Transistors, the building blocks of modern electronics, are typically made of silicon. Because it’s a semiconductor, this material can control the flow of electricity in a circuit. But silicon has fundamental physical limits that restrict how compact and energy-efficient a transistor can be.

MIT researchers have now replaced silicon with a magnetic semiconductor, creating a magnetic transistor that could enable smaller, faster, and more energy-efficient circuits. The material’s magnetism strongly influences its electronic behavior, leading to more efficient control of the flow of electricity. 

The team used a novel magnetic material and an optimization process that reduces the material’s defects, which boosts the transistor’s performance.

The material’s unique magnetic properties also allow for transistors with built-in memory, which would simplify circuit design and unlock new applications for high-performance electronics.

“People have known about magnets for thousands of years, but there are very limited ways to incorporate magnetism into electronics. We have shown a new way to efficiently utilize magnetism that opens up a lot of possibilities for future applications and research,” says Chung-Tao Chou, an MIT graduate student in the departments of Electrical Engineering and Computer Science (EECS) and Physics, and co-lead author of a paper on this advance.

Chou is joined on the paper by co-lead author Eugene Park, a graduate student in the Department of Materials Science and Engineering (DMSE); Julian Klein, a DMSE research scientist; Josep Ingla-Aynes, a postdoc in the MIT Plasma Science and Fusion Center; Jagadeesh S. Moodera, a senior research scientist in the Department of Physics; and senior authors Frances Ross, TDK Professor in DMSE; and Luqiao Liu, an associate professor in EECS, and a member of the Research Laboratory of Electronics; as well as others at the University of Chemistry and Technology in Prague. The paper appears today in Physical Review Letters.

Overcoming the limits

In an electronic device, silicon semiconductor transistors act like tiny light switches that turn a circuit on and off, or amplify weak signals in a communication system. They do this using a small input voltage.

But a fundamental physical limit of silicon semiconductors prevents a transistor from operating below a certain voltage, which hinders its energy efficiency.

To make more efficient electronics, researchers have spent decades working toward magnetic transistors that utilize electron spin to control the flow of electricity. Electron spin is a fundamental property that enables electrons to behave like tiny magnets.

So far, scientists have mostly been limited to using certain magnetic materials. These lack the favorable electronic properties of semiconductors, constraining device performance.

“In this work, we combine magnetism and semiconductor physics to realize useful spintronic devices,” Liu says.

The researchers replace the silicon in the surface layer of a transistor with chromium sulfur bromide, a two-dimensional material that acts as a magnetic semiconductor.

Due to the material’s structure, researchers can switch between two magnetic states very cleanly. This makes it ideal for use in a transistor that smoothly switches between “on” and “off.”

“One of the biggest challenges we faced was finding the right material. We tried many other materials that didn’t work,” Chou says.

They discovered that changing these magnetic states modifies the material’s electronic properties, enabling low-energy operation. And unlike many other 2D materials, chromium sulfur bromide remains stable in air.

To make a transistor, the researchers pattern electrodes onto a silicon substrate, then carefully align and transfer the 2D material on top. They use tape to pick up a tiny piece of material, only a few tens of nanometers thick, and place it onto the substrate.

“A lot of researchers will use solvents or glue to do the transfer, but transistors require a very clean surface. We eliminate all those risks by simplifying this step,” Chou says.

Leveraging magnetism

This lack of contamination enables their device to outperform existing magnetic transistors. Most others can only create a weak magnetic effect, changing the flow of current by a few percent or less. Their new transistor can switch or amplify the electric current by a factor of 10.

They use an external magnetic field to change the magnetic state of the material, switching the transistor using significantly less energy than would usually be required.

The material also allows them to control the magnetic states with electric current. This is important because engineers cannot apply magnetic fields to individual transistors in an electronic device. They need to control each one electrically.

The material’s magnetic properties could also enable transistors with built-in memory, simplifying the design of logic or memory circuits.

A typical memory device has a magnetic cell to store information and a transistor to read it out. Their method can combine both into one magnetic transistor.

“Now, not only are transistors turning on and off, they are also remembering information. And because we can switch the transistor with greater magnitude, the signal is much stronger so we can read out the information faster, and in a much more reliable way,” Liu says.

Building on this demonstration, the researchers plan to further study the use of electrical current to control the device. They are also working to make their method scalable so they can fabricate arrays of transistors.

This research was supported, in part, by the Semiconductor Research Corporation, the U.S. Defense Advanced Research Projects Agency (DARPA), the U.S. National Science Foundation (NSF), the U.S. Department of Energy, the U.S. Army Research Office, and the Czech Ministry of Education, Youth, and Sports. The work was partially carried out at the MIT.nano facilities.

In the 20th century, the United States built the world’s dominant manufacturing powerhouse. A thriving middle class grew, well into the 1970s. The U.S. was a beacon of democracy, defeating fascism in World War II and beating back communism and other forms of authoritarianism during the Cold War. 

To MIT economist Daron Acemoglu, there is a deep intertwining among these things. Democracy, his work has shown, helps economies grow. As the industrial economy expanded, in Britain, the U.S., and other countries in the 19th and 20th centuries, so did democratic participation, as people tried to stake out new rights, or make real the rights ascribed to them. 

“The industrial age created the tools for shared prosperity around which democracy organized,” says Acemoglu, a Nobel Prize-winning economist and Institute Professor at MIT. 

Today, though, income inequality has grown markedly in the U.S., starting around 1980. The U.S. has deindustrialized to a significant extent, offshoring production and hurting shop-floor workers and their families. As Acemoglu sees it, this economic realignment has had deep civic consequences: A stranded working class has become more alienated from the institutions and ideas traditionally buttressing democracy.

And for those around the world supporting democracy, he says, “You really need to have the working classes in your coalition for it to make any sense.”

Acemoglu explores these topics in a new book, “What Happened to Liberal Democracy? Remaking a Politics of Shared Prosperity,” published by Penguin Random House. In it, he looks broadly at the benefits of democracy, the tensions it faces in everyday life, and democracy’s trajectory in recent decades.

Broadly, Acemoglu favors rebuilding “working-class liberalism,” essentially seeking the largest coalition that favors self-government and the rule of law. “Working-class liberalism has strong communal roots, eschews social engineering, and prioritizes shared prosperity, jobs, and public services,” Acemoglu writes in the book.

After all, Acemoglu believes, democracy is the one form of rule that promotes rights and liberties, and allows the flexibility and “experimentation” we need to address all the challenges a complicated world throws at us.

“Democracy is the only way we can make progress in society,” Acemoglu says. “Trying to impose top-down solutions to all our problems will ultimately not work.” 

Along the narrow corridor

Acemoglu has long studied the relationship between economic growth, rights, and democracy. With economist Simon Johnson of MIT and political scientist James Robinson, now of the University of Chicago, Acemoglu published a landmark series of studies in the early 2000s demonstrating that economic growth is helped by the development of stable democratic institutions, including property rights. For that work, Acemoglu, Johnson, and Robinson later shared the 2024 Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel. 

Acemoglu’s 2012 bestseller “Why Nations Fail”— written before democracy’s current challenges seemed as acute — synthesized his research on these topics. His 2019 book “The Narrow Corridor,” co-authored with Robinson, casts democratic governments as essential to liberty because they protect people simultaneously from overreach by an authoritarian state, on the one hand, and from domination by other groups in society, on the other. 

However, as Acemoglu has consistently emphasized, self-governance is an ongoing effort; this machine does not run on its own. Relatedly, in the new book, Acemoglu critiques some famous attempts to formulate governance as a neat “social contract,” including Jean-Jacques Rousseau’s conception of a “general will” in society. 

Those ideas helped make the case for political rights, but actual governance in a pluralistic society will always be a messy process. 

“You have to allow communities, and societies in aggregate, to build rules around shared values for anything to stick as institutions, norms, or aspirations,” Acemoglu says. “When you go down the social contractarian path, you sometimes fool yourself into thinking there are clear solutions to dilemmas that in reality don’t quite have such obvious ways of being resolved.” 

For instance, Acemoglu notes, democracy itself “is built on tolerance and acceptance of plural perspectives, but how do you deal with people who are intolerant?” In the book, he largely regards interventions to stamp out seemingly intolerant thought as being unwise and politically counterproductive.

“You’re not going to have a clear-cut solution to all cases,” he says.

The economic realignment

Even with leaders backing a pragmatic, flexible approach to self-governance, democracy faces another challenge: supporting the material welfare of citizens. And here, “What Happened to Liberal Democracy?” takes an unflinching look at the postwar economy, finding fault lines that have shaken the political order. 

The roughly three decades after World War II were fantastic for many workers in democracies, and certainly in the U.S., where middle-class incomes grew by 2.5 percent annually into the 1970s. 

There were always going to be forces pushing back on this trend, and U.S. companies started offshoring and outsourcing production work to clamp down on wage growth. But one other technological and economic trend occurred just as the middle classes of the industrial economy were reaching new heights. 

“Then computers happened,” writes Acemoglu in the book — referring to a complex set of economic and civic realignments involving technology-driven shifts in work. 

