USC researchers have developed an approach that uses artificial intelligence to generate detailed maps that highlight differences in how distinct parts of the brain age.

The new model also sheds light on how patterns of brain changes correlate with changes in cognitive function across the lifespan, according to a USC study published Monday in the journal Proceedings of the National Academy of Sciences.

The researchers, led by Associate Professor Andrei Irimia of the USC Leonard Davis School of Gerontology, used magnetic resonance imaging from nearly 15,000 cognitively healthy individuals to train a deep learning AI model. The data provided a baseline against which the model could measure “local brain age,” or how old specific regions of the brain appear. When the AI model was then used to analyze MRI images from people with mild cognitive impairment and Alzheimer’s disease, it revealed distinct patterns of accelerated aging in brain regions known to be affected early in neurodegeneration.

While most studies of brain age measure this phenomenon using a single number, the new model provides a much richer picture of typical aging and neurodegeneration. Rather than assigning a single “brain age” to an individual, the approach generates a detailed map showing how old different parts of the brain appear relative to what is typical for someone of the same chronological age.

“Not all brain regions age at the same rate,” Irimia said. “Some areas appear to be more resilient, while others are more vulnerable to aging and disease. By measuring local brain aging, we can identify where the brain is aging faster than expected and how those changes relate to cognitive function.”

Brain age as a biomarker

The research builds on previous efforts to estimate brain age, an emerging neuroimaging biomarker that compares a person’s brain structure to patterns seen in healthy people across the lifespan. Traditional methods typically reduce the brain to a single age estimate, which can obscure important regional differences. The new approach instead measures local brain age at the voxel level — the three-dimensional units that make up an MRI scan — producing a much more detailed picture of structural aging throughout the brain.

“This more nuanced understanding of how the brain ages could pave the way for earlier identification of dementia, a better understanding of what factors affect risk and new ideas for treatment approaches,” Irimia said.

To develop the model, the researchers trained a deep-learning neural network using MRI scans from 14,748 cognitively normal adults ages 19 to 100 drawn from six large public datasets, including the UK Biobank, the Human Connectome Project and the Alzheimer’s Disease Neuroimaging Initiative. They then tested the model using MRI scans from more than 1,900 additional participants in the Alzheimer’s Disease Neuroimaging Initiative, including cognitively normal adults, people with mild cognitive impairment and people with Alzheimer’s disease.

Across healthy adults, the model consistently found that the frontal and temporal lobes — regions involved in decision-making, memory and other higher cognitive functions — appeared biologically older than the parietal and occipital regions, which are involved in spatial awareness and sensory processing functions. The researchers also found that the brain’s right hemisphere tended to show slightly more advanced aging than the left, a pattern that persisted regardless of whether participants were right- or left-handed.

As cognitive impairment progressed, the differences became even more pronounced. Compared with cognitively normal adults, participants with mild cognitive impairment or Alzheimer’s disease showed significantly older local brain ages in structures that are among the first affected by Alzheimer’s pathology, including the hippocampus, amygdala and several deep brain regions involved in memory and cognitive processing.

The researchers also found that older local brain age was associated with poorer performance on cognitive assessments, strengthening the link between structural brain changes and real-world function. The strongest relationships appeared in people with Alzheimer’s disease, suggesting that regional brain aging may become increasingly informative as neurodegeneration advances.

What’s ahead: Brain aging

Because the model produces anatomically detailed maps, it could eventually help scientists better understand why some people experience faster decline in specific cognitive abilities than others. The approach may also prove useful for tracking disease progression or evaluating whether experimental therapies are slowing degeneration in targeted brain regions.

Although the findings are promising, Irimia emphasized that the method remains a research tool. The model was trained primarily on research-quality MRI data and will require additional validation using more diverse clinical datasets before it can be adopted in routine patient care. The study also relied largely on cross-sectional data, meaning that future longitudinal studies will be needed to determine whether local brain aging can reliably predict who will progress from healthy aging to mild cognitive impairment or Alzheimer’s disease.

Still, the researchers believe that moving beyond a single measure of brain age represents an important advance for neuroscience.

