI has
come from. It would be an honor for me if I can work with Professors Todd Constable, Robin A.
de Graaf or Dana C Peters. Your school has a strong cooperation with hospitals and industry,
which will facilitate the translation of experimental techniques to clinical practice. Besides, it is
great to work with people from different backgrounds, such as radiologists, physicians,
physicists, chemists and engineers, allowing an exchange of brilliant ideas. Earning a PhD will
be the very first step of my research, and I am prepared to start my journey.