Over time, computers started replacing significant numbers of clerical office workers, shop-floor industrial workers, and other types of employees who were earning middle-class wages without holding a college degree. In recent decades, middle-class incomes have only grown by about 0.5 percent annually. 

To be sure, computers have produced plenty of benefits, and created many new forms of work. But as research shows, those jobs have tended to go mostly to college-educated employees, creating a significant split in society between well-educated, well-paid, white-collar workers, and less-educated, worse-paid workers in blue-collar and service jobs. That national share of income hauled in by the top 1 percent of earners has basically doubled in this time, from 10 percent to nearly 20 percent.

Crucially, in Acemoglu’s analysis, this material gap between more-educated and less-educated social cohorts has translated to U.S. politics, with political groupings reshuffling along educational lines, and cultural politics following suit. That’s the dynamic the U.S. faces now — even as one also accounts for the effects of social media and other polarizing features of contemporary society. 

“We now live in a less-industrial world, and we also live in an age defined by social media, much greater levels of conflict, more polarization, and now AI, and all of that complicates things,” Acemoglu says. 

Always a work in progress

This precise feature of contemporary society — deindustrialization fueling an earnings gap that has led to more political polarization — is what shapes Acemoglu’s prescription in response, the idea that “working-class liberalism” is needed to strengthen democracy. 

There are many potential ingredients in this formula, from politicians determined to reach across class lines to workers regaining the impetus to organize in their workplaces. 

“Trade union participation itself is a very important form of local governance that’s very difficult or unimaginable in an authoritarian society,” Acemoglu says, even while noting that he has not always agreed with the actions of particular unions in the past. 

Still, Acemoglu adds, “I don’t think the economic aspect is the only one, in that you cannot just gain the trust of workers by ensuring there are wage gains. That is an important step but it is not sufficient.” Voters need asurances that politicians are thinking about them, at least share their concerns, and have a grounding in similar values. More candidates today need to seek a shared language about those things.

That’s not easy in a world characterized, in part, by global migrations, increasingly diverse national populations, and economic flux. But it is possible, Acemoglu thinks. 

“There are deep dilemmas faced by liberal democracy that were sometimes going to come to boiling points, and this becomes more heightened when societies such as the U.S. and some European ones are simultaneously becoming more complex, more globalized, and more hetereogeneous,” Acemoglu says. On the other hand, he adds, “Multiracial tensions, I would say, were much worse for the U.S. in the 1950s and 1960s. We made democracy work then, in the face of much more difficult race problems, so why not today?”

None of this is a straightforward task, of course. “Forging working-class liberalism is a tall order in the best of times and much more challenging in today’s polarized environment,” Acemoglu writes in the new book. Still, he adds, even in frustrating moments, the stakes are too important for people to relent.

“Democracy is a success,” Acemoglu says. “It’s easy to fall into a trap of painting the democratic project as being doomed to failure, and I want to avoid that.” He adds: “It continues to be a work in progress.” 

Iron is essential. Our cells need it to produce energy, carry oxygen throughout the body, and power countless chemical reactions that sustain life. But this metal has a dark side. When too much of it is left free inside cells, it can trigger destructive reactions that break down DNA, proteins, and even cell membranes.

Now, MIT associate professor of biology and Whitehead Institute for Biomedical Research member Ankur Jain; MIT assistant professor of biology and Koch Institute for Integrative Cancer Research member Whitney Henry; and Pushkal Sharma PhD '26 have discovered that cells rely on an unexpected protector against this threat: small molecules called polyamines.

The researchers’ detailed findings, published Aug. 14 in the journal Cell, reveal that polyamines act like storage lockers for iron, safely holding the metal in a non-reactive state until cells need it.

These findings solve a decades-old mystery about why cells maintain such extraordinarily high levels of polyamines and uncover a previously unknown defense mechanism that protects cells from toxic iron overload.

This work could also help scientists develop better cancer treatments, by allowing iron overload to trigger cancer cell death. It could also offer new clues about diseases like early-onset Parkinson’s disease, in which mutations affect polyamine levels within neurons.

The Jain Lab studies RNA, the intermediary between DNA and the tiny molecular machines called proteins that perform most of the essential tasks inside cells. The lab is particularly interested in how RNA folds, misfolds, and sometimes clumps inside cells.

Jain and Sharma first began studying polyamines because these molecules bind to RNA and help shape its structure. However, they suspected that polyamines must be playing other roles inside cells: they’re among the most abundant small molecules within cells, present at levels comparable to ATP, the molecule cells use as their energy currency.

“We’ve known that without polyamines, cells stop growing and dividing,” Jain says. “But their best-known function only requires a small fraction of the polyamine levels cells actually have.”

To uncover polyamines’ hidden function inside cells, the researchers used a large-scale genetic approach that allows them to screen the entire genome at once, rather than testing genes one-by-one, in order to find out which cellular processes are impacted when polyamine levels are changed within cells.

The screen revealed that when cells have reduced levels of polyamines, a protein called GPX4 becomes essential for survival. GPX4 is known to prevent harmful chemical reactions that damage the fatty molecules that make up cell membranes. 

The team also found that cells with lower polyamine levels have higher amounts of another protein that acts as an iron sponge and keeps the metal in a mineralized form. Together, these findings led the researchers to hypothesize that polyamines might be helping keep iron in a safe, non-reactive state within cells.

To test this idea, they developed a new fluorescent sensor that would allow them to measure chemically reactive iron inside living cells. The new sensor causes living cells to glow based on the amount of chemically reactive iron they contain, allowing researchers to track any changes under a microscope in real-time.

The team paired the new iron sensor with another sensor they had previously developed that measures polyamine levels within cells. By employing them simultaneously, they observed a striking pattern: As polyamine levels dropped within cells, the amount of chemically reactive iron went up, offering new evidence that polyamines play a key role in preventing toxic iron build up inside cells.

Beyond answering a fundamental biological question, these findings could have implications for cancer treatment. Cancer cells often rely on high polyamine levels to support their rapid growth and division. However, cancer drugs designed to lower polyamine levels to stop cell division have had limited success.

“We saw that when polyamine levels fall, cells rely on GPX4 to protect themselves from iron toxicity,” says Sharma, who is also the first author of the study. “This could mean that combining drugs that lower polyamine levels with those that block GPX4 might be more effective for killing cancer cells than targeting either pathway alone.”

The discovery may also have implications beyond cancer. Mutations in genes that help move polyamines around cells are linked to a rare form of early-onset Parkinson’s disease, and scientists have long observed unusually high levels of iron in the brains of Parkinson’s patients. 

While it is still unclear whether excess iron directly contributes to neuron death in Parkinson’s, the discovery that polyamines help buffer reactive iron inside cells offers a possible explanation for this link and opens new directions for future investigation.

In addition, the researchers expect the new iron sensor to be a valuable tool for other scientists. By allowing them to track chemically reactive iron inside living cells, it could power new discoveries in aging, cancer, and neurodegeneration.

“There are a lot of promising future directions for this work,” Jain says. “It’s exciting to think about how these tools and findings could help answer further questions about disease pathways and potentially help design better therapies.”

This work is supported by grants from the National Institutes of Health, Bumpus Foundation, and Pew Charitable Trusts. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

A former MIT class project is becoming an $850 million effort to manufacture a new kind of aircraft in Ohio.

Electra began as an idea for a hybrid plane that could take off from shorter runways than traditional airplanes but have more range and speed than all-electric aircraft. Now, like other great MIT projects, it’s making an impact far beyond campus.

The company’s fixed-wing aircraft is designed to make travel easier, especially for people who don’t live in the immediate vicinity of a major airport. The plane features a smaller, more efficient engine than traditional planes, along with batteries to give it added power during takeoffs and landings.

With a range of around 1,200 miles and a cruising speed of around 200 miles per hour, the plane could improve the travel experience for many kinds of trips while cutting down on fuel use. And, given the much shorter runway needs and quieter operation than traditional planes, the plane can leverage unique access points such as barges, parking lots, and soccer fields to take off and land instead of traditional airports. 

“Helping people travel between 50 and 250 miles is the sweet spot for this technology,” says Electra Director of Technology Development Chris Courtin SM ’19, PhD ’24. “This can be a better option than driving or commercial airlines for many kinds of trips. There’s a lot of people traveling in that range and a huge amount of friction in existing ground and air transport systems. This could be a big benefit to those people.”

Courtin has worked on the hybrid plane concept since its inception, first as part of a class project at MIT, then as a teacher’s assistant, and finally as part of his PhD thesis. The company was founded by another alumnus, John Langford ’79, SM ’83, SM ’85, PhD ’87, and counts two MIT professors — Mark Drela and John Hansman — as its founding technical advisors.

“The company has really benefited from a strong collaboration with MIT,” Courtin says. “One of the compelling things about MIT is it gives people space to marry the theoretical side with the practical side — to actually go build the airplane and see if people will buy it.”

Electra has already built and flown a two-seated version of its plane. Last month, the company announced an $850 million investment to scale production of its nine-passenger aircraft in Springfield and Clark County, Ohio. The investment, which is expected to create 1,975 new jobs, means Electra will be building the next chapter of aviation in the state where engine-powered human flight first began.