“Brain aging isn’t uniform,” Irimia said. “By understanding how individual regions age, as well as how those patterns differ from person to person, we’re moving toward a much more precise understanding of healthy aging and neurodegenerative disease. Ultimately, that could help us identify people at risk earlier and develop more personalized approaches to preserving brain health.”

---

About the study: Irimia’s co-authors include first author Nikhil N. Chaudhari, Owen M. Vega Huerta, Samayan Bhattacharya and Nahian F. Chowdhury, all of USC. The study received support from the National Institutes of Health (R01 AG 079957 to Irimia), the Hanson-Thorell Family Research Scholarship Fund, the Center for Undergraduate Research in Viterbi Engineering (CURVE) at USC and from anonymous donors.

About 10 years ago, when Jason Kutch first began surfing, he paddled out to ride the waves in considerable pain. Then a postdoctoral researcher at the USC Viterbi School of Engineering, Kutch had suffered for years with several chronic pain conditions, including migraines, low back pain and pelvic pain.

But after each novice surf session, he experienced an extraordinary shift.

“I got out of the water and I was like, ‘Where did the pain go?’” says Kutch, now a professor in the USC Division of Biokinesiology and Physical Therapy at the Herman Ostrow School of Dentistry of USC.

Each time, the pain didn’t return for several days. “Once I recognized how stable and reliable that pattern of pain relief was, it took all of the anxiety out of chronic pain,” Kutch says. “Finally, there was something that I could do to control it.”

That revelation touched off a decade of research into the neurobiology of pain and the promise of surfing as a therapy. Early on, Kutch gathered data at the shoreline, tracking chronic pain sufferers’ self-reported pain before and after surf sessions.

Today, his research participants don’t zip into wetsuits or paddle into the Pacific Ocean. They don VR headsets and catch digital waves at the USC Health Sciences Campus, thanks to an immersive surfing simulator Kutch and his collaborators designed and built in Kutch’s basement lab with support from the Southern California Clinical and Translational Science Institute at the Keck School of Medicine of USC. Users sit, kneel or stand on a surfboard mounted atop a motion platform that responds to movements they make in the virtual seascape.

“It simulates the momentary feeling of weightlessness you get when you’re caught by the wave, and then you can slide down the wave and carve back and forth,” Kutch says. “But unlike waves in the real world, in VR we can keep them perfect and going on forever.”

The project is one of several endeavors led by USC researchers that leverage realistic wave simulations to advance scientific discovery about ocean-related phenomena, from surfing to tsunamis. These technologies include both human-made waves in wave pools and interactive digital wave experiences, offering unprecedented opportunities for scholars to study — and users to immerse in — waves without the potential dangers, accessibility barriers and unpredictability of the ocean itself.

Making waves

“The technology of artificial waves goes back over 100 years,” says Peter Westwick, professor of the practice of thematic option and history at the USC Dornsife College of Letters, Arts and Sciences, and co-author of The World in the Curl: An Unconventional History of Surfing. Westwick cites such notable attempts as a 1903 wave pool in Germany that used steam-driven mechanical agitators and a hydraulic-propelled wave machine in Tempe, Ariz., that introduced America to its first surfable artificial waves in 1969.

In recent years, wave simulations have become increasingly sophisticated — in part due to breakthrough digital and mechanical technologies pioneered by USC researchers.

Adam Fincham ’89, PhD ’94, adjunct research associate professor of aerospace and mechanical engineering at USC Viterbi, collaborated with professional surfer Kelly Slater to design what is widely regarded as the world’s most perfect wave-pool wave. The hydrofoil system Fincham engineered displaces water in a way that closely approximates the ocean’s natural wave-creation force. A submerged 100-ton plow is dragged through the water by a vehicle on a track adjacent to a manmade lagoon, creating a swell of water.

The Kelly Slater Wave Co., where Fincham has been the chief scientist since 2010, debuted the technology in 2015 in Lemoore, Calif. — more than 100 miles inland. They turned a former artificial waterski lake into the Surf Ranch, a practice and competition hub for professional and aspiring surfers the world over. The team used computer simulations to design the contours of the lake floor, which, like a shallow reef in the ocean, coaxes the swell created by the hydrofoil to break into a surfable wave.

