Changing the Game
How biomedical engineer and neuroscientist Lilianne Mujica-Parodi is revolutionizing the study and treatment of brain disorders
By Meghan Goff
As Jan Ellison Baszucki looks back on her son Matt’s journey with bipolar disorder, sadly, she knows their story is not unique. At 19 years old, Matt suffered a manic episode. In the years that followed, he was hospitalized four more times, treated by 40 mental healthcare professionals and prescribed nearly 30 different medications during a five-year period, and still, according to Jan, he was struggling to manage his own mind.
“Matt’s experience sent me on a journey to understand what was happening in his brain.
What was causing these symptoms?” she recalled. “What I found out was that we don’t know.”
For years, the complexities of brain-based disorders — including bipolar disorder, dementia and schizophrenia — have left millions of families, like the Baszuckis, searching for answers. And despite incredible advancements in modern medicine, understanding these mysteries of the mind has lagged behind traditional medicine, posing an incredible challenge — and an urgent one — for the physicians and researchers tasked with finding a cure.
Now, one Stony Brook University researcher is answering this call for help with a bold, new solution: A blueprint for individualized care that is poised to disrupt clinical neuroscience and revolutionize mental health, all inspired by an unlikely source — the Baszuckis themselves.
An Inspired Approach
It was the Baszuckis, Jan and her husband, David, who sought out Lilianne “Lily” Mujica-Parodi, PhD, a professor in Stony Brook’s Department of Biomedical Engineering and director of its Laboratory for Computational Neurodiagnostics, who also holds joint appointments in Stony Brook’s Graduate Program in Neuroscience, Department of Psychiatry, Department of Neurology and Department of Physics. Their interest was piqued after learning about Mujica-Parodi’s first-of-its-kind study exploring the role of ketosis on brain functioning. This was an area of particular significance for the Baszuckis, as it was a ketogenic diet that — when added to other treatments — put their own son’s bipolar disorder into remission.
“Here was a neuroscientist unveiling the mechanism by which ketones work to stabilize brain networks,” Jan recalled. “This explained why a ketogenic diet gave our son his mind and his life back. We had to wonder, could building on this knowledge be the first step toward helping others suffering from mental illnesses?”
Our family believes Neuroblox’s impact on understanding and treating brain-based disorders, including mental illness, will be transformative.
— David Baszucki, founder and CEO, Roblox
By early 2021, Mujica-Parodi and her team had developed an electrical engineering-inspired model of the prefrontal limbic system — which plays an important role in the regulation of emotion — to understand anxiety disorders, an area closely tied to her interest in the role neural control circuits play in neurological and psychiatric disorders. Hearing Mujica-Parodi speak about this study on a podcast, the Baszuckis were fascinated by the connections she had made, and they hoped she could take that same approach to study bipolar disorder, specifically.
“Throughout my career, I’ve moved between fields to get the tools I need to answer the questions I’m interested in,” Mujica-Parodi said. “My approach to understanding psychiatric disorders is grounded in questions of how neurobiological and physiological function maintain stability in the face of chaotic environments; exactly what it means for biological control circuits to be dysregulated in different ways; and why each of those different types of dysregulation leads to a disorder’s particular set of signs and symptoms.”
This engineering approach to psychiatric diseases resonated particularly with David Baszucki, a computer scientist and electrical engineer by training and, most notably, the founder and chief executive officer of the immersive communications platform Roblox.
I thought, if users can translate the principles of physics to become creators, then clinical neuroscientists, physicians and pharmacologists might be able to use Neuroblox in the same way: to create circuits, run simulations and, thereby, to develop testable hypotheses and optimize treatments in meaningful ways.
— Professor Lilianne Mujica-Parodi
The widely popular interactive platform, where users can design their own immersive experiences in a collaborative and creative space, began as Interactive Physics, a 2D physics engine intended to teach children physics by simulating physical phenomena.
Now known as Roblox, it has evolved into the global sensation it is today by empowering its users with the tools and the technology to create their own immersive experiences —from games to simulations to social experiences — and share them with others around the world.