From concept to company

The origins of Electra date back to a 2017 project among graduate students in MIT class 16.886 (Air Transportation Systems Architecting). Electric vertical takeoff and landing (eVTOL) aircraft were garnering excitement at the time, and Courtin’s group wanted to compare that approach to alternative designs.

“It was an open-ended, project-based class where you look at developments in aerospace,” Courtin says. “My group realized short takeoff and landing aircraft had a lot of advantages over eVTOL for getting people where they wanted to go. We started exploring using the same technology — lightweight, electric motors suitable for aviation — to make a new aircraft, which we now call the ultra-short takeoff and landing aircraft.”

The idea was to use batteries and small electric motors to shorten the runway and landing space of a fixed-wing aircraft while leveraging blown wind to travel farther distances in the sky than would be possible with electric motors alone.

The concept was developed further in several senior design classes co-taught by Drela and Hansman, while Courtin served as a teacher’s assistant. In the classes, student collaborators built a subscale model of the aircraft to prove it would work, testing it in MIT’s Wright Brothers Wind Tunnel and in flight. Courtin went on to work on parts of the concept for his PhD. 

In 2019, John Langford, who had been running the aircraft company Aurora Flight Sciences, which had recently been acquired by Boeing, got involved. Electra was officially formed that year.

As a first step, Electra’s team built the EL2, a two-seated version of its aircraft. That included designing and testing the hybrid propulsion system. The EL2 completed its first test flights in 2023 and has since completed over 200 flights.

The aircraft has a gas-powered generator located in its nose and two batteries under the floor, both of which feed the propellers during takeoff and landing. When cruising, the aircraft uses the generator, which can also charge the batteries.

“The gas generator is like a traditional turbine engine used in a conventional aircraft, only instead of driving a propeller or fan it drives an electric generator,” Courtin explains. “That feeds power to the eight motors on the wing. It allows you to have a smaller and more efficient engine because you can size it for cruising, not takeoff and landing conditions.”

The eight motors create a blown lift effect that allows the aircraft to take off and land in areas about the length of a soccer field, much shorter than the runways for conventional airplanes.

For travelers, “the big benefit is you can save a lot of time,” Courtin says. “You don’t need to go to an airport, and you don’t have to go to a train station.”

Operators could also maximize existing infrastructure at airports:

“If you’re three hours away from the nearest major airport, there’s a lot of friction in that,” Courtin says. “With Electra, we could fly you to the nearest major airport, and you don’t need to use a runway, so it doesn’t add to congestion at these very low-capacity places.”

Electra’s aircraft are also more affordable than traditional aircraft and far more quiet.

“The large number of propellers means you can make them much quieter than if you only had one or two,” Courtin explains. “That’s important because helicopters are restricted from operating in places they otherwise could because of the noise.”

Scaling up

Construction on Electra’s 96-acre Ohio manufacturing facility will begin next year. The facility’s initial phase will be capable of producing 400 of its nine-seat aircraft each year. The next phase will expand capacity to around 800 aircraft per year.

Electra’s team could see their aircraft shuttling people to major airports for longer trips or ultimately eliminating the need for conventional airports entirely.

“If you don’t have an existing airport, that’s a very difficult thing to build these days,” Courtin says. “But finding a soccer field-sized area is not hard, especially with our noise reductions.”

Electra’s team is also exploring applications around military logistics, cargo transport, and humanitarian missions.

For the passenger aircraft application, Electra’s team believes that as it scales production, it will be able to make the passenger aircraft accessible to a broad swath of travelers.

“If we can keep the fixed-wing design simplicity and make this large enough, then the per-seat cost could get to a range where a lot of people would have access to this,” Courtin says. “It wouldn’t just be a luxury product, so it could help a lot of people.”

Squid have occupied New England’s maritime imagination for centuries, from fishermen’s tales of strange tentacled creatures to the widely publicized giant squid reports along North Atlantic coastlines in the 1800s. Today, we turn a curious eye toward the longfin inshore squid (Doryteuthis pealeii) — the “Boston squid” — which has long been an important part of regional commercial fisheries. These squid thrive in the emerald green depths off the coast of Massachusetts, where low visibility is characteristic of the turbid, nutrient-rich waters.

Keith Ellenbogen, MIT Sea Grant resident artist and acclaimed underwater photographer, knows these waters well. On a good day, he might have 20 feet of visibility beneath the surface. But even with the clearest conditions, these swift squid evaded Ellenbogen’s lens for months. 

In 2023, over 2.85 million pounds of longfin squid — likely over 6 million individual squid — were landed commercially in Massachusetts, with a value exceeding $2.82 million. Consistently among the top 10 species landed commercially in Massachusetts, the longfin squid shares the ranks with the iconic American lobster, Atlantic surf clam, and sea scallop, underscoring its enduring significance in the state’s seafood industry. 

In 2024, landings came in under 750,000 pounds, roughly a quarter of the previous year’s catch. Likewise, the spring 2024 trawl survey documented in the Massachusetts Division of Marine Fisheries Annual Report recorded a historic low in longfin squid biomass, following a record high in 2023. 

But this drop, while significant, doesn’t necessarily spell trouble for the species. Longfin squid often show large fluctuations in abundance from year to year. Their short lifespan of about six to nine months, combined with their sensitivity to shifting ocean conditions, makes their population dynamics inherently unpredictable. 

Kimberly Hyde, a biological oceanographer with the National Oceanic and Atmospheric Administration (NOAA)’s Northeast Fisheries Science Center (NEFSC), notes that NEFSC research scientists have expanded squid research over the past few years, including a two-year study on longfin squid to better understand their life history and track their maturity and readiness to spawn. Across the Northeast, scientists are also collaborating with the fishing industry through efforts including the Longfin Squid Biological Sampling Program (SQUIBS), Collaborative SQUid Size Monitoring (SQUISM), and the Squid-Squad, a highly interdisciplinary team of scientists, industry members, and managers with a common goal to improve squid science. Still, much of this cephalopod’s story remains hidden from sight. 

After years of trying to find and photograph a squid spawning aggregation, Ellenbogen chartered a vessel and worked with a network of local fishermen to locate concentrations of squid beneath the surface. In New England’s emerald-green, nutrient-rich coastal waters, visibility was limited, and although he knew he was directly above the spawning grounds, finding a cluster of eggs and squid 30 to 40 feet below the surface was far from easy.

“As I descended through the emerald-green waters of Cape Cod, I couldn’t see anything at first,” Ellenbogen recalls. “I knew I was in the right area, but the seafloor seemed empty. Then I noticed a few faint shadows moving in the distance.” 

As he swam closer, the scene slowly revealed itself. 

“First came the squid, then the egg masses, and then hundreds more animals appearing out of the green water. Suddenly I found myself surrounded by squid flashing colors, courting, competing, and spawning,” Ellenbogen says. “It felt like being in a theater, watching a carefully choreographed performance unfold all around me.”

Dozens of squid hovered intently in a circle with arms and tentacles stretching toward the focus of attention: a large pale cluster of eggs fixed to a bed of slipper shells and fingerlike seaweed. Thousands of gelatinous, translucent egg capsules swayed softly in the current. Male and female pairs broke from the outer circle with ceremonial precision, darting inward to spawn. Their alienlike bodies pulsed and flickered with waves of color, from flares of rust-red, golden yellow, and iridescent pink to flashes of lightning white. Their brief lives had culminated here in this critical moment of coordination to give rise to the next generation. 

Longfin squid live fast and die young, exhibiting incredible exponential growth throughout their lifespan of less than one year. They emerge from egg capsules as planktonic hatchlings — paralarvae — that already look like tiny 1.5 millimeter simplified versions of their adult form. 

Still small but fast-growing, juveniles feast on planktonic prey in coastal, surface waters and move deeper in the water column as they grow larger, settling on a life closer to the seafloor. As natural-born ambush and pursuit predators, they begin hunting small crustaceans and other invertebrates — including other squid — with jet propulsion and 10 grasping arms and tentacles.

Like octopuses, squid have chromatophore organs, pigmented cells controlled by nerves and muscles through their central nervous system. Contraction or dilation of these sacs results in mesmerizing iridescence and color shifts. Their color-changing behavior serves several purposes, including camouflage, courtship, and communication through ancient visual language. And to add an extra layer of curiosity, squid themselves are effectively colorblind. Their eyes are well-adapted to detect contrast, brightness, movement, and even polarization, but not to perceive hues like we do.

According to a report from the NEFSC, longfin squid can reach sexual maturity at a mantle length of just 8 centimeters. Nearing adulthood, they migrate offshore to overwinter in deeper, warmer waters along the continental shelf. At night, they form large schools grouped by body size and move upward in the water column to feed. Growth remains rapid and temperature-dependent, with males growing faster and larger than females.

Longfin squid spawn year-round with seasonal and geographic peaks. In New England waters, spawning has been reported from May to August. Egg clusters, or mops, like the one Ellenbogen photographed, act like a hub. As reported by NOAA Fisheries, female squid lay fertilized egg capsules that contain about 150 to 200 eggs each in clusters attached to the ocean bottom. They return repeatedly to deposit multiple clutches of eggs over several weeks, with a typical female laying a total of 3,000 to 6,000 eggs.

