Yu Veronica Sui, PhD, is a postdoctoral fellow at NYU Langone Health and a scientist at the Center for Advanced Imaging Innovation and Research. She holds a doctorate in biomedical imaging and technology from NYU Grossman School of Medicine, where she used MRI methods to develop quantitative assessments of myelin in neuropsychiatric disorders and aging. Our conversation was edited for clarity and length.
What is your research focus?
My focus is to apply advanced neuroimaging techniques in the study of neuropsychiatric and neurodegenerative disorders, such as schizophrenia and dementia. I’m interested in mapping different microstructural and biochemical features of the brain in diseases, finding biomarkers for disease progression, and tracking how the normal brain develops and degenerates across the lifespan.
For my doctoral thesis, I focused on myelin, which is the fatty substance around axons that speeds up signal transmission. For myelin imaging, I’ve been using a magnetization transfer method, which is sensitive to macromolecules in the brain. (Myelin is made of lipids, which are large molecules.)
What led you to this topic?
We chose to image myelin because it’s an important feature in the brain, and myelin deficits are present in many neurodevelopmental and degenerative diseases. We know that myelin content decreases in aging, and there is a hypothesis that earlier or abnormal demyelination is involved in people with dementia. Right now, there are no good treatments for these disorders. Sometimes I think that research or clinical trials fail because we are looking at people who already have full-blown disease, and if you test an experimental treatment only on patients with an advanced condition, you might not learn whether it’s effective at an early stage. That’s why we want to look at potential early biomarkers that are more specific to a biological mechanism of disease and then look for another perspective to target these conditions.
In your research on myelination, what did you discover?
In my thesis work, I’ve analyzed cortical myelin in healthy aging populations. We found that changes in cortical myelin density during normal aging are layer-specific and independent of local cortical thinning.
Originally my thinking was that if I could establish what a typical healthy aging brain looks like in terms of myelination in the cortex, it could help identify the points where potential high-risk dementia patients deviate from the normal trajectory. The idea seemed straightforward but the project has turned out to be quite complex, and I’m still working on it. We did develop a method to profile the cortical myeloarchitecture and found differences in cortical myelin in schizophrenia, which is another focus of my thesis.
When people think of the cerebral cortex, they often think of cell bodies that project axons into the white matter, but there’s also a fair amount of myelinated axons within the cortex that run through and across the cortical layers. We developed a method not only for looking at average myelin content in the entire cortical region but also for sampling different depths of the cortex and producing a cortical myelin profile.
In most places the cortex has six distinctive layers, and the amount of cortical myelin increases with proximity to white matter, although it’s not a linear accumulation across all layers because of differences in cell types. We produced a cortical myelin profile and quantified its shape with a “nonlinearity index.” Normally, the cortical profiles of prefrontal regions are highly nonlinear. We’ve shown that this nonlinearity is gone in schizophrenia patients. Patients in our cohort were pretty young, between ages 18 and 35, and this makes our findings more informative because we know that schizophrenia—now considered a neurodevelopmental disorder—usually has an onset in adolescence or young adulthood. That’s also the time when there is a boost in cortical myelination, when the brain really matures. We found cortical myelin differences mainly in the prefrontal region, in areas that are typically related to high-order cognitive functions. So it seems like in early disease stages there are already changes in myelin, and we think that what we found shows that the developmental deficits in myelination relate to schizophrenia symptoms.
How is imaging and measuring myelin in grey matter different from looking at myelin in white matter?
They’re similar in the sense that the myelin measurement is coming from the same magnetization transfer imaging technique, but the way you extract image values must be tailored to the underlying biological structure. In white matter, you usually look at the tracts or fiber bundles that connect different brain regions, and you need tractography or tractometry to analyze the data. For the cortex, people often take an average value of, say, the frontal cortex or some predefined functional region. But because of the special, layered structure, I think it’s really important to account for the different information at various depths. That’s what this nonlinearity index is really picking up: unique information across cortical depth. One difficulty is that the cortex is very thin, between 2 and 5 millimeters. Researchers are now going higher and higher in image resolution with high-field MRI, and I think this is a really important direction in studying the cortex.
What were some of the challenges that you had to overcome to be able to look at myelination in these cortical layers?
I studied psychology in college, so I didn’t have a strong background in physics or programming. The first challenge for me was to build expertise and understand how something like myelin can be estimated from MRI, to really understand the mechanism and limitations of this technique. And then, after I formulated the idea, the practical challenge was to build a programming pipeline to reconstruct the image to surfaces and locate each cortical depth, and then test whether I was getting meaningful values.
How did you take the step from psychology—a very different realm in terms of methods—to biomedical imaging?
I feel like I’ve taken a beeline from psychology to biomedical imaging. I studied psychology at Beijing Normal University—it’s a teachers college, and in China there’s a tradition of calling teachers colleges “normal.” I was in the cognitive psychology track, where we learn theories of attention, memory, perception. Many researchers in the field are using functional MRI to study the neuromechanisms of these processes, so it’s a very short step from cognitive psychology to cognitive neuroscience. Then I did a master’s program in psychology at NYU. We were required to do a thesis, and the program had a list of hundreds of faculty members who were running psychology-related studies. And Mariana Lazar was on that list.
Mariana Lazar, PhD, is an associate professor of radiology at NYU Langone and a neuroimaging researcher who studies psychiatric disorders. She advised your master’s thesis, later also advised your PhD thesis, and now supervises your postdoctoral research. When as a master’s student you were first looking through that list of hundreds of faculty members, how did you narrow it down to Dr. Lazar?
I filtered the list by areas of interest, like “brain” and I think I searched for schizophrenia. I was fascinated by how psychotic disorders present themselves. I always feel like it’s really difficult to study the human brain because the brain compensates so well. Even if you have a missed connection, something will make up for that. There’s a high variability in how your brain functions compared to mine.