This accidental innovation meant that the complex programming typically required for building video games could be broken down into a system of “blocks”— premade pieces of code that could be combined to create games and be manipulated by anyone, even those with no prior experience in programming.
It was David’s brainchild that quickly inspired Mujica-Parodi’s own research, even sending her back to the drawing board to tackle the challenge of bipolar disorder on a larger scale with a solution that — if executed correctly —could be applied to brain-based disorders more broadly.
With the right tools, she could determine the brain’s modular functional components, or “blocks,” allowing neuroscientists to visualize and streamline the identification and simulation of brain circuits through a revolutionary new approach: Neuroblox, a software platform for modeling neural circuits and their regulation.
Just as Roblox relies on developers to create user-generated content for games and experiences available on the platform, putting a Neuroblox platform out into the world would allow researchers worldwide to share, refine and test each other’s models across populations and disorders, providing seamless integration of data that informs models and experiments.
According to Mujica-Parodi, this could open a world of possibility, accessible even to neuroscientists who do not have training in computational approaches.
“I thought, if users can translate the principles of physics to become creators, then clinical neuroscientists, physicians and pharmacologists might be able to use Neuroblox in the same way: to create circuits, run simulations and, thereby, to develop testable hypotheses and optimize treatments in meaningful ways.”
But embarking on a groundbreaking endeavor like Neuroblox would require help. While the “laws of physics” are well-established, the “laws of neuroscience” are still mostly unknown, according to Mujica-Parodi.
“Right now, there is a disconnect between the aims of clinical research and the computational tools we have to exploit that research,” she said. “Neuroblox is doing something fundamentally different. It’s trying to bridge that gap.”
As Mujica-Parodi unpacked this new approach, she quickly realized that the potential for impact extended far beyond bipolar disorder. She was no longer creating just one solution by taking a circuit-based approach to the problem. She was developing an infrastructure that could be applied to brain-based disorders more widely — a range of diagnoses from Alzheimer’s disease to anorexia nervosa.
“My process of figuring out one specific piece of the prefrontal limbic circuit took nearly a decade, so if I were to approach a different circuit or disorder, I would want to automate that scientific process in ways that would markedly accelerate how we harness data to create and test models,” she explained.
One of the key advantages of Neuroblox is it would allow researchers to work with an artificial brain, creating an opportunity to test the full set of alternative hypotheses all at once, a breakthrough that could provide an incredible benefit for the “trial and error” method to treating brain-based disorders that many patients face, according to Mujica-Parodi.
What is currently lacking in the treatment of brain-based disorders is a clear understanding of which interventions — from lifestyle changes like therapeutic nutritional ketosis to medications and supplements — will work, in what combination, and for which patient. With Neuroblox, Mujica-Parodi and her team are working to create the tools that will give doctors the knowledge and confidence to identify the best course of treatment in a matter of minutes.
“When dealing with the complexities of brain-based disorders, innovations like this can help support the evolution of crucial research,” said Harold Paz, MD, executive vice president for health sciences, Stony Brook University, and chief executive officer, Stony Brook University Medicine. “We are encouraged by the incredible potential of Neuroblox, and we are immensely grateful to donors like the Baszuckis who help support Stony Brook’s commitment to research and encourage innovations that can lead to improvements in the lives of our patients here and around the world.”
For the millions of families struggling to understand the complexities of brain-based disorders, a breakthrough of this magnitude would not only provide a clear treatment path, but it also would alleviate their worry and uncertainty and replace it with something much better — hope.
“When a patient has a psychiatric episode, you don’t know what’s going to work, so many times they’re given multiple medications and encouraged to try a host of interventions in the hopes that at least one will reduce symptoms. The problem is that each of these has its own potential side effects, and they can mutually interact in sometimes surprising ways,” Mujica-Parodi said. “Now, we’re creating a way to test hundreds, or even thousands, of hypotheses in parallel. This means we can eliminate options through computation instead of just hoping we’ve made the right guess.”
By building a computational neuroscience model of a brain using patient-specific inputs, and then developing and testing drugs or different treatment strategies in that model, Neuroblox would create an opportunity to predict how that brain would respond to various interventions and at what rate over a particular timescale.