Spawning typically occurs in seasonal pulses tied to water temperature and other conditions, sometimes triggering large spawning aggregations — dense gatherings of squid depositing and fertilizing eggs. Males and females gather in the spawning grounds, and density increases until individuals are constantly interacting.

But reproductive biology and behavior are complicated for longfin squid. The NEFSC report highlights unique behavior: Females can store sperm from spawning events for later use, and eggs in the same capsule from a single female may have multiple males from multiple spawning events. 

Consort males, typically the larger males, pair up with a single female during spawning. They swim closely alongside her, often guarding her with their arms, and use their size to intimidate competition. Sneaker males, smaller unpaired squid, use stealth and speed to sneak into position with paired females. Males deposit bundles of spermatophores into the female’s mantle cavity or in a pouch located near her head. 

The competition is guided by visual and chemical signals, including pheromones indicating reproductive readiness, rapid color and pattern changes, as well as strategic arm postures and whole-body displays. A 1999 paper published by the Marine Biological Laboratory catalogued 34 visual displays involving colors, brightness, spots, stripes, iridescence, and polarization signals, as well as 17 specific positions and movements used by squid on the spawning grounds. 

Squid were seen flashing white as a signal to repel other squid. Squid displaying this component are almost always engaged in mate guarding, egg laying, or competitive confrontations. Like humans, female squid can blush, displaying a dark patch on one side of the mantle. But when a squid blushes, it functions as a repellent to courting males. And as captivating as these colorful displays can be, squid communication largely remains an encrypted secret to even the most curious scientists. 

“Keith’s rare glimpse of squid in their natural environment beautifully bridges art and science, reminding us that we still have much to learn,” says Hyde, who has centered her research at NOAA on marine ecosystems and fisheries. 

For most people, squid are known solely as seafood or as shadowy legends, chameleons of the sea. Yet each spring, beneath the cool waters of Cape Cod, millions of longfin squid gather in this remarkable reproductive event, a reminder of the rich biodiversity and productivity of Massachusetts marine ecosystems. 

MIT Sea Grant works to support sustainable fisheries and to promote environmental education and stewardship of our coastal and ocean resources. As MIT Sea Grant’s resident artist, Keith Ellenbogen is working on a long-term project to document extraordinary species and ocean events that unfold just off our coast, revealing hidden moments that few have the opportunity to witness. The longfin squid is one of our region’s most valuable marine resources, and these images help reveal one of the most extraordinary events unfolding right off our coast.

A brain must constantly cope with the highly variable, fast-paced nature of the world when trying to judge what’s going on around it. On one hand, it has to be open to whatever new sensory information may come its way, but on the other hand, just to keep up, it has to try to leverage prior experience to make predictions about what seems to be happening. 

In a new study published in Science, MIT neuroscientists identify a circuit that links a sensory decision-making region with one that advises it on how much sensory information just changed.

“This circuit organizes a comparison between what has just happened versus what is happening now in the sensory world in a manner that can be used to act,” says study senior author Mriganka Sur, Newton Professor in The Picower Institute for Learning and Memory and MIT’s Department of Brain and Cognitive Sciences.

Study lead author Ning Leow Phd ’23, a former graduate student in Sur’s lab who is now a postdoc at A*STAR in Singapore, says the study in mice sheds light on closely analogous circuitry in humans, in which an area of the prefrontal cortex (the anterior cingulate cortex, or ACC) makes sensory decisions. The new study shows it bases those decisions on advice about immediate past history from an area of the thalamus called the pulvinar (though in mice, it’s called the lateral posterior thalamus, or LP).

“The brain does not evaluate each new event from scratch,” Leow says. “The pulvinar has traditionally been studied for its role in attention and filtering visual information, but we found that it was also important for comparing present information with the immediate past and highlighting meaningful changes to influence whether we maintain or update a decision.”

As part of Sur’s long-standing interest in how the brain’s cortex integrates sensory perception and learning to produce behavior, Leow and Sur began comprehensively mapping the copious inputs to the LP-ACC circuit, culminating in a paper in 2022. It was clear from that study how the circuit would seem well-positioned to help focus attention, which is what it was known for at the time.

But in thinking more deeply about what focused attention is for, and about how these well-connected regions seemed to sit at the center of not only attention but also perception and action, Sur and Leow hypothesized that they might also have a hand in guiding decisions based on sensory information. The new study presents multiple lines of evidence that it does.

The findings not only shed light on a fundamental function of the brain, Sur says, but could also be applicable to studies of autism, in which many patients show significant differences in the predictions they make about the sensory world. Often, this manifests as difficulty filtering out stimuli that neurotypical people are able to regard as recurring, and therefore mundane.

Which way?

To conduct the study, the researchers trained lab mice to play a video game in which dots on a screen would drift around, but at least some would move together in the same direction (left or right). In each trial, the mice had to discern that trend. From one trial to the next, then, the sensory cue could vary not only by the direction of movement, but also by how what proportion of dots were participating. For instance, on one trial maybe 64 percent of the dots would move left and on the next trial maybe 16 percent of the dots would move right. In this way, the researchers could measure a whole continuum of differences from one trial to the next. 

Meanwhile, as mice played the game, the scientists used a two-photon microscope to record the activity of the LP-ACC circuit and the response of neurons in the ACC. In some experiments, they used a technique called optogenetics to artificially activate the circuit.

By tracking how mice performed the task trial after trial, the researchers were able to see that the mice indeed factored in not only what they were seeing in the moment, but also what they had just seen previously. For instance, when mice guessed right, they were very likely to repeat their guess if the new cue was very similar to the prior one, and very unlikely to if the cue was very different. But if they guessed wrong, then the opposite was true: They wouldn’t repeat that decision if the cue was similar to the last, but would if it looked very different.

Looking in the brain

Of course, behavioral observations only indicated that the mice indeed compared new cues to prior ones. Determining whether that was indeed because of the LP-ACC circuit required the researchers to use optogenetics to perturb it (by stimulating extra activity in the LP’s inputs into the ACC). For instance, optogenetic perturbation of the circuit in the left brain hemisphere made mice less likely to guess that dots were moving right, and perturbation in the right hemisphere made mice more likely to guess dots were moving to the right. But in both cases, the extent of these deviations from normal behavior was directly proportional to the difference between the current cue and the previous one. In other words, perturbing the circuit disrupted how mice used recent sensory history when evaluating new evidence, Leow says.

“That showed the pathway is causally involved in the comparison process that influences how current evidence is interpreted, rather than merely carrying the information,” Leow says.

Moreover, using the microscope imaging (which visualizes calcium levels in neurons, a close proxy of the electrical activity), the researchers extensively analyzed the activity patterns of the LP input into the ACC and how ACC neurons reacted to that input.

“The main takeaway is that the LP and ACC were performing different jobs,” Leow says. “The pulvinar doesn’t appear to be making the decision itself. Instead, it sends that history-referenced sensory comparison to the frontal cortex. The ACC then transforms that information into the neural activity that predicts the animal’s final choice.”

Essentially, the pulvinar advises the ACC on the degree of change so that the frontal cortex can consider whether it’s time to change a guess. After all, if a mouse is guessing right and little is changing, why not keep on trucking? But if there’s a big change, then it might make sense for the mouse to re-evaluate what it’s thinking.

It turns out, the brain has this dedicated circuit for doing so.

In addition to Leow and Sur, the paper’s other authors are Arundhati Natesan, Alexandria Barlowe, Sofie Ährlund-Richter, Tianyu (Cindy) Luo, and Mehrdad Jazayeri.

The National Institutes of Health, a MURI grant, the Simons Foundation Autism Research Initiative, A*STAR, and the Freedom Together Foundation funded the research.

Around 4.6 billion years ago, the solar system was little more than a giant ball of gas and dust. Over the next few million years, this “solar nebula” underwent a huge transformation, flattening into a disk of matter that then condensed to form the central sun and orbiting planets. 

Scientists have assumed that the early solar system was shaped mainly through gravity. But a new study finds that magnetism also likely played a role. 

MIT scientists have discovered records of ancient magnetism in the oldest samples of meteorites known today. The team analyzed microscopic grains embedded in a meteorite that was discovered in Antarctica in 2008. These grains, called calcium-aluminum-rich inclusions, or CAIs, originally formed during the solar system’s first 200,000 years, making the samples the oldest known solar system material.

The findings suggest that a magnetic field existed very early on, during the time of the solar nebula. The researchers estimate that this nebular magnetic field was stronger than Earth’s magnetic field today, and likely played a significant role in pulling together primordial matter to form the early sun.

“This transition, from a spherical cloud to a protoplanetary disk, is one of the most significant events in all of solar system history,” says Benjamin Weiss, the Robert R. Shrock Professor of Earth and Planetary Sciences at MIT. “It has long been theorized that gravity caused this, but our measurements show magnetism likely played a role.”

