Research Interests

Last updated: August 2026

Current

Life experiences are transient, yet they often leave lasting footprints on the mind and brain. Two people can live through nearly identical moments and come out changed in entirely different ways. Why are you and I different? And before I can even ask that seriously, what does “different” mean in the first place? These two questions are where my current interest begins.

a. Why do individuals weigh the same experience so differently, to the point where one person forgets it and another carries it for years?

I keep noticing that what we perceive is never a direct copy of the world. it’s already a reconstruction built from whatever the senses picked up in the first place. From there, attention keeps deciding what gets amplified internally and externally. Executive control governs how that balance between external and internal attention shifts over time, and together I think of these as the main levers behind how much weight an experience ends up carrying. Some events are shed almost immediately, others become impossible to let go of. I want to understand why individuals differ in how these levers operate, and I’d like to dig further into vulnerability and resilience through that same lens.

b. How should we represent that weight-assignment process so that it's something we can actually interpret?

State trajectories, network dynamics, modular organization, gradients: there are several candidate languages for describing how the brain processes and updates in response to experience over time. Each is really an attempt to turn raw brain data into something interpretable, a format that keeps enough of the underlying signal to say something about behavior. While developing a cross-atlas connectome transformation model, I found that the choice of representation is not neutral: which one you pick changes how much brain-behavior association, and phenotypic information more broadly, survives the transformation. I want to try a range of approaches like these to get a fuller picture of how this cognitive process actually unfolds in the brain.

c. Of what we call individual differences, how much is a shared computational principle expressed differently across people, how much is genuine individual signal, and how much is simply noise?

These three may never be fully separable. What looks like noise in one analysis can turn out to be someone else’s signal, and a lot of what gets called “individual” may just be a shared principle wearing a different face. If people were simply different, with no shared basis at all, thought and behavior would look random. They don’t: there’s a shared principle underneath, and we diverge from it just enough to become who we are. Any claim about what makes someone different also depends on what we’re calling shared or normal to begin with, so defining that baseline is crucial. What interests me is that the line between the two doesn’t stay put. It holds fairly steady in basic sensory processing, and loosens considerably in higher-order cognition, like interpreting a story or reasoning about someone else’s mind. I want to know where the shared principle still dominates, where individual signal takes over, and what’s actually happening in the brain when it does.

Precision approaches are only as good as the data they’re built on. What raises that ceiling, I think, is a combination of deeply sampled datasets with larger subject numbers and methods that help us compare across people without smoothing away the very differences we’re trying to find. I hope to eventually contribute to better-timed and more personalized mental-health intervention. But that’s a destination, not the starting point. I’m also open to the means shifting, toward brain-computer interfaces, or further toward building the large-scale datasets this kind of question actually requires. What’s settled is that I want to study this in the human brain with human MRI.

Brief descriptions of my current projects: Click ‘Current’ section of the Project menu.

Post-master's (early)

a. How do individual differences in cognitive state transitions shape how people interpret and adapt to internal and external experiences?

Even when facing identical environments, people perceive, evaluate, and internalize experiences differently. These variations likely arise from distinct configurations and transitions among cognitive processes such as attention allocation, mind wandering, and executive control. Cognition dynamically interacts with external inputs and gradually alters individual’s priors (or biases) carved as large-scale brain networks. I aim to understand how differences in cognitive state transitions, especially those governing internally oriented attention, give rise to stable behavioral and emotional patterns across individuals.

b. How can biologically grounded, multivariate representations of cognition capture latent individual differences?

Moving beyond categorical descriptions, I seek to represent individuals along latent cognitive dimensions derived from multivariate modeling of brain–behavior association. Such representations can reveal continuous axes of phenotypic variability and give cornerstone of neural mechanisms underlying individual uniqueness in cognitive dynamics described in question a. By anchoring these representations in biological substrates, I hope to study representations that are both predictive of individual behavior and interpretable in terms of underlying brain network organization.

c. How can integrating latent cognitive dimensions with brain network dynamics advance interpretable and personalized interventions?

By embedding latent cognitive factors within the dynamic architecture of brain networks, we may obtain more interpretable and biologically grounded markers of mental function. This integrated approach may offer interpretable markers of mental function that inform when, where, and for whom an intervention is most effective. Ultimately, my goal is to refine theoretical models of cognition while contributing to personalized strategies for mental-health assessment and intervention.

