IntroductionMultiple sclerosis (MS) affects different cognitive domains, including social cognition. Immersive Virtual Reality (VR) may provide a novel rehabilitative approach to treat motor and cognitive symptoms of MS. This exploratory pilot study evaluated the effects of immersive VR rehabilitation on social cognition in MS patients and explored related cortical neurophysiological signatures.MethodsSeven MS patients underwent immersive VR rehabilitation with the CAREN system (3 sessions/week, approximately 45 min of active training per session, about 1 h including preparation, 8 weeks), while seven healthy controls (HC) did not undergo any intervention. Patients were evaluated at baseline (T0) and post-treatment (T1) with standardized measures of cognitive, emotional, and motor functioning. EEG data were acquired from all participants, and, after artifact removal, spectral parameterization decomposed signals into aperiodic (exponent, offset) and periodic oscillatory components (alpha and beta power). Power spectral density was analyzed using group comparisons and Pearson correlations with neuropsychological measures.ResultsCompared with HC, MS patients showed reduced alpha-band power, mainly over frontal and parieto-occipital regions, whereas aperiodic parameters did not differ between groups. In patients, alpha and beta power correlated with the Positive Emotions Self-Efficacy Scale (alpha: r = 0.92, p = 0.003; beta: r = 0.83, p = 0.020). Alpha power is also correlated with RAO SRT–LTS (r = 0.85, p = 0.016), and beta with EQ-CE (r = 0.82, p = 0.023). Overall, alpha and beta power were correlated with emotional self-efficacy, balance, memory, and empathy, suggesting that oscillatory markers are potential indicators of clinical outcomes.DiscussionRehabilitation via immersive VR has shown promising clinically significant effects in the cognitive, emotional, and motor domains, supported by convergent EEG spectral signatures. Future studies employing predictive modeling approaches will be required to assess their prognostic value.
The problem with thinking you’re part Neanderthal
You’ve probably heard some version of this idea before: that many of us have an “inner Neanderthal.” That is to say, around 45,000 years ago, when Homo sapiens first arrived in Europe, they met members of a cousin species—the broad-browed, heavier-set Neanderthals—and, well, one thing led to another, which is why some people now carry a small amount of Neanderthal DNA.
This DNA is arguably the 21st century’s most celebrated discovery in human evolution. It has been connected to all kinds of traits and health conditions, and it helped win the Swedish geneticist Svante Pääbo a Nobel Prize.
But in 2024, a pair of French population geneticists called into question the foundation of the popular and pervasive theory.
Lounès Chikhi and Rémi Tournebize, then colleagues at the Université de Toulouse, proposed an alternative explanation for the very same genomic patterns. The problem, they said, was that the original evidence for the inner Neanderthal was based on a statistical assumption: that humans, Neanderthals, and their ancestors all mated randomly in huge, continent-size populations. That meant a person in South Africa was just as likely to reproduce with a person in West Africa or East Africa as with someone from their own community.
Archaeological, genetic, and fossil evidence all shows, though, that Homo sapiens evolved in Africa in smaller groups, cut off from one another by deserts, mountains, and cultural divides. People sometimes crossed those barriers, but more often they partnered up within them.
In the terminology of the field, this dynamic is called population structure. Because of structure, genes do not spread evenly through a population but can concentrate in some places and be totally absent from others. The human gene pool is not so much an Olympic-size swimming pool as a complex network of tidal pools whose connectivity ebbs and flows over time.
This dynamic greatly complicates the math at the heart of evolutionary biology, which long relied on assumptions like randomly mating populations to extract general principles from limited data. If you take structure into account, Chikhi told me recently, then there are other ways to explain the DNA that some living people share with Neanderthals—ways that don’t require any interspecies sex at all.
“I believe most species are spatially organized and structured in different, complex ways,” says Chikhi, who has researched population structure for more than two decades and has also studied lemurs, orangutans, and island birds. “It’s a general failure of our field that we do not compare our results in a clear way with alternative scenarios.” (Pääbo did not respond to multiple requests for comment.)
The inner Neanderthal became a story we could tell ourselves about our flaws and genetic destiny: Don’t blame me; blame the prognathic caveman hiding in my cells.
Chikhi and Tournebize’s argument is about population structure, yes, but at heart, it is actually one about methods—how modern evolutionary science deploys computer models and statistical techniques to make sense of mountains upon mountains of genetic data.
They’re not the only scientists who are worried. “People think we really understand how genomes evolve and can write sophisticated algorithms for saying what happened,” says William Amos, a University of Cambridge population geneticist who has been critical of the “inner Neanderthal” theory. But, he adds, those models are “based on simple assumptions that are often wrong.”
And if they’re wrong, what’s at stake is far more than a single evolutionary mystery.
A captivating story of interspecies passion
Back in 2010, Pääbo’s lab pulled off something of a miracle. The researchers were able to extract DNA from nuclei in the cells of 40,000-year-old Neanderthal bones. DNA breaks down quickly after death, but the group got enough of it from three different individuals to produce a draft sequence of the entire Neanderthal genome, with 4 billion base pairs.
As part of their study, they performed a statistical test comparing their Neanderthal genome with the genomes of five present-day people from different parts of the world. That’s how they discovered that modern humans of non-African ancestry had a small amount of DNA in common with Neanderthals, a species that diverged from the Homo sapiens line more than 400,000 years ago, that they did not share with either modern humans of African ancestry or our closest living relative, the chimpanzee.
WELLCOME COLLECTION
Pääbo’s team interpreted this as evidence of sexual reproduction between ancient Homo sapiens and the Neanderthals they encountered after they expanded out of Africa. “Neanderthals are not totally extinct,” Pääbo said to the BBC in 2010. “In some of us, they live on a little bit.”
The discovery was monumental on its own—but even more so because it reversed a previous consensus. More than a decade earlier, in 1997, Pääbo had sequenced a much smaller amount of Neanderthal DNA, in that case from a cell structure called a mitochondrion. It was different enough from Homo sapiens mitochondrial DNA for his team to cautiously conclude there had been “little or no interbreeding” between the two species.
After 2010, though, the idea of hybridization, also called admixture, effectively became canon. Top journals like Science and Nature published study after study on the inner Neanderthal. Some scientists have argued that Homo sapiens would never have adapted to colder habitats in Europe and Asia without an infusion of Neanderthal DNA. Other research teams used Pääbo’s techniques to find genetic traces of interbreeding with an extinct group of hominins in Asia, called the Denisovans, and a mysterious “ghost lineage” in Africa. Biologists used similar tests to find evidence of interbreeding between chimpanzees and bonobos, polar and brown bears, and all kinds of other animals.
The inner-Neanderthal hypothesis also took a turn for the personal. Various studies linked Neanderthal DNA to a head-spinning range of conditions: alcoholism, asthma, autism, ADHD, depression, diabetes, heart disease, skin cancer, and severe covid-19. Some researchers suggested that Neanderthal DNA had an impact on hair and skin color, while others assigned individuals a “NeanderScore” that was correlated with skull shape and prevalence of schizophrenia markers. Commercial genetic testing companies like 23andMe started offering customers Neanderthal ancestry reports.