“If we just worked to identify the circuits and their dynamics that underlie metabolic dysregulation in bipolar disorder, then on the one hand, any success would be a fantastic achievement,” said Mujica-Parodi. “But on the other hand, anything we were able to achieve would only be applicable to this very specific problem.
“However, if we also worked to create tools flexible enough to model and validate any neural circuit relevant to any brain-based disorder, as well as to compare and combine our models with those of other scientists, it could vastly accelerate our progress as medical researchers, both individually and collaboratively,” she added.
During the past two years, the Baszuckis have become essential partners in bringing Mujica-Parodi’s vision to life through key investments totaling more than $6.2 million to fund breakthroughs and endowed research in neuroscience.
Thanks to an initial research grant supported by the Baszuckis, Mujica-Parodi has begun mapping this first-of-its-kind blueprint to provide individualized treatment for a host of brain-based disorders.
“We were drawn to Lily’s vision for neuroscience because it is analogous to the approach we took at Roblox, providing a platform that enables millions of independent developers around the world to build unique experiences for our users,” David said.
“With Neuroblox, Lily is building a software platform where neuroscience researchers worldwide can refine, test and share models to help us understand how the brain regulates energy — a critical driver of mental health,” he continued.
Matt’s experience sent me on a journey to understand what was happening in his brain. What was causing these symptoms?” she recalled. “What I found out was that we don’t know.
— Jan Ellison Baszucki
This work is possible only with the support of a diverse, highly multidisciplinary team comprising leading neuroscientists, computer scientists, programmers and engineers at the top of their fields.
“When you move between fields as much as I do, you have to rely on people with much greater domain expertise than you have,” Mujica-Parodi said. “Once I started to think through what pieces of the puzzle would be necessary, I wanted to find the best people in the world in each of these key areas to attack the problem most efficiently.
“I hadn’t worked with any of these scientists before — but luckily, they were as excited about this project and its potential as I was, and we were able to forge an amazing team very quickly.”
While this latest partnership spans institutions, it undoubtedly builds on an inherent culture of collaboration at Stony Brook, said Mujica-Parodi, one that has informed her work and some of her proudest accomplishments since joining the institution in 2003.
“The research that I’m most passionate about, and which has consistently garnered some of our most highly competitive funding mechanisms and provocative results, tends to be ambitious and highly multidisciplinary,” she said.
“As a faculty member, my research group and I have benefited from the fact that the departments of Biomedical Engineering, Neuroscience, Computer Science and Physics, as well as the medical school and hospital, are not only physically close but also administratively linked via the Center for Engineering-Driven Medicine,” she explained. “I believe our ability to bridge fields reflects something unique and essential about Stony Brook University.”
In addition to their initial $3.2 million investment for the research grant, Jan and David have supplemented that support with $3 million to establish the Baszucki Endowed Chair for Metabolic Neuroscience, which will recognize an exceptional researcher in neuroscience. Based on her outstanding track record of cutting-edge research in the field, Mujica-Parodi will be the inaugural holder. The $3 million gift from the Baszuckis will be enhanced by an additional $550,000 from Stony Brook’s Presidential Innovation and Excellence Fund. These funds are designed specifically to accelerate the university’s highest ambitions.
“Lily’s innovative approach to one of our most pressing societal issues — our mental health and well-being — is inspiring. It underscores our commitment as an institution to advance knowledge that will have a long-term, significant impact on the world,” said President Maurie McInnis. “We could not be prouder of these efforts, and we are thrilled that the Baszuckis have chosen to invest in Lily’s trailblazing work in a way that will undoubtedly change lives.”
As Mujica-Parodi and her team work to map and better understand these circuits, they hope that Neuroblox, like the platform it was modeled after, will continue to evolve and be strengthened by the contributions of others. For scientists, researchers and physicians, this means building on the latest knowledge and hypotheses to accelerate the power of these models to advance cures in the next five to 10 years and, ultimately, establish a more direct pathway to individualized treatment of brain-based disorders.
For the scientific and medical communities, Mujica-Parodi’s work has the potential to be radical, groundbreaking and even revolutionary. But for countless other families just like the Baszuckis, it could be life-changing.
