A paper describing the research appeared today March 17 in the Astronomical Journal. The findings offer clues about the properties of sub-Neptunes and help address long-standing questions about their origins, the team said.

“The majority of the 5,500 or so exoplanets discovered to date have a very close orbit to their stars, closer than Mercury to our sun, which we call ‘close-in’ planets,” said Rachel Fernandes, President’s Postdoctoral Fellow in the Department of Astronomy and Astrophysics at Penn State and leader of the research team. “Many of these are gaseous sub-Neptunes, a type of planet absent from our own solar system. While our gas giants, like Jupiter and Saturn, formed farther from the sun, it’s unclear how so many close-in sub-Neptunes managed to survive near their stars, where they are bombarded by intense stellar radiation.”

To better understand how sub-Neptunes form and evolve, the researchers turned to planets around young stars, which only recently became observable thanks to TESS.

“Comparing the frequency of exoplanets of certain sizes around stars of different ages can tell us a lot about the processes that shape planet formation,” Fernandes said. “If planets commonly form at specific sizes and locations, we should see a similar frequency of those sizes across different ages. If we don’t, it suggests that certain processes are changing these planets over time.”

Observing planets around young stars, however, has traditionally been difficult. Young stars emit bursts of intense radiation, rotate quickly and are highly active, creating high levels of “noise” that make it challenging to observe planets around them.

“Young stars in their first billion years of life throw tantrums, emitting a ton of radiation,” Fernandes explained. “These stellar tantrums cause a lot of noise in the data, so we spent the last six years developing a computational tool called Pterodactyls to see through that noise and actually detect young planets in TESS data.”

The research team used Pterodactyls to evaluate TESS data and identify planets with orbital periods of 12 days or less — for reference, much less than Mercury’s 88-day orbit — with the goal of examining the planet sizes, as well as how the planets were shaped by the radiation from their host stars. Because the team’s survey window was 27 days, this allowed them to see two full orbits from potential planets. They focused on planets between a radius of 1.8 and 10 times the size of Earth, allowing the team to see if the frequency of sub-Neptunes is similar or different in young systems versus older systems previously observed with TESS and NASA’s retired Kepler Space Telescope.

The researchers found that the frequency of close-in sub-Neptunes changes over time, with fewer sub-Neptunes around stars between 10 and 100 million years of age compared to those between 100 million and 1 billion years of age. However, the frequency of close-in sub-Neptunes is much less in older, more stable systems.

“We believe a variety of processes are shaping the patterns we see in close-in stars of this size,” Fernandes said. “It’s possible that many sub-Neptunes originally formed further away from their stars and slowly migrated inward over time, so we see more of them at this orbital period in the intermediate age. In later years, it’s possible that planets are more commonly shrinking when radiation from the star essentially blows away its atmosphere, a process called atmospheric mass loss that could explain the lower frequency of sub-Neptunes. But it’s likely a combination of cosmic processes shaping these patterns over time rather than one dominant force.”

The researchers said they would like to expand their observation window with TESS to observe planets with longer orbital periods. Future missions like the European Space Agency’s PLATO may also allow the research team to observe planets of smaller sizes, similar to that of Mercury, Venus, Earth and Mars. Expanding their analysis to smaller and more distant planets could help the researchers refine their tool and provide additional information about how and where planets form.

Additionally, NASA’s James Webb Space Telescope could permit the characterization of the density and composition of individual planets, which Fernandes said could give additional hints to where they formed.

“Combining studies of individual planets with the population studies like we conducted here would give us a much better picture of planet formation around young stars,” Fernandes said. “The more solar systems and planets we discover, the more we realize that our solar system isn’t really the template; it’s an exception. Future missions might enable us to find smaller planets around young stars and give us a better picture of how planetary systems form and evolve with time, helping us better understand how our solar system, as we know it today, came to be.”

In addition to Fernandes, the research team at Penn State includes Rebekah Dawson, Shaffer Career Development Professor in Science and professor of astronomy and astrophysics at the time of the research and now a physical scientist at NASA. The research team also includes Galen J. Bergsten, Ilaria Pascucci, Kevin K. Hardegree-Ullman, Tommi T. Koskinen and Katia Cunha at the University of Arizona; Gijs Mulders at Pontifical Catholic University of Chile; Steven Giacalone, Eric Mamajek, Kyle Pearson, David Ciardi, Preethi Karpoor, Jessie Christiansen and Jon Zink at the California Institute of Technology; James Rogers at the University of Cambridge, Los Angeles; Akash Gupta at Princeton University; Kiersten Boley at the Carnegie Institution for Science; Jason Curtis at Columbia University; Sabina Sagynbayeva at Stony Brook University; Sakhee Bhure at the University of Southern Queensland in Australia; and Gregory Feiden at the University of North Georgia.

Funding from NASA, including through support of the “Alien Earths” grant; Chile’s National Fund for Scientific and Technological Development; and the U.S. National Science Foundation supported this research. Additional support was provided by the Penn State Center for Exoplanets and Habitable Worlds and the Penn State Extraterrestrial Intelligence Center. Computations for this research were performed with Penn State’s University’s Institute for Computational and Data Sciences’ Roar supercomputer.

The model, called patch foundational transformer for sleep (PFTSleep), analyzes brain waves, muscle activity, heart rate, and breathing patterns to classify sleep stages more effectively than traditional methods, streamlining sleep analysis, reducing variability, and supporting future clinical tools to detect sleep disorders and other health risks.

Current sleep analysis often relies on human experts manually scoring short segments of sleep data or using AI models that are not capable of analyzing a patient’s entire night of sleep. This new approach, developed using thousands of sleep recordings, takes a more comprehensive view. By training on full-length sleep data, the model can recognize sleep patterns throughout the night and across different populations and settings, offering a standardized and scalable method for sleep research and clinical use, say the investigators.