The inner Neanderthal became a story we could tell ourselves about our flaws and genetic destiny: Don’t blame me; blame the prognathic caveman hiding in my cells. Or as Latif Nasser, a host of the popular-science program Radiolab, put it when he was hospitalized with Crohn’s disease, another Neanderthal-associated condition: “I just keep imagining these tiny Neanderthals … just, like, stabbing me and drawing these little droplets of blood out of me.”
“These things become meaningful to people,” Chikhi says. “What we say will be important to how people view themselves.”
The pitfalls of simplistic solutions
When population geneticists built the theoretical framework for evolutionary biology in the early 20th century, genes were only abstract units of heredity inferred from experiments with peas and fruit flies. Population genetics developed theory far more quickly than it accumulated data. As a result, many data-driven scientists dismissed the study of evolution as a form of storytelling based on unexamined assumptions and preconceived ideas.
By the ’90s, though, genes were no longer abstractions but sequenced segments of DNA. Genomic sequencing grounded evolutionary studies in the kind of hard data that a chemist or physicist could respect.
Yet biologists could not simply read evolutionary history from genomes as though they were books. They were trying to determine which of a nearly infinite number of plausible histories was the most likely to have created the patterns they observed in a small sample of genomes. For that, they needed simplified, algorithmic models of evolution. The study of evolution shifted from storytelling to statistics, and from biology to computer science.
That suited Chikhi, who as a child was drawn to the predictable laws and numerical precision of math and science. He entered the field in the mid-’90s just as the first big studies of human DNA were settling old debates about human origins. DNA showed that Africa harbored far more genetic diversity than the entire rest of the planet. The new evidence supported the idea that modern humans evolved for hundreds of thousands of years in Africa and expanded to the other continents only in the last 100,000 years. For Chikhi, whose parents were Algerian immigrants, this discovery was a powerful challenge to the way some archaeologists and biologists talked about race. DNA could be used to deconstruct rather than encourage the pernicious idea that human races had deep-seated evolutionary differences based on their places of origin.
At the same time, though, he was wary of the tendency to treat DNA as the final verdict on open questions in evolution. Chikhi had been surprised when, back in 1997, Pääbo and his team used that small amount of mitochondrial DNA to rule out hybridization between Homo sapiens and Neanderthals. He didn’t think that the absence of Neanderthal DNA there necessarily meant it wouldn’t be found elsewhere in the Homo sapiens genome.
Chikhi’s own research in the aughts opened his eyes to the gaps between historical reality and models of evolution. For one, despite the assumption of random mating, none of the animals Chikhi studied actually mated randomly. Orangutans lived in highly fragmented habitats, which restricted their pool of potential mates, and female birds were often extremely picky about their male partners.
These factors could confound an evolutionary biologist’s traditional statistical tool kit. Scientists were starting to apply a mathematical technique to estimate historical population sizes for a species from the genome of just a single individual. This method showed sharp population declines in the histories of many different species. Chikhi realized, though, that the apparent declines could be an artifact of treating a structured population as one that evolved with random mating; in that case, the technique could indicate a bottleneck even if all the subgroups were actually growing in size. “This is completely counterintuitive,” he says.
That’s at least partly why, when Pääbo’s 2010 Neanderthal genome came out, Chikhi was impressed with the sheer technical accomplishment but also leery of the findings about hybridization. “It was the type of thing we conclude too quickly based on genetic data,” he says. Pääbo’s work mentioned population structure as a possible alternative explanation—but didn’t follow up.
Just a couple of years later, a pair of independent scientists named Anders Eriksson and Andrea Manica picked up the idea, building a model with simple population structure that explicitly excluded admixture. They simulated human evolution starting from 500,000 years ago and found that their model produced the same genomic patterns Pääbo’s group had interpreted as evidence of hybridization.
“Working with structured models is really out of the comfort zone of a lot of population geneticists,” says Eriksson, now a professor at the University of Tartu in Estonia.
Their research impressed Chikhi. “At the time, I thought people would focus on population structure in the evolution of humans,” he says. Instead, he watched as the inner-Neanderthal hypothesis took on a life of its own. Scientists produced new methods to quantify hybridization but rarely examined whether population structure would yield the same results. To Chikhi, this wasn’t science; it was storytelling, like some of the old narratives about the evolution of racial differences.
Chikhi and Tournebize decided to take a crack at the problem themselves. “I’ve always been very skeptical about science, and population genetics in particular,” says Tournebize, now a researcher at the French National Research Institute for Sustainable Development. “We make a lot of assumptions, and the models we use are very simplistic.” As detailed in a 2024 paper published in Nature Ecology & Evolution, they built a model of human evolution that replaced randomly mating continent-wide populations with many smaller populations linked by occasional migration. Then they let it run—a million times.
At the end of the simulation, they kept the 20 scenarios that produced genomes most similar to the ones in a sample of actual Homo sapiens and Neanderthals. Many of these scenarios produced long segments of DNA like the ones their peers argued could only have been inherited from Neanderthals. They showed that several statistics, which other scientists had proposed as measurements of Neanderthal DNA, couldn’t actually distinguish between hybridization and population structure. What’s more, they showed that many of the models that supported hybridization failed to accurately predict other known features of human evolution.
“A model will say there was admixture but then predict diversity that is totally incompatible with what we actually know of human diversity,” Chikhi says. “Nobody seems to care.”
So how did Neanderthal DNA wind up in living people if not via interspecies passion? Chikhi and Tournebize think it’s more likely that it was inherited by both Neanderthals and some sapiens groups in Africa from a common ancestor living at least half a million years ago. If the sapiens groups carrying those genetic variants included the people who migrated out of Africa, then the two human species would have already had the DNA in common when they came into contact in Europe and Asia—no sex required.
“The interpretation of genetic data is not straightforward,” Chikhi says. “We always have to make assumptions. Nobody takes data and magically comes up with a solution.”
Embracing the uncertainty
Most of the half-dozen population geneticists I spoke with praised Chikhi and Tournebize’s ingenuity and appreciated the spirit of their critique. “Their paper forces us to think more critically about the model we use for inference and consider alternatives,” says Aaron Ragsdale, a population geneticist at the University of Wisconsin–Madison. His own work likewise suggests that the earliest Homo sapiens populations in Africa were probably structured—and that this is the likely reason for genomic patterns that other research groups had attributed to hybridization with a mysterious “ghost lineage” of hominins in Africa.
Yet most researchers still believe that modern humans and Neanderthals did probably have children with each other tens of thousands of years ago. Several pointed to the fact that fossil DNA of Homo sapiens who died thousands of years ago had longer chunks of apparent Neanderthal DNA than living people, which is exactly what you would expect if they had a more recent Neanderthal ancestor. (To address this possibility, Chikhi and Tournebize included DNA from 10 ancient humans in their study and found that most of them fit the structured model.) And while the Harvard population geneticist David Reich, who helped design the statistical test from Pääbo’s 2010 study, declined an interview, he did say he thought Chikhi and Tournebize’s model was “weak” and “very contrived,” adding that “there are multiple lines of evidence for Neanderthal admixture into modern humans that make the evidence for this overwhelming.” (Two other authors of that study, Richard Green and Nick Patterson, did not respond to requests for comment.)