Weiss and his colleagues report their discovery in a paper appearing this week in the Proceedings of the National Academy of Sciences. The study’s MIT co-authors are first author Cauê Borlina PhD ’22, Elias Mansbach PhD ’24, and Nilanjan Chatterjee, along with Xue-Ning Bai of Tsinghua University, Po-Yen Tung and Richard Harrison of Cambridge University, François Tissot of Caltech, and Kevin McKeegan of the University of California at Los Angeles.

Spinning grains

Magnetic fields are generated by matter that is electrically charged and moving around. In the very early solar system, the collapsing cloud of gas and dust could have whipped up a plasma of charged particles. As these charges spun through the developing disk, they could have produced and sustained a magnetic field.

If this were the case, Weiss and his colleagues reasoned that such early magnetism would have affected material in the disk. As this material condensed, tiny magnetic minerals would have locked in the strength of the magnetic field, preserving its original intensity over billions of years. If these minerals somehow made it to Earth for scientists to measure, their “remanent magnetization” would be evidence that a magnetic field indeed existed and could have played a role in shaping the solar system. 

In fact, the team has previously discovered evidence of a magnetic field, as early as 2 million years into the solar system’s formation. At that time, scientists believe that the sun was already in place, and that the planets were just starting to come together. Thus, the magnetic field Weiss measured likely played a part in the formation of the early planets.

“Nowadays people don’t debate whether magnetism is present when planets are forming. But the debate is around the very early solar system, before planets are forming, when there’s just a disk,” says Borlina, who led the new study as an MIT graduate student and is now an assistant professor at Purdue University. “That’s where the debate still resides, and that’s where we’re operating now.”

Magnetic records

For their new study, the team investigated whether a magnetic field could have existed even earlier in the solar system, when the sun was first coming together. They analyzed samples of DOM 08006, a meteorite that was discovered in 2008 in Dominion Range, a mountain range located along the East Antarctic Ice Sheet. Since it was first recovered, the meteorite has been studied extensively. 

DOM 08006 is one of the most primitive meteorites discovered, and it contains mineral grains that date back to the earliest stages of solar system development, possibly even before the sun was formed. Surprisingly, the meteorite has managed to keep its original composition and minerals.

“Other meteorites went through many different processes over this 4.5 billion year history,” Weiss says. “They were formed in the solar nebula, then added to bodies with water, then got destroyed, moved to the asteroid belt, and then landed here. But somehow, DOM has experienced less alteration than any other meteorite.”

If the early solar system did harbor a magnetic field, records of that field could still be in place in some of DOM’s ancient mineral grains, including CAIs. 

“We know they are the oldest things we have of the early solar system,” Borlina says. “But CAI’s are very complex and are not all the same, even within a 1-millimeter piece of the meteorite. So we have to carefully identify what types they are.” 

From small samples of the parent meteorite, the team isolated tiny grains and identified a handful of CAIs that contained inherently magnetic minerals such as iron. They then put the grains through a series of tests to measure any magnetism they still carry. 

The team identified traces of a magnetic field in the ancient grains. Based on their measurements, they estimate that a magnetic field, of about 150 to 600 microteslas, existed in the early solar system. This field strength is about three to 12 times greater than the Earth’s magnetic field today. 

“We think these kinds of magnetic fields were helping to move gas from the protoplanetary disk, in toward this central star, the sun,” Borlina says. “Gravity is also playing a role. But we are now showing that, if you want to fully understand how the sun and planets formed, you should include magnetic fields in the ingredients that make them.”

This research was supported, in part, by NASA.

Can a city’s seawall stand up to a blockbuster storm? Will a region’s power grid hold against record-breaking heat? And can a town’s fire-fighting resources contain a major wildfire? 

To answer these questions, communities will first need to know how such extreme events could unfold. How far is a wildfire likely to spread? How much of a region might a storm impact? How long could a heat wave last? 

But extreme events are notoriously difficult to anticipate. By their nature, they are outliers. In the history of record keeping, extreme events are sporadic and rare. Yet most methods that assess a region’s risk depend on extreme events of the past to characterize even more extreme, worst-case scenarios in the future. 

Now, MIT engineers have developed a tool that generates plausible extreme events and worst-case scenarios, and maps their characteristics, such as an extreme storm’s likely duration, intensity, and area of impact. The key to their method is that it does not need to know about previous extreme events in order to generate plausible future extreme events.

Instead, the method, in the form of a machine-learning algorithm, learns from a dataset, such as a region’s daily weather records and maps. This record may or may not contain past extreme deviations, such as record-setting heat or rain. The team’s algorithm takes a statistical approach to learn from the available data, to exclude implausible weather scenarios. The method then generates plausible extreme events that are likely to occur in a region with a given frequency (such as once every 100 years), and projects how those extreme events might look in terms of their size, intensity, and duration.

“We are trying to model extreme, unprecedented events that no one has seen before, that are not in the dataset,” says Kai Chang, an MIT graduate student in mechanical engineering and affiliate of the MIT Center for Computational Science and Engineering. 

“An event like Hurricane Katrina is something that happens every 30 to 40 years,” adds Themis Sapsis, the William I. Koch Professor of Mechanical and Ocean Engineering at MIT, a core member of the Center for Computational Science and Engineering, and an affiliate of the MIT Institute for Data, Systems, and Society. “What will be the Katrina that happens every 100 years? How bad will it be? That’s exactly what we’re trying to quantify, to help planners prepare for plausible extreme scenarios.”

Beyond weather events, the approach, which the team has dubbed Extreme Event Aware, or “η-learning,” can be applied to other fields, such as robotic navigation and financial markets.

“Financial market crashes are extreme events that are a complicated combination of things, involving many different sectors,” Chang says. “What is the interaction that leads to a market crash? That is something that this method could explore.” 

Sapsis and Chang detail their new method in an open-access paper that appeared on Aug. 20 in the journal Nature Communications. 

“Riskier than everything”

To estimate a region’s risk of an extreme weather event, planners, policymakers, and insurance companies typically ask questions such as “What does a once-every-100-year storm look like for New York City?” For answers, they use computer simulations that must be trained on data that includes extreme, once-in-a-century events, in order to learn the conditions leading up to those events and generate scenarios of how those events might look in the future. 

“These methods assume there are very disastrous events that we have seen in the dataset, and they build a method to either estimate the risk of those events, or they try to predict exactly the events that have happened,” Chang says. “We are trying to see: What do unprecedented extreme events look like that are riskier than everything that has happened before and yet are still plausible?”

For example, if the most extreme rainfall measurement ever recorded in New York City is 200 millimeters, what kind of storm would produce an even more extreme measurement, of 300 millimeters? Such an event has never been recorded before and yet could still be plausible. City planners would want to know where such a storm would hit, how big an area it would cover, and how intense it would be. A simulation of the storm could help them assess infrastructure and plan reinforcements. 

“We want to predict maps of these worst-case scenarios,” Sapsis says. “There is no method that does this efficiently to predict events that happen rarely.”

Extreme learning

The team’s new algorithm generates plausible, unprecedented extreme scenarios, without needing to train on previous extreme event data. To do so, the algorithm combines and learns statistics, or probabilities, about the relationships between two types of data: point statistics and spatial maps.

To demonstrate, the researchers applied the method to generate maps of future extreme precipitation events over the continental United States. The researchers began with 25 years of hourly precipitation maps, which they pooled into daily maps. From the full record, they computed point statistics describing how often the maximum rainfall across a map reached a given level. They then trained the algorithm on paired low- and high-resolution spatial maps from just the first six months of the record, which contained few or no examples of the most extreme rainfall levels.

From these data, the algorithm learned how patterns in low-resolution maps correspond to detailed, high-resolution precipitation maps. It then used the point statistics to constrain the rainfall extremes represented in those maps. This combination enables the algorithm to generate plausible spatial patterns for events more extreme than those represented in the training data — for instance, the possible locations, sizes, and intensities of a once-in-a-century rainfall event with a maximum of 300 millimeters.

A user can prompt the trained algorithm with a question such as, “What could a once-in-a-century storm look like in New York City?” The algorithm then generates maps of statistically plausible storms that are likely to occur with that frequency, including characteristics such as the storm’s size, area of coverage, and intensity of rainfall.

“Someone can say, ‘I’m interested in building things to withstand the risk of an event that happens every 100 years,’” Chang says. “What we can do then is produce thousands of possible realizations that will happen with this sort of rare frequency.”

As long as relevant point statistics and spatial data are available, the method could be applied to visualize other unprecedented events such as extreme floods and wildfires.

“Extreme events have become a strategic concern, not just an environmental one — we’ve optimized global systems for efficiency, and the price of that efficiency is that there’s very little slack left anywhere. A single extreme event propagates through supply chains, energy markets, and food systems in weeks,” Sapsis says. “Being able to put a probability on an event that hasn’t happened yet is now a question of national and economic resilience.”

This research was supported, in part, by a Vannevar Bush Faculty Fellowship and the U.S. Air Force Office of Scientific Research. 

As people age, many cognitive functions tend to decline. Brain scanning studies have revealed corresponding changes in the function of a brain network that is involved in many of these cognitive functions, including working memory and problem-solving.