Second half of the master's program

My research began to take a new direction toward human neuroimaging and network-level approaches. It felt like entering an entirely new world. I didn’t even know how to open a raw T1 file at first. Everything felt foreign and overly abstract. But as I learned and practiced, I realized that what initially felt abstract was simply a different way to describe the same brain from a network/system level. It was interesting to see how distal brain regions synchronize and collectively generate human behavior as an output. Working with human MRI felt closer to real life. Instaed of focusing on a single circuit, I learned to analyze structural networks and closely followed functional analyses from non-invasive human MRI data.

Detailed information about the Parkinson’s disease project with Yonsei Severance Hospital: Click here

How would the individuals behave, think or feel? What would it mean to live with reduced drive or emotional flatness?
And how could those phenotype originate from the brain's architecture?

While working on this project, I began to think deeply about disease complexity and individual heterogeneity. Patient with the same diagnosis could differ significantly in their symptoms and experiences. When I saw columns labeled ‘apathy’ or ‘motivation’ in the dataset (represented as numerical scores), I found it difficult to envision what those numbers actulally represented in an actual human being. Even after reading the questionnaires that was used to derive those scores, I kept wondering what these humans with that score/lables felt like from the inside. That gap between quantified measures and realities became one of the most interesting and unresolved questions for me.

First half of the master's program

In the early part of my master’s program, I was interested in how how the brain distributes energy with precise spatiotemporal coordination. Neurons fire within milliseconds, while vessels adjust over seconds, yet the resulting hemodynamic signal remains coherent across space and time.

How is such temporal precision and region specific coordination maintained in the neurovascular system?
What happens when the mechanical condition of this system begin to fail?

These questions led me to view neurovascular coupling as a dynamic control problem embedded in a hierarchical vascular network. I learned that the hierarchical organization of the brain vessel especially interesting. Vessels close to proximal branches (or small-order branches) are surrounded by smooth muscle cells with active contractility, but as vessels bifurcate into smaller perforating branches that penetrate deeper into the parenchyma, they are no longer wrapped by smooth muscle cells but by less contractile pericytes and even endothelial cells. This graudal loss of contractility makes deeper vessels respond more paassively. Then how does a neuronal response occurring deep in the tissue communicate upstream to secure the needed blood supply?

Before tackling these big questions, I wanted to start by examining one aspect of the system that mechanical properties of vessels being the boundary conditions for neurovascular coupling. I focused on how mechanical stress (originating from arterial stiffness) could distort the fine balance between neuronal demand and vascular response, even when pathology didn’t originate in the brain. Although I had to discontinue this work after developing severe mouse allergies, this experience was a turning point for me to shift toward network neuroscience.

For early stage ideas of that time: you can take a look here.

Undergraduate years

I majored biology, studying fundamental disciplines such as biochem, genetics, molecular, cellular, and developmental biology. My first exposure to neuroscience came during a short summer session at UC Berkeley, where I found how differently the brain could be studied compared to the biology I have known.

Later, during a fully funded exchange semester, returning at Berkeley, I took Genetic Factors of Neuropsychology, and this course changed how I thought about science. Until then, I had viewed “intervention” only through pharmacological means. But learning about the transition from DSM-IV to DSM-5 and how redefining autism as a spectrum could expand access to care and insurance coverage, was a eureka moment. I realized that research doesn’t need to end with a prescription to have an impact. Sometimes, understanding a disorder more accurately and carefully measuring its risk factors could directly improve quality of lives. That was the moment I began to see neuroscience not just as molecular mechanisms of the brain, but as a way to connect scientific understanding to real human outcomes.

At that time, I was obsessed with the idea:

What makes something truly interventible? Are there modifiable and systmic factors that can indirectly have an impact on brain health in broad populations?

Returning to Korea, I began my first full-time internship with this question in mind. For detailed project description: click here

Through this project, I began to study brain vasculature and realized how dynamic it is. I learned that blood flow can contribute to the clearance system (such as glympathic pathway). And the blood flow are altered by neurovascular coupling and pathological proteins. Growing interest in coordination between neural activity and local blood supply drove me to apply master program.