Nevertheless, most scientists these days welcome the development of structured, or “spatially explicit,” models that account for the fact that any given member of a population is usually more closely related to individuals living nearby than to those living far away.
Loosening our attachment to certain narratives of evolution can create space for wonder at the sheer complexity of life’s history.
Other scientists also say that random mating isn’t the only assumption in population genetics that merits scrutiny. Models rarely factor in natural selection, which can also create genetic patterns that look like hybridization. Another common assumption is that everyone’s DNA mutates at the same, constant rate. “All the theory says the mutation rate is fixed,” says Amos, the Cambridge population geneticist. But he thinks that rate would have slowed drastically in the group of Homo sapiens that expanded to Europe around 45,000 years ago. This, too, could have created genomic patterns that other scientists interpret as evidence of interbreeding with Neanderthals.
COURTESY OF 23ANDME
The point here isn’t that a complex model of evolution with many moving pieces is necessarily better than a simple one. Scientists need to reduce complexity in order to see the underlying processes more clearly. But simple models require assumptions, and scientists need to reevaluate those assumptions in light of what they learn. “As you get more data, you can justify more complex models of the world,” says Mark Thomas, a population geneticist at University College London, who wrote a history of random mating in population genetics that highlighted how the field was starting to see it as “a limiting assumption as opposed to a simplifying one.”
It can feel discouraging to couch conversations about the past in confusing terms like “population structure” and “mutation rates.” It seems almost antithetical to the spirit of science to talk more about uncertainty at the same time we are developing powerful technologies and enormous data sets for analyzing evolution. These tools often yield novel answers, but they can also limit the questions we ask. The French archaeologist Ludovic Slimak, for example, has complained that the idea of the inner Neanderthal has domesticated our image of Neanderthals and made it difficult to imagine their humanity as distinct from our own. Investigating Neanderthal DNA is sexier to many young researchers than searching for archaeological and fossil evidence of how Neanderthals actually lived.
Loosening our attachment to certain narratives of evolution can create space for wonder at the sheer complexity of life’s history. Ultimately, that’s what Chikhi and Tournebize hope to do. After all, they don’t believe the question of population structure versus hybridization is either-or. It’s possible, and even likely, that both played a role in human evolution. “Our structured model does not necessarily mean that no admixture ever took place,” Chikhi and Tournebize wrote in their study. “What our results suggest is that, if admixture ever occurred, it is currently hard to identify using existing methods.”
Future methods might disentangle the different factors, but it’s just as important, Chikhi says, for scientists to be up-front about their assumptions and test alternatives. “There’s still so much uncertainty on so many aspects of the demographic history of Neanderthals and Homo sapiens,” he notes.
Keep that in mind the next time you read about your inner Neanderthal. The association between this DNA and some diseases may be real, of course—but would journals publish these studies without the additional claim that the DNA is from Neanderthals? Any good storyteller knows that sex sells, even in science.
Ben Crair is a science and travel writer based in Berlin.
Kent and Medway mental health appointments launch online
Kent and Medway Mental Health NHS Trust has a launched a portal to give patients 24/7 online access to their mental health appointments.
Want to understand the current state of AI? Check out these charts.
If you’re following AI news, you’re probably getting whiplash. AI is a gold rush. AI is a bubble. AI is taking your job. AI can’t even read a clock. The 2026 AI Index from Stanford University’s Institute for Human-Centered Artificial Intelligence, AI’s annual report card, comes out today and cuts through some of that noise.
Despite predictions that AI development may hit a wall, the report says that the top models just keep getting better. People are adopting AI faster than they picked up the personal computer or the internet. AI companies are generating revenue faster than companies in any previous technology boom, but they’re also spending hundreds of billions of dollars on data centers and chips. The benchmarks designed to measure AI, the policies meant to govern it, and the job market are struggling to keep up. AI is sprinting, and the rest of us are trying to find our shoes.
All that speed comes at a cost. AI data centers around the world can now draw 29.6 gigawatts of power, enough to run the entire state of New York at peak demand. Annual water use from running OpenAI’s GPT-4o alone may exceed the drinking water needs of 12 million people. At the same time, the supply chain for chips is alarmingly fragile. The US hosts most of the world’s AI data centers, and one company in Taiwan, TSMC, fabricates almost every leading AI chip.
The data reveals a technology evolving faster than we can manage. Here’s a look at some of the key points from this year’s report.
The US and China are nearly tied
In a long, heated race with immense geopolitical stakes, the US and China are almost neck and neck on AI model performance, according to Arena, a community-driven ranking platform that allows users to compare the outputs of large language models on identical prompts. In early 2023, OpenAI had a lead with ChatGPT, but this gap narrowed in 2024 as Google and Anthropic released their own models. In February 2025, R1, an AI model built by the Chinese lab DeepSeek, briefly matched the top US model, ChatGPT. As of March 2026, Anthropic leads, trailed closely by xAI, Google, and OpenAI. Chinese models like DeepSeek and Alibaba lag only modestly. With the best AI models separated in the rankings by razor-thin margins, they’re now competing on cost, reliability, and real-world usefulness.
The index notes that the US and China have different AI advantages. While the US has more powerful AI models, more capital, and an estimated 5,427 data centers (more than 10 times as many as any other country), China leads in AI research publications, patents, and robotics.
As competition intensifies, companies like OpenAI, Anthropic, and Google no longer disclose their training code, parameter counts, or data-set sizes. “We don’t know a lot of things about predicting model behaviors,” says Yolanda Gil, a computer scientist at the University of Southern California who coauthored the report. This lack of transparency makes it difficult for independent researchers to study how to make AI models safer, she says.
AI models are advancing super fast
Despite predictions that development will plateau, AI models keep getting better and better. By some measures, they now meet or exceed the performance of human experts on tests that aim to measure PhD-level science, math, and language understanding. SWE-bench Verified, a software engineering benchmark for AI models, saw top scores jump from around 60% in 2024 to almost 100% in 2025. In 2025, an AI system produced a weather forecast on its own.
“I am stunned that this technology continues to improve, and it’s just not plateauing in any way,” says Gil.
However, AI still struggles in plenty of other areas. Because the models learn by processing enormous amounts of text and images rather than by experiencing the physical world, AI exhibits “jagged intelligence.” Robots are still in their early days and succeed in only 12% of household tasks. Self-driving cars are farther along: Waymos are now roaming across five US cities, and Baidu’s Apollo Go vehicles are shuttling riders around in China. AI is also expanding into professional domains like law and finance, but no model dominates the field yet.
But the way we test AI is broken
These reports of progress should be taken with a grain of salt. The benchmarks designed to track AI progress are struggling to keep up as models quickly blow past their ceilings, the Stanford report says. Some are poorly constructed—a popular benchmark that tests a model’s math abilities has a 42% error rate. Others can be gamed: when models are trained on benchmark test data, for example, they can learn to score well without getting smarter.