When it comes to language skills, however, the picture is different. Unless impaired by a stroke or dementia, most older people retain their language skills and may even improve them as they steadily gain vocabulary throughout their lives.

A new brain imaging study from an MIT-Boston University collaboration now reveals the neural activity underlying this observation. The researchers found that in older adults, activity of the language processing network is nearly identical to that seen in the brains of younger adults during language tasks.

In contrast, the researchers found that activation patterns in the multiple demand network, a brain system involved in executive control tasks such as decision-making, were very different in older and younger adults. 

“In the language network, we couldn’t find any differences between older and younger groups. In contrast, the executive system showed decline across almost all of the measures. The network synchronization declined in older adults, the extent of activation was reduced, and the magnitude of activation was reduced as well,” says Anne Billot, one of the lead authors of the new study, who carried out this work while doing her PhD at BU and is now a postdoc at Harvard University.

The findings suggest that parts of the brain that are specialized for specific functions, such as language processing, may be more resilient to aging than the multiple demand network, a more general-purpose network that has greater flexibility in its function, the researchers say.

Former MIT research assistant Niharika Jhingan is also a lead author of the study, which appears today in Nature Communications. Evelina Fedorenko, an MIT associate professor of brain and cognitive sciences and member of MIT’s McGovern Institute for Brain Research, and Swathi Kiran, the James and Cecilia Tse Ying Professor in Neurorehabilitation at BU, are the paper’s senior co-authors.

The resilience of language

To study the effects of aging on the brain, the researchers looked at two groups of people, ages 17-39 and 41-80. Based on previous studies, they expected that the multiple demand network, which includes several regions in the frontal and parietal lobes of the brain, would look different in the brains of older people. 

“It’s well known that executive functions, such as attention, working memory, and cognitive control, tend to decline with age. And it’s also known that in opposition to that, language skills typically tend to remain quite stable or even improve with age,” Billot says. “These two types of functions really go in opposite directions in healthy aging. In terms of behavior, that’s quite well-established, and we wanted to see if that was also the case in the brain.”

In previous studies of the multiple demand network, scientists have found that as people age network activity becomes less synchronized. Some neuroscientists have hypothesized that this may also happen in the language network, but studies haven’t found definitive evidence for this.

The MIT researchers were able to look at both networks by designing tasks that elicit responses primarily in either the language network or the multiple demand network. This allowed them to identify, for each participant, the brain areas that belong to each network.

During a spatial memory task — remembering the location of squares in a grid — the researchers confirmed that the multiple demand network showed altered activity in older adults. Compared to the younger subjects, their networks were smaller and less well-synchronized, and the overall activation level was weaker.

To identify the language network, the researchers had participants listen to stories and read sentences. They found that in both groups, brain activity in response to language showed similar levels and spatial distribution across the network.

They also found that younger and older subjects showed similar brain responses when they encountered an unfamiliar word or an unusual grammatical construction.

“We have previously used similar kinds of materials to show that young adults show strong sensitivity to these points of linguistic difficulty: activity in the language areas goes up. Here we found that in older adults, you also see this sensitivity, which suggests that there’s nothing fundamentally different about how they process language,” Fedorenko says. 

A language boost

The researchers also showed that in older people, the language network did not show any signs of becoming less synchronized. Additionally, the network did not show signs that it was blurring together with the multiple demand network, as some neuroscientists have hypothesized might happen.

While this study did not evaluate language ability, other studies have shown that not only do language skills not decline with age, for some people, their language processing improves in older age. This might be because vocabulary and reading skill can continually grow over time, the researchers say.

“Vocabulary keeps increasing as long as people have been measuring, which makes sense. People get exposed to more and more language, and older people sometimes start reading more, so they get an extra boost — it’s like a large language model trained on increasingly more data,” Fedorenko says. 

Given these findings, one possible generalization is that parts of the brain that are specialized for particular functions such as language may be less susceptible to age-related decline than the multiple demand network. That network is unique in its ability to give the human brain the flexibility to learn new skills and adapt to new situations. 

“The multiple demand network is a different system in the sense that it’s not accumulating knowledge over time. It’s more like a flexible resource that you can deploy in all sorts of ways. And somehow that’s the thing that is more vulnerable to aging,” Fedorenko says. “Why it’s so vulnerable — that is a very good question.”

The research was funded by the National Institute on Deafness and Other Communication Disorders, as well as MIT’s McGovern Institute, Simons Center for the Social Brain, Poitras Center for Psychiatric Disorders Research, and Quest for Intelligence.

Tuberculosis (TB) is a leading cause of infectious disease deaths, claiming over 1 million lives every year. It spreads through the air when an infected person coughs, sneezes, or exhales, and drug-resistant strains and asymptomatic spreading are growing concerns. Curbing TB transmission is an urgent public health challenge, yet scientists still don’t understand how airflow and other environmental factors influence that spread.

One problem is that studies of infectious disease transmission have focused mainly on population-level assessments or individual immune responses. But understanding how airflow and mixing influence transmission in indoor spaces requires expertise in fluid physics and computational modeling.

An interdisciplinary team including researchers at MIT and the University of Texas Southwestern Medical Center has now combined animal transmission experiments with quantitative particle tracking and flow modeling to understand how some lab-based environments can promote the spread of respiratory infectious diseases such as TB, while others mitigate that spread.

A key factor in predicting infectious transmission was not just the total ventilation rate but, more importantly, the local pattern of airflow driven by the design — such as air leakage, inflow and outflow locations, and forces created by an infected individual.

“The local airflow patterns turn out to be pivotal,” says Lydia Bourouiba, the Japan Steel Industry Chair Professor at MIT and faculty lead of the Fluid Dynamics of Disease Transmission Laboratory, part of the Fluids and Health Network within the Institute for Medical Engineering and Science (IMES). “Our team’s findings provide some of the clearest evidence I’m aware of showing the importance of accounting for [airflow] inhomogeneity and its effects when designing for airflow detailed patterns. This insight is critical when building or retrofitting an indoor space to mitigate airborne transmission, or when designing an airborne transmission study.”

The research is an important step toward connecting laboratory infectious disease studies with how people spread such diseases in the real world. The team hopes their insights can extend beyond their model system and show the importance of flow physics in building designs to prevent the spread of airborne diseases indoors.

“Despite recent pandemics and epidemics, there is still resistance to incorporating airflow in routine infectious disease prevention tools,” Bourouiba says. “Infrastructure could be retrofitted at relatively low cost, but the paucity and difficulty of gathering direct evidence prevents broader adoption of flow physics as a tool for indoor health. This study helps provide such evidence.”

Joining Bourouiba on a paper about the work are Yash Kulkarni, a postdoc at IMES, who led the fluid and aerosol physics components; Kubra Naqvi, lead author and a postdoc at UT Southwestern; Michael Shiloh, a professor at UT Southwestern, who led the multiyear effort to reestablish a classic tuberculosis transmission model; Hui Ouyang, an assistant professor of aerosol engineering at UT Dallas; Yuhui Guo, Deepak Sapkota, and Arabella Martin, all UT Southwestern PhD students; Pei Lu and Victoria Ektnitphong, research associates at UT Southwestern; Shibo Wang, a University of Minnesota researcher; Beatriz Dias, a UT Southwestern instructor; Bret Evers, an associate professor at UT Southwestern; and Lenette Lu, assistant professor at UT Southwestern.

Opening the black box

When people exhale, talk, cough, or sneeze, tiny microdroplets and bioaerosols launch from their mouths, carried forward by a cloud. If infected by a respiratory disease, these bioaerosols can contain pathogens that can infect others. Disease transmission depends on pathogen survival in the air, which is influenced by temperature, humidity, and ventilation.

In 1882, German physician and microbiologist Robert Koch first established an animal model for the study of tuberculosis pathogenesis. Decades later, researchers demonstrated airborne transmission of tuberculosis between people and animals.

These early experiments have proven difficult to replicate in today’s modern, biosafety-grade facilities. This new study reveals the difficulty comes from stringent containment and ventilation requirements, which can dramatically influence airflow in experiments.

“Host-to-host transmission is an obligatory evolutionary phase of respiratory pathogens, yet it has been considered too intractable or complex to be amenable to systematic investigation, hence is commonly relegated to a black box. Our work opens that black box,” says Bourouiba, who is professor in MIT’s departments of Mechanical and Civil and Environmental Engineering, and an IMES core faculty member.

To quantify how local airflow patterns impact infectious disease transmission, the researchers redesigned and modelled the early studies for modern high-containment lab facilities — including their seal, inflow, outflow, and exhaust pathways — and quantified particle and bacteria-laden particle release and dispersal. They released tracer particles and bacteria into a compartment and modelled recovery from air sampled on the other side under differing airflow rates, designs, and leak configurations.

The MIT team carried out computations, benchmarked against particle release experiments. The results revealed how important seemingly small details such as leakage paths could be.

“Even a small leak could short-circuit the airflow by drawing fresh air directly toward the exhaust, rather than drawing contaminated air across the containment chambers,” says Kulkarni. 

Advancing TB research

To date, uneven indoor airflow patterns have not been fully harnessed as part of a risk mitigation strategy. 

