Decoding the Mind:
An Engineering Perspective on Neuroscience with Stanford's Prof. Jin Hyung Lee
Advice for students & aspiring neuroscientists
Question 7:
"You have a background in Electrical Engineering (BS from Seoul National, PhD from Stanford) before moving into Neuroscience. What do you think the next generation of neuroscientists should start by learning? Biology, engineering, or something else?"
“It’s hard to tell where the future will go, even in just the next 10 years,” says Professor Lee. “It’s a time of fast development. It’s important for high schoolers to know that the future is going to be different, and that the pace of change is going to be high.”
Professor Lee claims that a lot of fields will be valuable to study for those looking to get into neuroscience, from biology to chemistry to engineering. However, what is most important is computational skills, which include decomposition of problems, logical reasoning and analysis, and understanding of algorithms. She says that if there is one most important area to master to harness the future, it is data science and computational skills.
The future of brain-computer interaction
Question 6:
"A lot of high schoolers hear about 'Brain-Computer Interfaces' (BCIs) from companies like Neuralink. Your work focuses heavily on non-invasive or circuit-level understanding. Do you think we will eventually be able to 'read' or 'write' to the brain with high precision without needing surgery or invasive measures?"
“Non-invasive brain connection to cure brain disorders is the goal of LVIS,” says Professor Lee, “but ‘reading’ and ‘writing’ the brain? There are a lot of ethical issues with that, and a lot of debate around it.” But, she states, her personal belief is that all of these brain-computer procedures should be non-invasive.
LVIS & NeuroMatch
For some background:
Professor Lee founded a company called LVIS, releasing a software called NeuroMatch that assists in quickly and effectively reading and analyzing electroencephalogram (EEG) data with the assistance of AI.
Question 4:
"You founded a company called LVIS to bring your work to patients. What is the hardest part about translating a discovery from a controlled lab environment at Stanford to a real-world product that doctors can actually use?"
“The hardest part is figuring out how to get a clean signal from the real environment,” says Professor Lee, “though moving from a lab to the outside world isn’t just the hardest part.” Getting through the regulations to release the product in the first place is a significant part of the battle as well.
Question 5:
"With your platform NeuroMatch, you are modernizing how we read EEGs. Why has EEG analysis remained somewhat stuck in the past for so long, and how does your AI approach differ compared to the older technology?"
NeuroMatch works by modeling the function of brains with various disorders and uses AI to help reduce irrelevant artifacts in EEG data, detect trends in patient reports over time, and recognize and map spikes and seizures directly onto a model of the brain. This automation increases the efficiency of EEG reading compared to manual data analysis.
Modeling the brain allows for visualization of the problems so that those studying the scanned brain can know what to “fix.” Coming back to the computer analogy, this visualization allows for researchers to know: “What’s missing from the optimal circuit? How can we improve signal flow? Do we need to amplify the signal itself? Is there a wire that can be repaired?”
Optogenetic functional MRI (ofMRI) & mechanogenetic functional ultrasound (MfUS)
Question 2:
"You are a pioneer in optogenetic fMRI (ofMRI). Before this technology, it seems like we could either see the whole brain (with standard fMRI) or single neurons (with electrodes), but not both. Can you explain why bridging that gap was so critical for understanding how different parts of the brain talk to each other?"
There are two extremes to this—looking at single units (neurons) versus on a global scale (whole-brain function). However, either is simply not enough. Imagine attempting to study a computer by monitoring a single wire or an individual signal and expecting it to explain everything.
The strength of ofMRI is that it puts together multiple scales of information, from local to global, to create a thorough view of the brain.
Question 3:
"Your recent work has moved into mechanogenetic functional ultrasound (fUS). For students who are just getting used to the idea of fMRI, what does ultrasound allow you to see that MRI doesn’t? Is it about speed, resolution, or something else?"
The greatest upside of mechanogenetic fUS is that it’s much more portable than fMRI is. MRIs use large magnets that are difficult to transport. In contrast, ultrasound is more portable and can be used on people during social interaction to study empathy, complex behavior, and more in real time. However, the downside is that it is a little bit more invasive than fMRI is.
Conclusion
The world of engineering in neuroscience is largely unexplored and quickly evolving, from the technology to the frameworks of thinking. As increasingly efficient algorithms and more advanced scanning technology develops, researchers can study the brain more effectively, even in live situations. However, despite what the present has brought, the future is still unknown, and all we can do for now is prepare however we can.
The engineering perspective on neuroscience
Question 1:
"I read that your lab’s mission is to understand how the brain works at a systems level. How does viewing the brain from engineering perspective vs a biological perspective potentially change how you approach diseases like Alzheimer's, Parkinson’s or epilepsy compared to a traditional biologist?"
“There are two different things to consider,” says Professor Lee. “Firstly, we analyze the system with engineering principles, but we also use engineering technology.”
Biology focuses on the testing of hypotheses; in contrast, engineering centers around designing and testing systems, checking input and output, and observing the organization of the system (both the designed systems and the brain) as a whole.
There are different lenses through which problems are viewed overall. In Alzheimer’s, for instance, biology studies various chemical and biological interactions within the organism, from the mechanisms of drugs to Aβ plaques (misfolded protein fragments that build up in the brains of Alzheimer’s patients). Engineering will instead frame the brain as a system with
Introduction
In a rapidly developing world of new technology, what role does engineering play in neuroscience? In what direction are brain-computer interfaces moving today, and how can students prepare for the future this progress will pave?
To learn more about the answers to these questions, I recently interviewed Professor Jin Hyung Lee, an Associate Professor of Neurology and Neurological Sciences, Bioengineering and more at Stanford University. She leads the Lee Lab at Stanford, which focuses on understanding the brain at the systems level and studying it through an engineering perspective rather than a purely biological one. The Lee Lab has done research on the circuits underlying disorders within the brain, developed technologies to take closer looks, and used brain modeling all throughout with the goal of curing brain disorders and diseases.
Professor Lee and I spoke about the engineering approach to neuroscience, the hardware and software she uses for her work, and what aspiring neuroscientists need to know in order to prepare for the rapidly changing future.
Professor Jin Hyung Lee
ASSOCIATE PROFESSOR OF NEUROLOGY AND NEUROLOGICAL SCIENCES (NEUROLOGY RESEARCH), OF NEUROSURGERY AND OF BIOENGINEERING
AND, BY COURTESY, OF ELECTRICAL ENGINEERING
← Blog Post 19
Blog Post 21 →
To be continued