Because AI is rarely used the same way it’s tested, strong benchmark performance doesn’t always translate to real-world performance. And for complex, interactive technologies such as AI agents and robots, benchmarks barely exist yet.
AI companies are also sharing less about how their models are trained, and independent testing sometimes tells a different story from what they report. “A lot of companies are not releasing how their models do in certain benchmarks, particularly the responsible-AI benchmarks,” says Gil. “The absence of how your model is doing on a benchmark maybe says something.”
AI is starting to affect jobs
Within three years of going mainstream, AI is now used by more than half of people around the world, a rate of adoption faster than the personal computer or the internet. An estimated 88% of organizations now use AI, and four in five university students use it.
It’s early days for deployment, and AI’s impact on jobs is hard to measure. Still, some studies suggest AI is beginning to affect young workers in certain professions. According to a 2025 study by economists at Stanford, employment for software developers aged 22 to 25 has fallen nearly 20% since 2022. The decline might not be pinned on AI alone, as broader macroeconomic conditions could be to blame, but AI appears to be playing a part.
Employers say that hiring may continue to tighten. According to a 2025 survey conducted by McKinsey & Company, a third of organizations expect AI to shrink their workforce in the coming year, particularly in service and supply chain operations and software engineering. AI is boosting productivity by 14% in customer service and 26% in software development, according to research cited by the index, but such gains are not seen in tasks requiring more judgment. Overall, it’s still too early to understand the bigger economic impact of AI.
People have complicated feelings about AI
Around the world, people feel both optimistic and anxious about AI: 59% of people think that it will provide more benefits than drawbacks, while 52% say that it makes them nervous, according to an Ipsos survey cited in the index.
Notably, experts and the public see the future of AI very differently, according to a Pew survey. The biggest gap is around the future of work: While 73% of experts think that AI will have a positive impact on how people do their jobs, only 23% of the American public thinks so. Experts are also more optimistic than the public about AI’s impact on education and medical care, but they agree that AI will hurt elections and personal relationships.
Among all countries surveyed, Americans trust their government least to regulate AI appropriately, according to another Ipsos survey. More Americans worry federal AI regulation won’t go far enough than worry it will go too far.
Governments are struggling to regulate AI
Governments around the world are struggling to regulate AI, but there were some minor successes last year. The EU AI Act’s first prohibitions, which ban the use of AI in predictive policing and emotion recognition, took effect. Japan, South Korea, and Italy also passed national AI laws. Meanwhile, the US federal government moved toward deregulation, with President Trump issuing an executive order seeking to handcuff states from regulating AI.
Despite this federal action, state legislatures in the US passed a record 150 AI-related bills. California enacted landmark legislation, including SB 53, which mandates safety disclosures and whistleblower protections for developers of AI models. New York passed the RAISE Act, requiring AI companies to publish safety protocols and report critical safety incidents.
But for all the legislative activity, Gil says, regulation is running behind the technology because we don’t really understand how it works. “Governments are cautious to regulate AI because … we don’t understand many things very well,” she says. “We don’t have a good handle on those systems.”
Opinion: I’m a MAHA activist. I went into the public health lion’s den — and it changed how I think
The past few weeks have been nothing but discouraging for those of us who helped create the Make America Healthy Again movement, including a silly executive order on glyphosate that feels anathema to what we have fought for. I’d be lying if I said that my heart hasn’t been bent toward repentance for my part in the whole thing. I helped champion Bobby Kennedy as a campaign volunteer, and when he joined up with then-candidate Donald Trump, I reluctantly decided that the trade-offs were worth what I believed Kennedy could advocate for within the walls of a Trump White House: the best fixes for a very sick and broken nation.
Yet I found myself recently, and reluctantly, headed to the citadel of arrogance: Washington (well, Arlington, Va., to be more specific). At the invitation of Brinda Adhikari — one of the hosts of the podcast “Why Should I Trust You?” — I attended the Association of Schools and Programs of Public Health’s annual meeting, where I spoke on a panel about engaging in civil conversation in a session called “A Dialogue Between Academic Public Health and MAHA.”
In Memoriam: Judith L. Rapoport, MD
Dr. Judith L. Rapoport has left an indelible mark on the field of obsessive compulsive disorder (OCD) — not only through her extraordinary scientific contributions, but through the compassion, curiosity, and humanity she brought to her work. For countless individuals and families, her legacy is not just measured in research breakthroughs, but in hope restored and lives changed.
At a time when OCD was widely misunderstood, often hidden, and rarely discussed, Dr. Rapoport helped bring it into the light. Through her pioneering work at the National Institute of Mental Health, she gave shape and voice to a condition that many struggled to name. She was among the first to recognize that OCD could affect children, and that these young people deserved understanding, accurate diagnosis, and effective care. This insight alone transformed the trajectory of the field and opened doors for earlier intervention and support for families who had long felt alone.
What set Dr. Rapoport apart was not only her intellect, but her deep commitment to the people behind the science. She approached each question with both rigor and empathy, helping to establish treatments that have since become the gold standard, including exposure and response prevention (ERP) and medication. Her work helped shift the narrative—away from blame or misunderstanding, and toward recognition of OCD as a real, treatable medical condition.
Beyond the lab and clinic, Dr. Rapoport had a rare gift for storytelling. Her book, The Boy Who Couldn’t Stop Washing, brought readers into the lived experience of OCD with clarity and care. For many, it was the first time they saw their own struggles reflected with such honesty and dignity. It helped families feel seen, understood, and less alone — an impact that continues to ripple outward today. The Boy Who Couldn’t Stop Washing impacted professionals as well, providing an eye-opening introduction and gateway to the world of working with OCD.
For these accomplishments and more, Dr. Rappaport received the IOCDF’s 2018 Career Achievement Award. Her influence extends through the many clinicians and researchers she has mentored, each carrying forward her dedication to both excellence and empathy. Through them, her work continues to grow, shaping the future of OCD research and care in ways that are both profound and deeply human.
To honor Dr. Judith Rapoport is to honor a career defined not only by discovery, but by kindness and purpose. She helped the world better understand OCD — but more importantly, she helped people living with OCD feel understood. And in doing so, she changed lives in ways that will endure for generations.
The post In Memoriam: Judith L. Rapoport, MD appeared first on International OCD Foundation.
Is fake grass a bad idea? The AstroTurf wars are far from over.
A rare warm spell in January melted enough snow to uncover Cornell University’s newest athletic field, built for field hockey. Months before, it was a meadow teeming with birds and bugs; now it’s more than an acre of synthetic turf roughly the color of the felt on a pool table, almost digital in its saturation. The day I walked up the hill from a nearby creek to take a look, the metal fence around the field was locked, but someone had left a hallway-size piece of the new simulated grass outside the perimeter. It was bristly and tough, but springy and squeaky under my booted feet. I could imagine running around on it, but it would definitely take some getting used to.