“By systematically defining how airflow and design influence biological exposure, we were ultimately able to restore transmission and create a system that can now be used to ask fundamental questions about the bacterial, host, and environmental factors that determine tuberculosis spread,” says Naqvi.

“I began working to reestablish this seminal TB animal transmission model nearly 10 years ago, and it proved far more challenging than I anticipated,” says Shiloh. “I hope this work serves as a reminder that meaningful scientific advances often require patience and perseverance.”

“This work illustrates how crucial it is to support synergistic collaborations integrating complementary disciplines to tackle research bottlenecks — and to standardize reporting norms across laboratories,” Bourouiba says. “If different labs have varying airflow patterns from uncontrolled leaks or seal details, that physical variability can overwhelm the biological signals researchers seek. Beyond its foundational impact for TB transmission studies, our work shows that opening the black box of transmission provides mechanistic insights: Detailed airflow pattern control can enhance or mitigate airborne transmission — making it exploitable as a prevention measure in crowded gathering spaces.”

This work was supported, in part, by the National Institutes of Health, the National Science Foundation, the Burroughs Wellcome Fund, MathWorks, and the Translational Research Institute for Space Health.

Ammonia is one of the most important chemicals produced in the world, ranking second only to sulfuric acid in the total volume produced each year. It is used mostly to make fertilizer, which is essential to feeding the world’s population. Yet its production accounts for up to 2 percent of the world’s energy consumption and about 1.5 percent of greenhouse gas emissions, so the search has been underway for ways to produce ammonia more sustainably.

The traditional way of making ammonia, in use for more than a century and accounting for the vast majority of production, is the Haber-Bosch process, which relies on fossil fuels to provide the needed heat. Hydrogen used in the process is also largely produced from fossil fuels.

There is another way, using electrochemistry instead of heat and pressure, but so far this method has not been anywhere near economically competitive at the scales needed.

Now, researchers at MIT have developed a way to predict which materials could be most promising as catalysts in electrochemical ammonia production. Catalysts help drive chemical reactions, and their properties determine how efficiently those reactions proceed. Rather than using trial and error to test each possible combination out of the millions of possible alloys — which can take years — the new approach could greatly speed up the search for materials that could make this low-emissions method competitive with the Haber-Bosch process. 

“Our approach identifies the key physical properties that drive catalytic activity in ammonia production,” says Bilge Yildiz, the Breen M. Kerr Professor in the departments of Nuclear Science and Engineering and Materials Science and Engineering (DMSE). The results can guide the search for new and more effective catalyst compounds.

The open-access findings were published Aug. 11 in the Royal Society of Chemistry journal EES Catalysis, in a paper by Yildiz and doctoral students Constantine Athanitis of DMSE and Filip Grajkowski of the Department of Chemistry. 

The challenge of greener ammonia

As the world’s population grows, Athanitis says, “we’re just going to need more and more food, and the only reason why we’re able to sustain so many people is because of fertilizer.” But more than 90 percent of the ammonia needed for fertilizer is still made by that energy-intensive Haber-Bosch process, which “has been hyper-optimized since it first came out more than a century ago,” he says.

“If we’re trying to keep in line with society’s sustainability and energy targets and climate change targets, we really need to come up with another alternative,” he explains. The world currently uses about 200 million metric tons of ammonia each year, “so ideally we want to be able to find a way to produce the same amount of ammonia, or even more, but in a more energy-efficient way and also with lower CO2 emissions,” he says.

Using electricity to produce ammonia is not a new idea. “It’s really just the electrochemical reaction between proton-electron pairs and nitrogen gas. And these technologies exist,” he says. The approach uses the same basic principles as electrolyzers, which use electricity to drive chemical reactions in devices.

But while the process works, it’s not efficient enough for industrial-scale production. “Production rates and yields are still too low,” Athanitis says. “Even though a technology might be better for the world or for the climate, companies and capitalism won’t really allow it unless it’s cost competitive.”

How to make it more competitive? The key ingredient in the electrochemical process is a metallic catalyst, whose properties govern the reaction that takes place on its surface. “If we can somehow find a catalyst that reduces the energy needed and is more selective for ammonia production,” Athanitis says, “then we could essentially hit the jackpot.” A more selective catalyst would produce more ammonia while reducing unwanted side reactions.

Finding better catalysts

But finding that ideal catalyst is not simply a matter of identifying one perfect material. Different materials can improve different parts of the reaction, and researchers are seeking combinations that can make ammonia production efficient, affordable, and practical at large scale.

“Metal nitride compounds make an ideal material system for this reaction and for identifying the electronic, chemical, and structural properties that determine reactivity in nitrogen reduction and ammonia electrosynthesis,” Yildiz says. 

Transition metals could form promising nitride alloys for this purpose, and historically, “materials research has been pretty much trial and error,” Athanitis says.

The usual process is to take some existing material and “tweak it in some way,” he says. “It’s all somewhat guided by scientific and chemical intuition.” 

Now, increasingly, computational tools are being used to model the physical interactions and predict outcomes. A method called density functional theory uses quantum mechanics to simulate the properties and behavior of materials, allowing researchers to predict how different atomic arrangements may perform before making them in the lab. Rather than searching randomly through every possible alloy combination, Yildiz says, “we first assessed what microscopic properties of the material make them tick for nitrogen reduction.” 

For ammonia-producing catalysts, “we’re looking at transition metal nitrides,” Athanitis says, because they have been found to be effective in these electrochemical nitrogen reactions. They are especially effective because “the nitrogen inherent to the catalyst itself becomes part of the reaction.”

This produces a series of chemical steps in which one step provides part of the energy needed to drive the next, reducing the amount of input energy needed. This helps solve one of the major bottlenecks in the nitrogen reduction reaction: the high energy required to break the strong bonds in nitrogen molecules, he says.

But the process is far from perfect, Athanitis says. It is “still limited by certain steps throughout the reaction pathway, including nitrogen dissociation and hydrogen transfer.” The study attempted to identify those bottlenecks and, with the help of machine learning, determine which alloys of these metals might overcome them.

With that understanding, “it can give us insights and open up potential strategies for how we can tune these materials to create next-generation better nitride catalysts,” Athanitis says.

Pushing past theory

The approach is “exciting work” that could help develop a foundation for designing new catalysts for ammonia production, says Dane Morgan, a professor of engineering at the University of Wisconsin who was not involved in this study.

“This work helps clarify how fundamental electronic properties of a material relate to its role as a catalyst in making ammonia,” Morgan says. “Such understanding can help guide researchers in designing new catalysts, both through better qualitative understanding and by accelerating computational screening.” 

So far, the study is purely theoretical: The researchers have used computer models to identify promising alloys, but those materials still need to be made and tested. Morgan notes that “translating these calculations into practical catalysts will require many additional steps, so meaningful real-world impact is likely still some distance away.”

The next step will be to build a working reaction cell, a laboratory device that uses the catalyst to produce ammonia and test its performance under real operating conditions. “For this to really make an impact in society, we need to bring it to the experimental lab,” Athanitis says.

“There have always been pushes at the frontiers of what’s possible,” he adds. “We like to think we’ve pushed the boundary of candidate materials here beyond what was thought of before, and hopefully we’re almost there. But even if we’re not almost there, we’re still pushing in the right direction.”

Anyone who’s ever wrestled with a claw machine at an arcade can appreciate the difficulty in pulling and pushing on joysticks, in just the right way, to get a mechanical arm to scoop up that one special toy. Coordinating the joysticks and connecting their movement to the claw’s motion is a type of “mental mapping” that is not immediately intuitive. (And it’s what arcade owners depend on to bring players back, again and again). 

In fact, the mechanics of a claw machine are broadly similar to driving an excavator: An operator uses joysticks to control the digger’s boom, arm, and bucket, and the direction of its cab. But an excavator’s maneuvers are far more complex than anything an arcade claw can do. Operators must learn more complicated mental maps to direct a digger to move rocks, grade soil, clear debris, and dig foundations, among other essential on-site jobs. Indeed, it can often take years for operators to build up expertise in maneuvering the heavy machines.

MIT engineers are looking to shorten the learning curve for excavator operators with a new training interface. Instead of using joysticks, the team has designed a more intuitive controller, which itself resembles a miniature excavator’s arm and bucket. Trainees grasp the device and use their arm and hand to make it move the way an excavator does. A digital excavator projected on an immersive six-screen display mirrors the trainee’s movements in a virtual environment.

“This is a more intuitive way to command the machine,” says Hermano Krebs, principal research scientist in MIT’s Department of Mechanical Engineering. “With this new interface, we can eliminate a lot of the mental maps that an operator would need to build in order to operate an excavator.” 

Krebs sees the interface as a faster way to train excavator operators, as well as a new way to physically operate the machines, both on-site and remotely. 

“Instead of having joysticks, you might have this miniature arm on the side, where the operator would place their own arm, kind of like an exoskeleton, which would allow them to operate the excavator in the cab,” Krebs says. “If work has to be done in a difficult or unsafe environment, you could have an operator sitting off-site in a trailer and using this arm to remotely tele-operate the excavator.”