Slater’s signature configuration is a six-foot barreling wave that travels more than 2,300 feet, allowing for rides longer than one minute. Wave preferences can be tailored to the skill level of each visitor. The company’s technology also powers Surf Abu Dhabi, which opened in the United Arab Emirates in 2024 and holds the Guinness World Record for the world’s highest wave-pool wave (about 12.3 feet).

For Fincham — an expert in geophysical fluid dynamics, which is the study of flow and motion in large bodies of liquid — the Surf Ranch offers a unique scientific testbed. For the past several years, he and his collaborators have used the wave generator to make new discoveries about how the wind shapes waves.

Fincham notes that studying the effects of wind in the ocean can be challenging because natural conditions continually shift, and every wave is different. “The Surf Ranch serves as a laboratory where you can have the exact same wave again and again to perfect your measurements,” he says.

Waves without water

Advances in computing have paved the way for digital waves that look and behave like the real thing. To create a lifelike VR surfing experience, Jason Kutch collaborated with Heather Culbertson, associate professor of computer science, biomedical engineering and aerospace and mechanical engineering at USC Viterbi. Culbertson is an expert in haptics, which infuses virtual environments with tactile, force and motion sensations. The surfing simulator project has expanded her lab’s work into designing multisensory experiences that meld tactile and motion cues with visual and auditory ones.

Premankur Banerjee, a computer science doctoral student in Culbertson’s lab, developed the motion platform hardware that shifts the surfboard in space. He used a complex technique called motion mapping to coordinate the board’s movement with what users are doing and seeing in VR. Without this integration, riders would quickly get motion sickness.

Culbertson’s team designed algorithms to customize the behavior of the waves in Unity, a video game development engine. “Unity’s Crest engine only handles basic wave physics, so we’ve been doing a lot of adjustments in order to get to a wave that’s actually surfable,” Culbertson says. The motion platform is good at simulating waves up to two feet high, comparable to what you might see at many iconic surf breaks, Kutch says.

Kutch worked with Culbertson’s team to design the multisensory elements of the virtual seascape: a vivid coastal scene complete with leaping dolphins, swaying palm trees, sea caves and even a pirate ship. Users feel the wind in their face from a fan that tailors the force of the gusts to users’ velocity in the VR environment.

Paddling around, they hear the slosh of the water and the calls of seagulls; if they dive off the board and go underwater, the soundscape is muffled as if through liquid. Culbertson’s lab is developing haptic gloves to make users feel like their hands are touching and displacing water.

“We spent a lot of time making sure that the virtual ocean environment was engaging enough that we could compare the effect of surfing waves versus just being on the water and paddling around,” Kutch says. “We can really dial in on exactly which part of the experience affects neural activity.”

Kutch’s preliminary data suggests that surfing in VR affects a measure of brain activity called peak alpha frequency, which quantifies the speed of the brain’s resting oscillation.

“People’s brains oscillate at slightly different frequencies, and this baseline frequency is reliable and stable over time,” Kutch says. “The lower it is, the more pain-sensitive you are.”

Early data show that those who enter the VR surfing simulator with a low peak alpha frequency experience a temporary jump in this measure after completing a surfing session. Kutch hypothesizes that this boost may be one common mechanism contributing to the benefits reported in studies of surfing-based interventions for pain and other conditions, including depression, post-traumatic stress disorder and autism. These conditions have also been associated with lower peak alpha frequency, suggesting one possible avenue for future research.

Warning: Tsunami ahead

Though waves hold therapeutic promise, they also have destructive power. Last October, Patrick Lynett, a professor of civil and environmental engineering at USC Viterbi, traveled with a group of researchers to a remote Alaskan fjord to study the aftermath of a mega-tsunami caused by a massive landslide.

The tsunami began the morning of Aug. 10, 2025, after a chunk of rock with the volume of a small city detached from a mountainside adjacent to a glacier and slid into the sea. Like an anvil dropped into a bathtub, the rock made a colossal splash, generating a giant wave that ran nearly 1,600 feet up the mountainside on the opposite side of Tracy Arm fjord within a minute — the second-highest tsunami run-up ever recorded.