My companion on this walk seemed even less favorably disposed to the thought. Yayoi Koizumi, a local environmental advocate, has been fighting synthetic-turf projects at Cornell since 2023. A petite woman dressed that day in a faded plum coat over a teal vest, with a scarf the colors of salmon, slate, and sunflowers, Koizumi compulsively picked up plastic trash as we walked: a red Solo cup, a polyethylene Dunkin’ container, a five-foot vinyl panel. She couldn’t bear to leave this stuff behind to fragment into microplastic bits—as she believes the new field will. “They’ve covered the living ground in plastic,” she said. “It’s really maddening.”
The new pitch is one part of a $70 million plan to build more recreational space at the university. As of this spring, Cornell plans to install something like a quarter million square feet of synthetic grass—what people have colloquially called “astroturf” since the middle of the last century. University PR says it will be an important part of a “health-promoting campus” that is “supportive of holistic individual, social, and ecological well-being.” Koizumi runs an anti-plastic environmental group called Zero Waste Ithaca, which says that’s mostly nonsense.
This fight is more than just the usual town-versus-gown tension. Synthetic turf used to be the stuff of professional sports arenas and maybe a suburban yard or two; today communities across the United States are debating whether to lay it down on playgrounds, parks, and dog runs. Proponents say it’s cheaper and hardier than grass, requiring less water, fertilizer, and maintenance—and that it offers a uniform surface for more hours and more days of the year than grass fields, a competitive advantage for athletes and schools hoping for a more robust athletic program.
But while new generations of synthetic turf look and feel better than that mid-century stuff, it’s still just plastic. Some evidence suggests it sheds bits that endanger users and the environment, and that it contains PFAS “forever chemicals”—per- and polyfluoroalkyl substances, which are linked to a host of health issues. The padding within the plastic grass is usually made from shredded tires, which might also pose health risks. And plastic fields need to be replaced about once a decade, creating lots of waste.
Yet people are buying a lot of the stuff. In 2001, Americans installed just over 7 million square meters of synthetic turf, just shy of 11,000 metric tons. By 2024, that number was 79 million square meters—enough to carpet all of Manhattan and then some, almost 120,000 metric tons. Synthetic turf covers 20,000 athletic fields and tens of thousands of parks, playgrounds, and backyards. And the US is just 20% of the global market.
Where real estate is limited and demand for athletic facilities is high, artificial turf is tempting. “It all comes down to land and demand.”
Frank Rossi, professor of turf science, Cornell
Those increases worry folks who study microplastics and environmental pollution. Any actual risk is hard to parse; the plastic-making industry insists that synthetic fields are safe if properly installed, but lots of researchers think that isn’t so. “They’re very expensive, they contain toxic chemicals, and they put kids at unnecessary risk,” says Philip Landrigan, a Boston College epidemiologist who has studied environmental toxins like lead and microplastics.
But at Cornell, where real estate is limited and demand for athletic facilities is high, synthetic turf was a tempting option. As Frank Rossi, a professor of turf science at Cornell, told me: “It all comes down to land and demand.”
In 1965, Houston’s new, domed baseball stadium was an icon of space-age design. But the Astrodome had a problem: the sun. Deep in the heart of Texas, it shined brightly through the Astrodome’s skylights—so much so that players kept missing fly balls. So the club painted over the skylights. Denied sunlight, the grass in the outfield withered and died.
A replacement was already in the works. In the late 1950s a Ford Foundation–funded educational laboratory determined that a soft, grasslike surface material would give city kids more places to play outside and had prevailed upon the Monsanto corporation to invent one. The result was clipped blades of nylon stuck to a rubber base, which the company called ChemGrass. Down it went into Houston’s outfield, where it got a new, buzzier name: AstroTurf.
AP PHOTO/ED KOLENOVSKY, FILE
That first generation of simulated lawn was brittle and hard, but quality has improved. Today, there are a few competing products, but they’re all made by extruding a petroleum-based polymer—that’s plastic—through tiny holes and then stitching or fusing the resulting fibers to a carpetlike bottom. That gets attached to some kind of padding, also plastic. In the 1970s the industry started layering that over infill, usually sand; by the 1990s, “third generation” synthetic turf had switched to softer fibers made of polyethylene. Beneath that, they added infill that combined sand and a soft, cheap shredded rubber made from discarded automobile tires, which pile up by the hundreds of millions every year. This “crumb rubber” provides padding and fills spaces between the blades and the backing.
In the early 1980s, nearly half the professional baseball and football fields in the US had synthetic turf. But many players didn’t like it. It got hotter than real grass, gave the ball different action, and seemed to be increasing the rate of injuries among athletes. Since the 1990s, most pro sports have shifted back toward grass—water and maintenance costs pale in comparison to the importance of keeping players happy or sparing them the risk of injury.
But at the same time, more universities and high schools are buying the artificial stuff. The advantages are clear, especially in places where it rains either too much or not enough. A natural-grass field is usable for a little more than 800 hours a year at the most, spread across just eight months in the cooler, wetter northern US. An artificial-turf field can see 3,000 hours of activity per year. For sports like lacrosse, which begins in late winter, this makes artificial turf more appealing. Most lacrosse pitches are now synthetic. So are almost all field hockey pitches; players like the way the even, springy turf makes the ball bounce.
Furthermore, supporters say synthetic turf needs less maintenance than grass, saving money and resources. That’s not always true; workers still have to decompact the playing surface and hose it off to remove bird poop or cool it down. Sometimes the infill needs topping up. But real grass allows less playing time, and because grass athletic fields often need to be rotated to avoid damage, synthetic ground cover can require less space. Hence the market’s explosive growth in the 21st century.
The city and town of Ithaca—two separate political entities with overlapping jurisdiction over Cornell construction projects—held multiple public meetings about the university’s new synthetic fields: the field hockey pitch and a complex called the Meinig Fieldhouse. Koizumi’s group turned up in force, and a few folks who worked at Cornell came to oppose the idea too—submitting pages of citations and studies on the risks of synthetic grass.
At two of those meetings, dozens of Cornell athletes turned out to support the turf. Representatives of the university and the athletic department declined to speak with me for this story, citing an ongoing lawsuit from Zero Waste Ithaca. But before that, Nicki Moore, Cornell’s director of athletics, told a local newspaper that demand from campus groups and sports teams meant the fields were constantly overcrowded. “Activities get bumped later and later, and sometimes varsity teams won’t start practicing until 10 at night, you know?” Moore told the paper. “Availability of all-weather space should normalize scheduling a great deal.”
That argument wasn’t universally convincing. “It’s a bad idea, but that’s from the environmental perspective,” says Marianne Krasny, director of Cornell’s Civic Ecology Lab and one of the speakers at those hearings. “Obviously the athletic department thinks it’s a great idea.”
GETTY IMAGES
Members of Cornell on Fire, a climate action group with members from both the university and the town, joined in opposing the use of artificial turf, citing the fossil-fuel origins of the stuff. They described the nominal support of the project from student athletes as inauthentic, representing not grassroots support but, yes, an astroturf campaign.