“We are so fortunate that our son is thriving, so for us, it’s no longer about Matt’s health or even just bipolar disorder,” said Jan. “This is a frontier — maybe even the last frontier — of neuroscience, and Lily’s vision is moving the field forward so that other families don’t have to fumble in the dark to find interventions that work the way we did.
“If we imagine an underlying model that unifies these kinds of disorders, then suddenly there’s a way to not just understand these mechanisms in the brain, but to identify effective treatments. Now we have an idea that could improve the lives of millions of people around the world,” she said.
Meghan Goff is a communications consultant and founder of Onword Communications. Her focus on impact storytelling has helped a growing number of organizations create compelling, inspiring and impactful communications, all of which reflect their mission, vision and priorities.
Experts at Work: Building Neuroblox From the Ground Up
Professor Lilianne Mujica-Parodi has proposed a platform that provides user-friendly “blocks” of biologically meaningful neural circuit components, which can be manipulated and assembled in ways that provide outputs with which we can compare data.
Neuroblox is more than a concept; it’s a collaborative effort, bringing together some of the brightest minds in computing, neuroscience, biomedical engineering and beyond. Meet the team bringing this groundbreaking vision to life:
Alan Edelman, PhD
Professor of Applied Mathematics, Computer Science and AI Laboratories, Massachusetts Institute of Technology (MIT); Chief Scientist, Julia Computing Inc.; Co-Founder and Research Group Leader, Julia Lab, MIT
Optimized for high-performance computing, Neuroblox’s complex and interactive programming is possible only with the Julia programming language, a high-level, dynamic programming language for computational science co-created by Edelman.
Richard H. Granger Jr., PhD
Professor, Department of Psychological and Brain Sciences, Thayer School of Engineering, Dartmouth University; Director, The Brain Engineering Library, Dartmouth University
Granger is a computer scientist and neuroscientist, one of the few in the world working on computational primitives and microcircuits performing precise computations at the lowest level. Constrained by data, these modular “building blocks” will enable the flexible creation of a diverse array of circuits, scaling up brain models from the bottom up.
Earl K. Miller, PhD
Picower Professor of Neuroscience, The Picower Institute for Learning and Memory, MIT
Miller is a world expert on the cortico-striatal circuit, one of the first multiscale (micro-to-macro scale) circuits to be incorporated into Neuroblox. His research is on the neural basis of executive brain functions, and the ability to carry out complex mental processes has laid the groundwork for building more detailed accounts of dysfunction in brain disorders.
Helmut H. Strey, PhD
Associate Professor, Department of Biomedical Engineering, Stony Brook University
With a background in physics, Bayesian data analysis, physics and computer science — including the Julia programming language for high-performance computing — Strey is the lead software engineer for Neuroblox. His work involves the integration of computational neuroscience into the unique coding architecture required for Neuroblox’s modular design. Coded in Julia, Neuroblox is designed to be a comprehensive open-source tool for simulation and data analysis and a predictive tool for modeling brain circuit dynamics. As part of the software development, Strey’s team is also developing a graphical user interface for Neuroblox to broaden access to the tools without the need to write code.
Neuroblox and Its Potential for the Field
Computational neuroscience has the potential to advance psychiatry beyond a description of features. By running simulations with different initial conditions and using systems of coupled differential equations, it will be possible to test the power, accuracy and reliability of models in silico, including pharmaceutical interventions.
The advantage of this approach is that it will give the brain what Newtonian mechanics provided for physics — not only figuring out the difference in features between populations X and Y but also the fundamental rules governing the brain and its disorders.
But computational neuroscience requires deep expertise in many areas, including physics, electrical engineering, computer science and pure mathematics. This kind of breadth is rare, one critical reason “computational psychiatry” has yet to gain traction.
Mujica-Parodi has set out to create a collaborative platform, one that will allow researchers worldwide to share, refine and test each other’s models while seamlessly integrating data, models and experiments. While some tools exist today, none is as user-friendly, comprehensive or powerful as what Mujica-Parodi is proposing.