In Los Angeles terms, “the water got as high as the Hollywood sign,” Lynett says. “It scrubbed all of the trees and soil down to bare rock.”

After surveying the site in person, Lynett and his collaborators used computer models to recreate the landslide and the tsunami. Yet in terms of communicating about the event to the public, Lynett felt these state-of-the-art, highly accurate models failed to convey its sheer scale.

“I’ve been studying tsunamis my whole life, and I can’t really imagine what it would have been like to see water moving up a mountainside over 1,500 feet high in the course of a minute,” says Lynett, whose research focuses on building resilience to hazards like tsunamis and hurricanes in coastal communities.

<img decoding="async" class="size-full wp-image-73369" src="https://today.usc.edu/wp-content/uploads/2026/07/Tracy-Arm-Tsunami-Experience-game.jpg" alt="USC Viterbi professor Patrick Lynett is viewing The Tracy Arm Tsunami Experience, an immersive video game he built from computer-model data. ” width=”2400″ height=”1350″ srcset=”https://today.usc.edu/wp-content/uploads/2026/07/Tracy-Arm-Tsunami-Experience-game.jpg 2400w, https://today.usc.edu/wp-content/uploads/2026/07/Tracy-Arm-Tsunami-Experience-game-300×169.jpg 300w, https://today.usc.edu/wp-content/uploads/2026/07/Tracy-Arm-Tsunami-Experience-game-2048×1152.jpg 2048w, https://today.usc.edu/wp-content/uploads/2026/07/Tracy-Arm-Tsunami-Experience-game-320×180.jpg 320w, https://today.usc.edu/wp-content/uploads/2026/07/Tracy-Arm-Tsunami-Experience-game-768×432.jpg 768w, https://today.usc.edu/wp-content/uploads/2026/07/Tracy-Arm-Tsunami-Experience-game-1280×720.jpg 1280w, https://today.usc.edu/wp-content/uploads/2026/07/Tracy-Arm-Tsunami-Experience-game-1920×1080.jpg 1920w” sizes=”(max-width: 2400px) 100vw, 2400px” />

Lynett decided to create a video game based on the computer-model data to immerse people in the event, virtually speaking. The game, called The Tracy Arm Tsunami Experience, allows players to explore the tsunami via four different modes of transport: running or driving an ATV on the shoreline, flying overhead in a helicopter or riding a jet ski in the water. The jet ski view is the most powerful perspective, placing users face-to-face with the gargantuan wall of water moving toward them at more than 100 miles per hour. Spoiler alert: There’s no way to outrun it.

Lynett hopes the game draws attention to the risks that landslide-generated tsunamis pose to human life. During the summer, tourist boats frequent Tracy Arm and nearby fjords to give passengers a closeup view of the glaciers. As the planet warms and glaciers retreat, adjacent mountains are destabilizing, making landslides more frequent in the fjords. Had any boats been within a few miles of last August’s mega-tsunami, there would have been no survivors.

“The idea with the video game is, if we can make the tsunami realistic and immersive, maybe we can better convey the message of caution to people who spend time in these locations,” Lynett says.

Riding into the future

USC researchers’ wave simulations are expanding access to wave encounters in places far from the coast.

Kutch and Culbertson plan to make their VR surfing simulator widely available to a variety of patient populations. “We see VR as part of a broader ecosystem of surfing-based therapies,” Kutch says. “For some people, that may mean ocean surfing. For others, a wave pool or a simulator may be the most practical option. Together, these approaches can make the benefits of surfing accessible to far more people.”

Beyond patient care, the researchers are also exploring recreational applications. “We’re looking at adapting this as a training system for surfers,” Culbertson says. “We’ll adjust not just the waves but also the controllability of the surfboard to be easier or harder based on the individual’s skill level.”

The next iteration of the Kelly Slater Wave Co. technology is the Austin Surf Club in Austin, Texas, where a new 2,200-foot surf basin and luxury clubhouse will become the centerpiece of a condominium community currently under construction. “Think of the surf club like a golf club,” Fincham says.

As these simulations bring realistic waves to landlocked contexts, they’re realizing a human quest begun over a century ago: to replicate this force of nature.