Sorting out the actual science here isn’t simple. Over time, the plastic that synthetic turf is made of sheds bits of itself into the environment. In one study, published in 2023 in the journal Environmental Pollution, researchers found that 15% of the medium-size and microplastic particles in a river and the Mediterranean Sea outside Barcelona, Spain, came from artificial turf, mostly in the form of tiny green fibers. Back in 2020, the European Chemicals Agency estimated that infill material from artificial-turf fields in the European Union was contributing 16,000 metric tons of microplastics to the environment each year—38% of all annual microplastic pollution. Most of that came from the crumb rubber infill, which Europe now plans to ban by 2031.
This pollution worries the Cornell activists. Ithaca is famous for scenic gorges and waterways. The new field hockey pitch is uphill from a local creek that empties into Cayuga Lake, the longest of the Finger Lakes and the source of drinking water for over 40,000 people.
And it’s not just the plastic bits. When newer generations of synthetic turf switched to durable high-density polyethylene, the new material gunked up the extruders used in the manufacturing process. So turf makers started adding fluorinated polymers—a type of PFAS. Some of these environmentally persistent “forever chemicals” cause cancer, disrupt the endocrine system, or lead to other health problems. Research in several different labs has found PFAS in many types of plastic grass.
But the key to assessing the threat here is exposure. Heather Whitehead, an analytical chemist then at the University of Notre Dame, found PFAS in synthetic turf at levels around five parts per billion—but estimated it’d be in water running off the fields at three parts per trillion; for context, the US Environmental Protection Agency’s legal drinking-water limit on one of the most widespread and dangerous PFAS chemicals is four parts per trillion. “These chemicals will wash off in small amounts for long periods of time,” says Graham Peaslee, Whitehead’s advisor and an emeritus nuclear physicist who studies PFAS concentrations. “I think it’s reason enough not to have artificial turf.”
This gets confusing, though. There are over 16,000 different types of PFAS, few have been well studied, and different companies use different manufacturing techniques. Companies represented by the Synthetic Turf Council now “use zero intentionally added PFAS,” says Melanie Taylor, the group’s president. “This means that as the field rolls off the assembly line, there are zero PFAS-formulated materials present.”
Some researchers are skeptical of the industry’s assurances. They’re hard to confirm, especially because there are a lot of ways to test for PFAS. The type of synthetic turf going onto the new field hockey pitch at Cornell is called GreenFields TX; the university had a sample tested using an EPA method that looks for 40 different PFAS compounds. It came back negative for all of them. The local activists countered that the test doesn’t detect the specific types they’re most concerned about, and in 2025 they paid for three more tests on newly purchased synthetic turf. Two clearly found fluorine—the F in “PFAS”—and one identified two distinct PFAS compounds. (The company that makes GreenFields TX, TenCate, declined to comment, citing ongoing litigation.)
PFAS isn’t the only potential problem. There’s also the crumb rubber made from tires. A billion tires get thrown out every year worldwide, and if they aren’t recycled they sit in giant piles that make great habitats for rats and mosquitoes; they also occasionally catch fire. Lots of the tires that go into turf are made of styrene-butadiene rubber, or SBR. In bulk, that’s bad. Butadiene is a carcinogen that causes leukemia, and fumes from styrene can cause nervous system damage. SBR also contains high levels of lead.
But how much of that comes out of synthetic-turf infill? Again, that’s hotly debated. Researchers around the world have published suggestive studies finding potentially dangerous levels of heavy metals like zinc and lead in synthetic turf, with possible health risks to people using the fields. But a review of many of the relevant studies on turf and crumb rubber from Canada’s National Collaborating Centre for Environmental Health determined that most well-conducted health risk assessments over the last decade found exposures below levels of concern for cancer and certain other diseases. A 2017 report by the European Chemicals Agency—the same people who found all those microplastics in the environment—“found no reason to advise people against playing sports on synthetic turf containing recycled rubber granules as infill material.” And a multiyear study from the EPA, published in 2024, found much the same thing—although the researchers said that levels of certain synthetic chemicals were elevated inside places that used indoor artificial turf. They also stressed that the paper was not a risk assessment.
The problem is, the kinds of cancers these chemicals can cause may take decades to show up. Long-term studies haven’t been done yet. All the evidence available so far is anecdotal—like a series for the Philadelphia Inquirer that linked the deaths of six former Phillies players from a rare type of brain cancer called glioblastoma to years spent playing on PFAS-containing artificial turf. That’d be about three times the usual rate of glioblastoma among adult men, but the report comes with a lot of cautions—small sample size, lots of other potential causes, no way to establish causation.
Synthetic turf has one negative that no one really disputes: It gets very hot in the sun—as hot as 150 °F (66 °C). This can actually burn players, so they often want to avoid using a field on very hot days.
GETTY IMAGES
Athletes playing on artificial turf also have a higher rate of foot and ankle injuries, and elite-level football players seem to be more predisposed to knee injuries on those surfaces. But other studies have found rates of knee and hip injury to be roughly comparable on artificial and natural turf—a point the landscape architect working on the Cornell project made in the information packet the university sent to the city. Athletic departments and city parks departments say that the material’s upsides make it worthwhile, given that there’s no conclusive proof of harm.
Back in Ithaca, Cornell hired an environmental consulting firm called Haley & Aldrich to assess the evidence. The company concluded that none of the university’s proposed installations of artificial turf would have a negative environmental impact. People from Cornell on Fire and Zero Waste Ithaca told me they didn’t trust the firm’s findings; representatives from Haley & Aldrich declined to comment.
Longtime activists say that as global consumption of fossil fuels declines, petrochemical companies are desperate to find other markets. That means plastics. “There’s a big push to shift more petrochemicals into plastic products for an end market,” says Jeff Gearhart, a consumer product researcher at the Ecology Center. “Industry people, with a vested interest in petrochemicals, are looking to expand and build out alternative markets for this stuff.”
All that and more went before the decision-makers in Ithaca. In September 2024, the City of Ithaca Planning Board unanimously issued a judgment that the Meinig Fieldhouse would not have a significant environmental impact and thus would not need to complete a full environmental impact assessment. Six months later, the town made the same determination for the field hockey pitch.
Zero Waste Ithaca sued in New York’s supreme court, which ruled against the group. Koizumi and lawyers from Pace University’s Environmental Litigation Clinic have appealed. She says she’s still hopeful the court might agree that Ithaca authorities made a mistake by not requiring an environmental impact statement from the college. “We have the science on our side,” she says.
Ithaca is a pretty rarefied place, an Ivy League university town. But these same tensions—potential long-term environmental and public health consequences versus the financial and maintenance concerns of the now—are pitting worried citizens against their representatives and city agencies around the country.
New York City has 286 municipal synthetic-turf fields, with more under construction. In Inwood, the northernmost neighborhood in Manhattan, two fields were approved via Zoom meetings during the pandemic, and Massimo Strino, a local artist who makes kaleidoscopes, says he found out only when he saw signs announcing the work on one of his daily walks in Inwood Hill Park, along the Hudson River. He joined a campaign against the plan, gathering more than 4,300 signatures. “I was canvassing every weekend,” Strino says. “You can count on one hand, literally, the number of people who said they were in favor.”