“This is a step forward in AI-assisted sleep analysis and interpretation,” says first author Benjamin Fox, a PhD candidate at the Icahn School of Medicine at Mount Sinai in the Artificial Intelligence and Emerging Technologies Training Area. “By leveraging AI in this way, we can learn relevant clinical features directly from sleep study signal data and use them for sleep scoring and, in the future, other clinical applications such as detecting sleep apnea or assessing health risks linked to sleep quality.”

The model was built using a large dataset of sleep studies (polysomnograms) that measure key physiological signals, including brain activity, muscle tone, heart rate, and breathing patterns. Unlike traditional AI models, which analyze only short, 30-second segments, this new model considers the entire night of sleep, capturing more detailed and nuanced patterns. Further, the model is trained via a method known as self-supervision, which helps learn relevant clinical features from physiological signals without using human labeled outcomes.

“Our findings suggest that AI could transform how we study and understand sleep,” says co-senior corresponding author Ankit Parekh, PhD, Assistant Professor of Medicine (Pulmonary, Critical Care and Sleep Medicine) at the Icahn School of Medicine at Mount Sinai, and Director of the Sleep and Circadian Analysis Group at Mount Sinai. “Our next goal is to refine the technology for clinical applications, such as identifying sleep-related health risks more efficiently.”

The researchers emphasize that this AI tool, while promising, would not replace clinical expertise. Instead, it would serve as a powerful aid for sleep specialists, helping to speed up and standardize sleep analysis. Next, the team’s research aims to expand its capabilities beyond sleep-stage classification to detecting sleep disorders and predicting health outcomes.

“This AI-driven approach has the potential to revolutionize sleep research,” says co-senior corresponding author Girish N. Nadkarni, MD, MPH, Chair of the Windreich Department of Artificial Intelligence and Human Health at the Icahn School of Medicine, Director of the Hasso Plattner Institute for Digital Health, and the Irene and Dr. Arthur M. Fishberg Professor of Medicine. Dr. Nadkarni is also the inaugural Chief of the Division of Data-Driven and Digital Medicine and Co-Director of the Mount Sinai Clinical Intelligence Center. “By analyzing entire nights of sleep with greater consistency, we can uncover deeper insights into sleep health and its connection to overall well-being.”

The study, “Endothelial TDP-43 Depletion Disrupts Core Blood-Brain Barrier Pathways in Neurodegeneration,” was published on March 14, 2025. The lead author, Omar Moustafa Fathy, an MD/Ph.D. candidate at the Center for Vascular Biology at UConn School of Medicine, conducted the research in the laboratory of senior author Dr. Patrick A. Murphy, associate professor and newly appointed interim director of the Center for Vascular Biology. The study was carried out in collaboration with Dr. Riqiang Yan, a leading expert in Alzheimer’s disease and neurodegeneration research.

This work provides a novel and significant exploration of how vascular dysfunction contributes to neurodegenerative diseases, exemplifying the powerful collaboration between the Center for Vascular Biology and the Department of Neuroscience. While clinical evidence has long suggested that blood-brain barrier (BBB) dysfunction plays a role in neurodegeneration, the specific contribution of endothelial cells remained unclear. The BBB serves as a critical protective barrier, shielding the brain from circulating factors that could cause inflammation and dysfunction. Though multiple cell types contribute to its function, endothelial cells — the inner lining of blood vessels — are its principal component.

“It is often said in the field that ‘we are only as old as our arteries’. Across diseases we are learning the importance of the endothelium. I had no doubt the same would be true in neurodegeneration, but seeing what these cells were doing was a critical first step,” says Murphy.

Omar, Murphy, and their team tackled a key challenge: endothelial cells are rare and difficult to isolate from tissues, making it even harder to analyze the molecular pathways involved in neurodegeneration.

To overcome this, they developed an innovative approach to enrich these cells from frozen tissues stored in a large NIH-sponsored biobank. They then applied inCITE-seq, a cutting-edge method that enables direct measurement of protein-level signaling responses in single cells — marking its first-ever use in human tissues.

This breakthrough led to a striking discovery: endothelial cells from three different neurodegenerative diseases — Alzheimer’s disease (AD), amyotrophic lateral sclerosis (ALS), and frontotemporal dementia (FTD) — shared fundamental similarities that set them apart from the endothelium in healthy aging. A key finding was the depletion of TDP-43, an RNA-binding protein genetically linked to ALS-FTD and commonly disrupted in AD. Until now, research has focused primarily on neurons, but this study highlights a previously unrecognized dysfunction in endothelial cells.

“It’s easy to think of blood vessels as passive pipelines, but our findings challenge that view,” says Omar. “Across multiple neurodegenerative diseases, we see strikingly similar vascular changes, suggesting that the vasculature isn’t just collateral damage — it’s actively shaping disease progression. Recognizing these commonalities opens the door to new therapeutic possibilities that target the vasculature itself.”

The research team believes this newly identified subset of endothelial cells could provide a roadmap to targeting this endothelial disfunction to stave off disease, and also to develop new biomarkers from the blood of patients with disease.

Funding was provided by startup funds from the UConn School of Medicine and Department of Cell Biology, Center for Vascular Biology and Calhoun Cardiology Center, American Heart Association Innovative Project Award 19IPLOI34770151 (to P.A.M.); NIH National Heart, Lung, and Blood Institute Grants K99/R00-HL125727 and RF1-NS117449 (to P.A.M); American Heart Association Predoctoral award 23PRE1027078 (to O.M.F.O.) R01-AG046929 and R01-NS074256 (to R.Y.) and NIH GM135592 (to B.H.).