Could a “supernatural” wave that defies nature be next?

“Imagine if you could create a wave that allows an experienced surfer to do things that have never been done in the history of surfing, like a loop the loop in the barrel,” Fincham says. “That’s where we’re headed.”

A previously unknown microprotein hidden within the human mitochondrial genome may help explain certain forms of Type 2 diabetes and could point toward a new, precision-medicine approach to treating it, according to a new USC study.

Obesity and Type 2 diabetes are among the fastest-growing threats to human health, yet their genetic underpinnings remain only partly understood. While most disease-gene research focuses on the larger set of DNA found in the nucleus within cells, the much smaller genome found in mitochondria — cells’ energy factories — is now known to encode a family of microproteins with wide-ranging biological effects.

The new study adds a striking example to that list, said Pinchas Cohen, the study’s senior author, USC Distinguished Professor and dean of the USC Leonard Davis School of Gerontology. The findings were published July 20 in the journal Theranostics.

A genetic clue in a vulnerable population

In a mitochondrial genomewide interaction study using health and genetic data from more than 15,000 adults, the team identified a single-nucleotide polymorphism (SNP, or “snip”) associated with Type 2 diabetes. A SNP is a common type of genetic variation representing a difference in a single nucleotide, or individual “building block” of DNA. SNPs can play a role in individuals’ different responses to certain medications, environmental factors or pathogens.

This genetic variant sits within the gene for a newly identified mitochondrial-derived microprotein, which the researchers have named MENTSH (MDP Encoded in the ND-Two Subunit of Humans).

Notably, this common variant is found most frequently in populations indigenous to the Americas and in 20% of Mexican and Mexican American individuals. The SNP disables MENTSH’s “start codon,” the portion of the gene that would normally signal the cell to start producing the microprotein. The discovery suggests this SNP may be a genetic contributor to metabolic disease in a population that bears a disproportionate burden of diabetes.

From discovery to therapeutic candidate

The researchers confirmed that MENTSH is a genuine, biologically active microprotein using cell-culture experiments and detected it directly using mass spectrometry. They then tested MENTSH and more potent engineered analogues in mouse models of diabetes and obesity.

In these preclinical studies, administering MENTSH improved insulin signaling, while MENTSH analogues potently blocked weight gain in mice fed a high-fat diet. Analyses pointed to a tissue-specific mechanism: MENTSH activates signaling for an enzyme called AKT in muscle but simultaneously reduces AKT signaling in fat, a pattern consistent with improved metabolic health.

“What’s exciting is that this molecule appears to act differently in muscle versus fat, which is exactly the kind of targeted effect you’d want in a metabolic therapy,” said USC Leonard Davis School Research Associate Professor of Gerontology Kelvin Yen, the study’s first author.

Toward precision medicine

Together, the results identify a new biological cause of metabolic dysfunction and suggest that MENTSH-based therapies could one day offer a precision-medicine approach to Type 2 diabetes, particularly for individuals who carry this SNP. The presence of this genetic variant can be easily screened for and could serve as a test for diabetes risk.

The authors emphasize that these findings are preclinical. Further research, including additional safety and efficacy studies, will be required before MENTSH or its analogues can be evaluated in humans.

“For the first time, we’ve connected a mitochondrial microprotein to diabetes risk in a specific population, which opens the door to treatments tailored to the people who need them most,” said co-author Jerome Rotter, a professor at the Lundquist Institute for Biomedical Innovation.

MENTSH could represent an exciting new therapeutic target for diabetes and obesity: a muscle-sparing weight loss peptide, Cohen said.

“This discovery not only represents a potential novel therapeutic for diabetes and obesity, which are major problems around the world, but it also unravels a new cause of diabetes in Hispanics, who are known to be disproportionately affected by these conditions,” he said.

---

About the study: Additional authors included Ricardo Ramirez, Hiroshi Kumagai, Ana Silverstein, Roberto Vicinanza, Melanie Flores, Zeferino Reyna, Su-Jeong Kim, Noel Guerrero, Hemal H. Mehta, Junxiang Wan, Zhongzheng Niu, Carrie V. Breton, Thalida Em Arpawong and Eileen Crimmins of USC; Brendan Miller of the Salk Institute for Biological Studies; Jie Yao, Xiuqing Guo and Kent D. Taylor of the Lundquist Institute; Morgan Levine of Altos Labs; Jihui Sha and James Wohlschlegel of UCLA; and Maria C. Kenney of the University of California, Irvine.