But that doesn’t include the group that pushed for one of those fields in the first place: Uptown Soccer, which offers free and low-cost lessons and games to 1,000 kids a year, mostly from underserved immigrant families. “It was turning an unused community space into a usable space,” says David Sykes, the group’s executive director. “That trumped the sort of abstract concerns about the environmental impacts. I’m not an expert in artificial turf, but the parks department assured me that there was no risk of health effects.”
Artificial turf doesn’t go away. “You’re going to be paying to get rid of it. Somebody will have to take it to a dump, where it will sit for a thousand years.”
Graham Peaslee, emeritus nuclear physicist studying PFAS concentrations, University of Notre Dame
New York City councilmember Christopher Marte disagrees. He has introduced a bill to ban new artificial turf from being installed in parks, and he hopes the proposal will be taken up by the Parks Committee this spring. Last session, the bill had 10 cosponsors—that’s a lot. Marte says he expects resistance from lobbyists, but there’s precedent. The city of Boston banned artificial turf in 2022.
Upstate, in a Rochester suburb called Brighton, the school district included synthetic-turf baseball and softball diamonds in a wide-ranging February 2024 capital improvement proposition. The measure passed. In a public meeting in November 2025, the school board acknowledged the intent to use synthetic grass—or, as concerned parents had it, “to rip up a quarter million square feet of this open space and replace it with artificial turf,” says David Masur, executive director of the environmental group PennEnvironment, whose kids attend school in Brighton. Parents and community members mobilized against the plan, further angered when contractors also cut down a beloved 200-year-old tree. School superintendent Kevin McGowan says it’s too late to change course. Masur has been working to oppose the plan nevertheless—he says school boards are making consequential decisions about turf without sharing information or getting input, even though these fields can cost millions of dollars of taxpayer money.
In short, the fights can get tense. On Martha’s Vineyard, in Massachusetts, a meeting about plans to install an artificial field at a local high school had to be ended early amid verbal abuse. A staffer for the local board of health who voiced concern about PFAS in the turf quit the board after discovering bullet casings in her tote bag, she said, which she perceived as a death threat. After an eight-year fight, the board eventually banned artificial turf altogether.
What happens next? Well, outdoor artificial turf lasts only eight to 12 years before it needs to be taken up and replaced. The Synthetic Turf Council says it’s at least partially recyclable and cites a company called BestPLUS Plastic Lumber as a purveyor of products made from recycled turf. The company says one of its products, a liner called GreenBoard that artificial turf can be nailed into, is at least 40% recycled from fake grass. Joseph Sadlier, vice president and general manager of plastics recycling at BestPLUS, says the company recycles over 10 million pounds annually.
Yet the material is piling up. In 2021, a Danish company called Re-Match announced plans to open a recycling plant in Pennsylvania and began amassing thousands of tons of used plastic turf in three locations. The company filed for bankruptcy in 2025.
In Ithaca, university representatives told planning boards that it would be possible to recycle the old artificial turf they ripped out to make way for the Meinig Fieldhouse. That didn’t happen. An anonymous local activist tracked the old rolls to a hauling company a half-hour’s drive south of campus and shared pictures of them sitting on the lot, where they stayed for months. It’s unclear what their ultimate fate will be.
That’s the real problem: Artificial turf just doesn’t go away. “You’re going to be paying to get rid of it,” says Peaslee, the PFAS expert. “Somebody will have to take it to a dump, where it will sit for a thousand years.” At minimum, real grass is a net carbon sink, even including installation and maintenance. Synthetic turf releases greenhouse gases. One life-cycle analysis of a 2.2-acre synthetic field in Toronto determined that it would emit 55 metric tons of carbon dioxide over a decade. Plastic fields need less water to maintain, but it takes water to make plastic, and natural grass lets rainwater seep into the ground. Synthetic turf sends most of it away as runoff.
It’s a boggling set of issues to factor into a decision. Rossi, the Cornell turf scientist, says he can understand why a school in the northern United States might go plastic, even when it cares about its students’ health. “It was the best bad option,” he says. Concerns about microplastics and PFAS are “significant issues we have not fully addressed.” And they need to be.
Douglas Main is a journalist and former senior editor and writer at National Geographic.
A Social Justice Approach to Assistive Technology and Well-Being of People With Visual Disabilities in Low- and Middle-Income Countries: Qualitative Narrative Study
Background: The United Nations’ third Sustainable Development Goal emphasizes ensuring healthy lives and promoting well-being (WB) for all, which requires effective assistive technology (AT) for persons with disabilities. In low- and middle-income countries (LMICs), however, AT remains largely inaccessible, and high abandonment rates indicate that many existing solutions fail to meet users’ needs. To improve AT design and effectiveness, a deeper understanding of users’ lived experiences and the ways AT influences WB is essential. Objective: This study aimed to explore how technology creates opportunities or barriers in the daily lives of persons with visual disabilities in LMICs and how it affects their WB. Methods: We conducted a qualitative narrative study guided by deductive qualitative analysis, using the capability approach (CA) and disadvantage theory (DT) as theoretical frameworks. Nineteen adults with visual disabilities from Cali, Colombia, participated in in-depth, semistructured interviews. A focus group (n=5) deepened the exploration of shared experiences. Data analysis followed three stages: (1) deductive coding using Nussbaum list of central capabilities and key CA constructs (functionings, conversion factors, and agency); (2) recoding through DT concepts (insecure functioning, corrosive disadvantages, and fertile functionings); and (3) inductive analysis to capture emergent sociocultural themes. Results: AT shaped both opportunities and constraints in participants’ lives. While functionings such as employment, mobility, and affiliation were highly valued, they often remained insecure due to systemic barriers. Corrosive disadvantages—such as unemployment, exclusion, and limited spatial autonomy—undermined multiple capabilities simultaneously. Conversely, fertile functionings such as equitable employment, adaptive sports, and access to well-designed AT supported agency and resilience. The inductive analysis revealed 3 interconnected themes: the aspiration to explore and expand movement, the desire to appear attractive, and the adoption of nonconfrontational strategies to maintain social harmony. These findings highlight how emotional, aesthetic, and cultural dimensions shape the experience and meaning of AT. Conclusions: While AT research in LMICs often emphasizes availability, it rarely addresses how social norms, structural violence, and fear affect meaningful use. The combined CA and DT lens reveals that AT can either enable or constrain WB depending on how it aligns with users’ lived contexts. Designing for fertile functionings—those that support agency, safety, and resilience—is essential. Participatory, context-sensitive design must prioritize not only functionality, but also aesthetic dignity, cultural relevance, and emotional security. Including the voices—and silences—of persons with disabilities in the Global South is crucial for transforming AT from a mere tool into a catalyst for real freedom and WB.
<img src="https://jmir-production.s3.us-east-2.amazonaws.com/thumbs/b69af78b31160b1f7cc12c3b94f9456d" />
Faster Process Development via “Transfer Learning”
An emerging artificial intelligence technique called “transfer learning” could help drug makers use data to speed up the development of biopharmaceutical manufacturing processes, according to new analysis.
In transfer learning, predictive models that have been trained on historical data are used to improve the performance of a task.
Unlike machine learning (ML)—where the training process begins from scratch—transfer learning applies existing knowledge to new but related problems, reducing the amount of data and time required to build the model.