A lot of people in the neuroscience department study animal models, but animal models are raised in the lab and they’re almost all genetically identical, so with them you can’t really mimic the kind of natural variation among human brains. With psychotic patients, although there are many different symptom patterns, they fall into recognizable categories like delusion or hallucination. I think these symptoms are great targets to try to understand how the brain fails to work as intended and then infer how it would normally work. So that was my logic behind the interest that led me to working with Mariana; because she studies human subjects and examines the living brain with MRI.
Sounds like your interest in psychology has led you toward increasingly mechanistic questions.
The main research method in psychology is behavioral experiments: subjects are prompted to do something and perform some tasks. To me, these methods feel somewhat subjective, so I wanted to anchor on something more quantitative, and that’s how I feel about MRI. It’s maybe like the difference between measuring something with the length of your palm versus a ruler. Also, when I was in high school, physics was my favorite subject and I really wanted to choose physics as my undergraduate major but my dad, who has a PhD in physics, pretty much told me not to do it. He said that if I studied advanced physics, the math would keep coming at me.
So, was the physics aspect of MRI part of the attraction?
Definitely, it was very exciting for me. I also feel that if you want to understand some cognitive process, it’s just natural to want to understand its biological basis. And when I was in cognitive psychology, I liked the concept of metacognition, which is thinking about your own thinking. If I do brain research, it’s like my brain is trying to understand itself. I think it’s so meta and so cool.
It’s been about a year since you’ve defended your PhD and become a postdoctoral fellow. Are you continuing to investigate myelination or have you branched out to other aspects of brain composition, organization, and function?
Although my thesis work focused on myelination, we also had this surprising result of seeing changes in R1, the longitudinal relaxation rate. Although some people use R1 as a myelin measure, it also reflects other features, including paramagnetic substances like iron. The observation we made is that in people with psychotic spectrum disorders, when there’s no apparent myelin change as measured by our main myelin marker, there’s very significant decrease in R1. And we were like, why? What is that?
And that’s surprising because R1 would normally be expected to reflect the amount of myelin and therefore to stay in line with magnetization transfer.
Yes, and we saw that mainly in subcortical grey-matter nuclei—like the thalamus, globus pallidus, nuclei in the center of the brain that have the highest iron content. That led us to question whether there is a change in brain iron metabolism in schizophrenia patients. Mariana recently received an R01 grant to study this, and it is now our main investigation in the lab.
Can you say more about what makes these early findings about the relationship between subcortical iron and psychotic disorders intriguing?
Iron in the brain is used in many biochemical processes, and the major ones are dopamine synthesis and myelin production. Both of these are known to be affected in psychotic disorders, and this just shows how relevant brain iron is in the neuropathology of these conditions. But so far, studies of brain iron in psychotic patients have been very limited. Diffusion MRI research has shown a decrease in white matter integrity, but that can be tricky to target for treatment. But if patients turn out to have some kind of iron deficiency, you may be able to really treat that—it’s a more specific target.
What was it like for you to go from undergraduate study in China to graduate study in the U.S.?
In my third year of college, I came to Columbia University for an academic exchange. I liked the experience and later applied to psychology PhD programs in the U.S. but didn’t get in, so I did a master’s NYU. A lot of the courses turned out to be a repetition of what I had taken as an undergrad, and I was kind of unsure of whether I still wanted to do a PhD. It was really only after working with Mariana on my master’s thesis that I decided to pursue a doctorate. MRI was a whole new world to me, and I just couldn’t wait to get into that.
It’s brave to enter a new field and have to learn new, technical tools.
It was difficult, particularly learning the Fourier transform. In my cohort, most students had backgrounds in biomedical engineering or computer science. But I think our PhD program is very nicely designed such that people like me can still get the fundamentals and catch up with the current knowledge in MRI. I think not a lot of programs do that well, so I’m very grateful for that.
Was that a bit like the math that your dad had warned would come for you?
Yeah, the math was difficult. But I think for anything you want to do, you have to have some resilience with it. It’s also very interesting, and I really wanted to learn. That helped a lot.
Earning a PhD is a milestone. At this point, what do you see for yourself in the future?
I think I’m still in the process of finding my own voice in terms of research, and I feel like that’s particularly difficult for a lot of Chinese students. In the Chinese educational system, we get really good at solving problems that are given to us, but for me it feels very challenging when you can choose for yourself. I used to focus on short-term goals like finishing the thesis or publishing a paper, but more recently—I started my postdoc in January 2024—I’ve been trying to think of new areas I want to go into and new research topics I can write a grant proposal about. I definitely want to stay in academia as long as I feel comfortable.
In both going through a doctorate and living in a new place, there’s an element of identity building. Can you talk a little bit about how your middle name connects with this?
I picked out the name Veronica when I was in middle school, so I’ve been using it for many many years. I had an English dictionary with names listed alphabetically in the back, and I was looking for something starting with Y, same as my first name, but there were few and I didn’t really know how to pronounce them, so I went to V. My Chinese name is very short, and I wanted something longer, so that’s how I chose ‘Veronica.’
When I was publishing my first paper—based on the master’s thesis that I did with Mariana’s guidance—I decided to add my English name as a middle name. My parents didn’t like that, and I used to tell them that this is just how English names work—it’s already a different convention to say the given name first and the family name last, anyway. Then I realized that this made it really easy to search for my name. There are no words in Chinese that when transliterated into English start with a V, so that makes the combination of my initials and family name very unique. I feel like Veronica is a really big part of myself now, so I have to have that.
Related Story
Mariana Lazar, medical imaging researcher who investigates psychiatric disorders, talks about recent findings, mentoring students, and learning by doing.