The study was supported by grants from the National Institutes of Health (P30AG068345, R35 GM153408, R01AG068405, R01AG069698 and P30AG094848); the Navigage Foundation; the Ella Fitzgerald Charitable Foundation; the Hanson-Thorell Family Research Award; the Hinrich Foundation; the Hevolution Foundation (HF-AGE-23-1273964-51); and the Hinrich Endowment for Mitochondrial Genetics at the USC Leonard Davis School.

Across USC’s schools and campuses, researchers, faculty and practitioners are developing innovative solutions for societal issues, imagining new approaches to making art and finding ways for creators to prepare for a shifting landscape as AI increasingly intersects with different fields.

Under USC President Beong-Soo Kim’s leadership, advancing ethical AI that puts people first has become a university priority — building on USC’s legacy of innovation while preparing the next generation of researchers, scientists, engineers, entrepreneurs, artists and health professionals. 

This story is part of Trojan Family Magazine’s AI series highlighting the people and ideas driving USC’s vision forward. 

AI is helping USC researchers answer a question with life-or-death consequences: Why do medications work well for some patients but not others?

Gauri Rao, associate professor of clinical pharmacy at the USC Alfred E. Mann School of Pharmacy and Pharmaceutical Sciences, is the director of the Center for Quantitative Drug and Disease Modeling. At her lab, Rao and her team are using machine learning to find the answer, analyzing vast amounts of bacterial genomic and metabolomic data to better understand the mechanisms that drive antimicrobial resistance.

By identifying patterns in bacterial genomes and metabolic pathways, the team aims to predict how bacteria develop resistance to antibiotics and uncover strategies to preserve the effectiveness of existing therapies. Their work combines advanced computational approaches with experimental research to address one of the world’s most pressing public health challenges.

“Developing a new antibiotic is incredibly costly and can take more than a decade,” Rao says. “Given the urgent rise in antimicrobial resistance, we are focused on using machine learning and quantitative modeling to maximize the effectiveness of the antibiotics we already have and help slow the emergence of resistance.”

Given the growing threat of drug-resistant infections, a large portion of Rao’s research program is dedicated to bacteriophage (phage) therapy, an emerging approach that uses naturally occurring viruses to target and kill bacteria. Rao says machine learning is critical for helping researchers rapidly identify phages that are most likely to be effective against specific bacterial pathogens.

Rao is part of the National Institutes of Health’s first coordinated U.S. research network for phage therapeutics. Through this collaborative effort, computational approaches, including machine learning, are integrated with laboratory (in vitro) and animal model (in vivo) testing to identify promising phage candidates and accelerate their translation into clinical practice. The goal is to develop more precise and effective treatments for multidrug-resistant infections while preserving the effectiveness of current antibiotics.

Prashant Dogra, associate director of the center and assistant professor of clinical pharmacy at USC Mann, also integrates mechanistic mathematical modeling with AI to better understand disease mechanisms and drug delivery systems.

One of Dogra’s current projects, supported by funding from the NIH, studies how nanomedicines move through the body to design safer and more effective drug delivery systems to transform treatment for diseases ranging from cancer to chronic illnesses.

Using machine learning, he analyzes a large body of published studies to identify design principles for safer and more effective nanoparticles, potentially leading to better clinical outcomes for patients.

Dogra, who joined the USC Mann faculty last year, says he can’t imagine doing the projects he’s currently working on at any other university.

“I think USC has a very interdisciplinary environment that I find is really unmatchable,” Dogra says. “Having access to colleagues who are very strong in complementary domains facilitates newer ways of thinking and allows us to develop these novel methods of pursuing our scientific questions.”

---

Read the entire series: USC at the Leading Edge of AI

Part 1: The Business Students | Part 2: The Pharmaceutical Developers | Part 3: The Creatives | Part 4: The Filmmaker | Part 5: The Urban Planners