Researchers at the Karlsruhe Institute of Technology in Germany, who looked at the approach, identified several potential biopharma applications, according to lead author Daniel Barón Díaz, citing reactor modeling as an example.
“Transfer learning models can be used to predict critical outcomes like viable cell density (VCD) and product titre from online sensor data—for example, pH, temperature, gas flow—from historical data from a different, but related process.”
The approach can also optimize process monitoring. Díaz tells GEN that, “Transfer learning-enhanced soft sensors can be established to monitor protein concentrations in real-time by leveraging existing models from related fermentations.”
Data limitation
When compared with other model-building techniques, transfer learning offers potential cost and time savings, according to Díaz, who cites a reduced experimentation burden as an example.
“Conventional machine learning requires large, structured datasets that are often unavailable in biopharma due to the high cost and labor-intensive nature of experiments. Transfer learning allows companies to leverage historical data and existing models to build reliable predictors for new processes with very limited data.
“By reusing prior knowledge, transfer learning can significantly decrease the number of experiments required—sometimes needing only one to three batches to achieve robust simulations,” he says.
However, the ultimate benefit is that transfer learning speeds up process model development, according to Díaz, who adds, “It can make model adaptation faster than retraining from scratch, facilitating quicker process design and digital twin deployment.”
Challenges
So, transfer learning has the potential to create predictive models for manufacturing development. However, the key caveat is that the processes involved must be sufficiently similar for it to be effective, Díaz says.
“For transfer learning to be effective, the source and target domains must be meaningfully related. If the processes are too different, the assumptions and learned representations may not align, leading to negative transfer, where the transferred knowledge actually degrades the model’s performance.
“Data sets obtained at different scales or under varying conditions are often inconsistent, which can hinder the successful transfer of knowledge. Fine-tuning complex neural network architectures on very small target datasets can lead to overfitting, where the model fails to generalize to new data,” he says.
To address this, manufacturers will need to establish metrics to determine similarity, Díaz explains.
“There are currently no standardized metrics for measuring domain similarity in bioprocessing, nor are there comprehensive benchmark datasets to easily compare different transfer learning techniques.”
Another challenge is the current lack of AI expertise in the industry, Díaz says.
“There is often a disciplinary knowledge gap between process engineers and data scientists, and ML models without a mechanistic backbone may be perceived as opaque black boxes, hindering trust and industrial adoption,” he tells GEN.
The post Faster Process Development via “Transfer Learning” appeared first on GEN – Genetic Engineering and Biotechnology News.
Lung Screening Incidental Findings May Guide Follow-Up for Other Cancers
An analysis of the US National Lung Screening Trial (NLST) has found that the presence of certain types of abnormalities in regions outside of the lungs on low-dose computed tomography (LDCT) images may be associated with a significantly increased risk for extrapulmonary cancer.
The abnormalities, termed significant incidental findings (SIFs), could help clinicians decide when follow-up care is likely to catch extrapulmonary cancer early and when it may not be necessary.
“In this paper, we provide an evidence base for making decisions on abnormalities outside of the lungs that might be seen at lung screening,” said study author Ilana Gareen, PhD, a professor of epidemiology at Brown University School of Public Health. “The goal is to give physicians and patients better data so that they can make more informed choices about those abnormalities that should be considered for follow-up and those that most likely can be ignored.”
Writing in JAMA Network Open, Gareen and co-authors explain that LDCT lung cancer screening frequently detects SIFs unrelated to lung cancer; in the NLST, 34% of 26,455 patients screened with LDCT had SIFs reported but the nature of the SIFs varied.
And although there are recommendations for reporting and addressing SIFs, there is limited evidence for an association between SIFs detected at LDCT lung cancer screening and extrapulmonary cancer diagnoses.
To address this, Gareen and team analyzed data from 75,104 LDCT screening rounds performed in 26,445 individuals (mean age, 61 years; 59.0% men) who were randomly assigned to receive LDCT during the NSLT. The participants had a history of heavy smoking (≥30 pack–years), meaning they are also at high risk for several extrapulmonary cancers, including pancreatic, bladder, and kidney cancer.
The researchers focused on SIFs that were labelled as potentially indicative of extrapulmonary cancer (cancer SIF), rather than those that possibly indicated emphysema or cardiovascular disease.
They report that cancer SIFs were recorded for 2265 (3.0%) screening rounds in 1807 (6.8%) participants across the three screening rounds they received.
Participants with cancer SIFs were significantly older than those with no cancer SIF (mean 62.1 vs. 61.4 years) and significantly more likely to have a history of a smoking-related disease (68.6 vs. 65.7%).
Within one year of a screening round, 1025 participants were diagnosed with an extrapulmonary cancer. Of these, 67 (6.5%) had a SIF on LDCT. This corresponds to 3.0% of participants with a cancer SIF.
Overall, the risk for extrapulmonary cancer among the people with a cancer SIF was 29.6 per 1000 screening rounds compared with 13.3 per 1000 screening rounds in those without a cancer SIF. After adjustment for potential confounders, the marginal risk difference between the two groups was 13.9 per 1000 participants, suggesting that for every 1000 people screened, the presence of a cancer SIF is associated with 13.9 additional cases of extrapulmonary cancer.
When the researchers looked at specific cancer types, they found that the marginal risk difference was substantially higher for urinary cancers, at 17.0 per 1000 participants. It was 5.0 for digestive cancer, 12.3 for breast cancer, and 13.8 for other cancers including lymphoma and leukemia.
“In general, if an abnormality is found that might indicate cancer, the patient receives additional imaging to evaluate that abnormality,” Gareen told Inside Precision Medicine. “Our paper provides additional information as to those abnormalities that should be considered to increase the risk of a cancer diagnosis.”
Importantly, mortality from extrapulmonary cancer accounted for 22.3% of the certified deaths in the LDCT arm of the NLST. Therefore “early detection of these cancers may facilitate early treatment and potentially reduce associated morbidity and mortality,” the authors write. “Identification of cancer SIFs associated with extrapulmonary cancers in NLST participants could be used to plan appropriate diagnostic evaluations for patients undergoing lung cancer screening.”
Gareen said the next step will be to determine if the findings are replicated in lung screening in the community, or if the rate in community screening is higher or lower.
In accompanying comment, Patrick Senior and Andrew Creamer, both from Gloucestershire Hospitals NHS Foundation Trust, in Gloucester, United Kingdom, point out that the false positive rate for a cancer SIF was 97% but say “it is hard to imagine a scenario in which an incidental finding with even a possibility of representing cancer would be disregarded.”
However, they note that “when considered in the context of the numbers of people eligible for lung cancer screening programs around the world, acting on such findings poses a considerable additional burden on the health systems that must investigate them.”
Senior and Creamer say that the results “underscore the importance of both a robust health economics analysis of how screening programs manage such incidental findings and patient-centered research to understand the impact that such unexpected results may have on the individual. Further research is needed to ensure that screening programs are confident when faced with information they did not ask for.”
The post Lung Screening Incidental Findings May Guide Follow-Up for Other Cancers appeared first on Inside Precision Medicine.