The team reports its open-access results this week in the Journal of Computing and Civil Engineering. MIT co-authors include Moises Alencastre-Miranda, Joao Buzzatto, and Eran Beeri Bamani, along with collaborators from Sumitomo Heavy Industries, an industrial machinery manufacturer based in Japan. 

A machine mimic

At MIT, Krebs’ group works on human-robot interactions, with a longtime focus on physical rehabilitation. Through this work, the team has accumulated knowledge about the ways in which humans control their limbs and how they can most intuitively interact with machines. 

In 2018, Krebs struck up a collaboration with researchers at Sumitomo Heavy Industries, who were looking for a faster way to train excavator operators. They noted that in Japan, the population of heavy machinery operators is aging rapidly; training their replacements takes time. 

Operators typically learn by driving actual excavators on a controlled driving course. As they operate the machine, novices must learn to relate the actions of the excavator’s joysticks with the movements of the arm, bucket, and cab. Coordinating these actions to carry out actual tasks adds another level of complexity that can take months to years to master. 

The team reasoned that if they could eliminate the need for this mental map, they might significantly shorten the training process. To do so, they looked for a more natural way to control the machine, as an alternative to the traditional joysticks. They soon landed on the mechanical arm design, reasoning that the physical resemblance to the digger’s own arm and bucket could enable operators to mimic and control the excavator’s movements directly, without much mental translation. 

Over the next few years, the researchers worked to build the mechanical arm, along with the software to pair its movements with a virtual simulation of an excavator. The combination of the mechanical arm and the virtual simulator constitutes a new training and control platform for digger operators, which the team has named the “World-Space Interface.”

“‘World-space’ refers to everything in the world that is outside of yourself, or in this case, outside of the excavator’s cab,” Krebs explains. “Normally, operators have to build a mental map of how to manipulate things in the world-space. But now, we can just mime picking up rocks or dirt, and the computer will do that translation to the world-space for us.”

Construction on day one

For their new study, the team ran training experiments with volunteers who used the World-Space Interface (WSI) as well as a more traditional, joystick-based excavator simulator. The researchers developed virtual simulations of 15 realistic excavation environments, including construction sites, highways, forest roads, riverbanks, mining areas, and urban and rural settings. Each virtual environment was associated with various excavation tasks, such as scooping and dumping sand or gravel, digging and grading trenches, clearing debris from roads, removing tree branches from water edges, and breaking up rocks. 

The team designed the experiment to resemble the tasks that an operator typically performs during a weeklong excavator driving course. For one hour each day for seven days, volunteers — both expert and novice — operated the WSI and the joystick counterpart, training on tasks with increasing difficulty. 

The researchers then compared the volunteers’ performance before and after the training period. For the joystick simulator, they found that novices were consistently worse than experts, though they did improve over the training period. In comparison, the team found that with the new World-Space Interface, novices were just as good as experts from the start. 

“In this case, joysticks are a non-intuitive way to control and coordinate the machine,” says study co-author and MIT postdoc Joao Buzzatto. “This is the first interface that does not require me to command the excavator with joysticks.”

The team is now working to add haptics, or feeling to the WSI’s physical arm. The idea is that, as an operator uses the arm to mime an action such as picking up a pile of rocks, the arm will generate a force in response, as if the operator can feel the heaviness of the rocks, as confirmation that the excavator is indeed picking them up. 

“Haptics would make this an even more intuitive system,” says co-author and visiting engineer Solmon Jeong.

The team says the new training interface can be a more natural alternative to excavator simulators that the construction industry is currently exploring. Companies such as Caterpillar, Hyundai, and Komatsu are developing virtual simulators, both to help train operators before they go on-site, and to one day remotely control excavators from a distance. However, these simulators are largely based on traditional joystick controllers that still take time to learn. 

If the team’s new arm-and-bucket controller were incorporated, as an appendage in an excavator cab, or in a virtual, teleoperational simulator, the researchers envision that even first-time operators could get to work, from day one. 

This research was supported, in part, by Sumitomo Heavy Industries.

“I think that a game can be made about anything,” says Kayode Dada, a rising senior majoring in mechanical engineering. “But to make a good game,” he adds, “it takes a lot of time and effort and thinking about it.”

Dada has been thinking about games seriously since he was a teenager, but he took an interest in them much earlier. Growing up in Louisville, Kentucky, he and his brother would play card games, board games, and video games for hours. His face lights up as he recalls some of his favorites, including “Dungeons and Dragons,” “Catan,” “Yu-Gi-Oh!,” and “Risk.” During the Covid-19 pandemic, his family developed a nightly ritual. “For weeks and weeks, every single night we’d play a full game of ‘Monopoly.’ It was crazy!” he recalls.

Even as a kid, Dada wondered about the mechanics and decisions that went into the games he played, but it wasn’t until high school that he tried his hand at game design. He created his first game by constructing dice with different sides using arts and crafts materials. From there, the ideas kept coming. He figures during those years he designed about 25 games; he made nine of them and gave them to friends for their birthdays. 

Much of his application to MIT revolved around board game design, too. At first, he didn’t believe MIT would accept him because he didn’t “do any real tech stuff,” Dada says. But reading the MIT Admissions blogs changed that — particularly a post by a computer science major who wrote that she had no coding experience before coming to MIT. 

“That gave me motivation and hope that I could apply with a bunch of cardboard games, and I got in,” he says. “It was really cool to see, through the eyes of current students, that they are not all super-duper geniuses in an unattainable sense. They’re real.”

Now, Dada has created his biggest and most ambitious game yet: a trading-card game for 1,100 incoming first-year students during MIT Orientation in late August.

“Building up the next class”

Since he’s been at MIT, Dada has thrown himself into in a number of extracurricular activities and had less time to devote to game design; he’s now an MIT Admissions blogger himself, an orientation captain, a leader in MIT Cru (a campus Christian group), a drummer in the MIT Live student music group, and a DJ on WMBR 88.1 FM (his show, “Midwest Pizzeria,” features Midwest emo math rock).

For Dada, blogging and serving as an orientation captain are meaningful ways that he can help build community and increase awareness about the plethora of opportunities for students. He recalls one post he wrote about being a Christian at MIT that really struck a chord with readers. 

“I got a lot of emails from either students interested in the club or parents and high school counselors saying, ‘Thank you for posting this — I had no idea,’” he says. He also hears from a lot of students who don’t think they have what it takes to get into MIT. “I tell them, ‘I think you should still apply … and regardless if you get into MIT or not, you’re going to do amazing things.’”

This summer will be his third MIT Orientation. He loves it. 

“It’s a lot of fun to talk to students, hear what they are interested in, and then be like, ‘Oh, you should definitely go check out this thing. I’ll connect you to this person that you should talk to,” he says. 

In a fun twist of fate, some of the orientation leaders he works with read his blogs as prospective students. And he has met some of his best friends through orientation: “[They] are as interested as I am in helping out the people coming in … giving it back to MIT and contributing towards building up the next class.” 

Re-orienting orientation 

Last year, Dada’s interest in game design intersected with his role at MIT Orientation. The staff had been brainstorming games to try to keep the new students engaged between events, and Dada offered to create a trading-card game that would also serve as an ice-breaker. His game, “Campus Trade,” capitalized on the fact that the students are divided into 12 groups designated by a different color. Each student would start with 12 trading cards the color of their group, and by the end of the week they had to collect one card from every other group — in other words, 12 different colors.

The game was a hit. “They got so invested immediately, and we couldn’t get them to stop playing and go back inside to the next meeting,” Dada says. 

Nonetheless, he made some key observations watching them play — including the fact that many of the students would just trade cards and walk away, without talking much.

He decided to redesign “Campus Trade” and connected with Cassandra Lee, a research designer in the MIT Media Lab’s Center for Constructive Communication, whose interests include the connection between games and social interaction. Under her supervision, he spent his junior year doing an Undergraduate Research Opportunities Program project through the realtalk@MIT program, which supports students who design and organize their own creative community infrastructure. 

After months of prototyping, interviewing students, testing, and iterating, Dada is ready to roll out the game. This year’s “Campus Trade” cards are bigger — tarot-sized — and feature current MIT students’ artwork on one side, depicting a person, place, or thing from campus. On the other side there’s an icebreaker question. The game is designed in such a way that there are multiple ways to win, to encourage collaboration.

Working on “Campus Trade” with Dada was a “sincerely rewarding collaboration,” Lee says. “Community development work really requires a student who understands and is willing to fight for the value of people, stories, and being in community. Consistently throughout this project, it was clear that Kayode is more than capable of balancing his own spirit with the grueling organizing work that makes a project like ‘Campus Trade’ possible.”

Ultimately, Dada hopes to pursue a career in game design after he graduates. He thinks about games a lot, lying awake at night, or even in church. 

“Before, I didn’t really believe I could do it,” he says, “but now that I have this project under my belt, some real experience working on a game on a large scale, I definitely have a lot of ideas.” 

He figures he’ll either apply for a job at a game design company like Wizards of the Coast, or maybe even start his own company and “just see how it goes.”

And why not? After all, he took a chance when he decided to apply to MIT — and that turned out to be exactly the right move.

Toxoplasma gondii, or Toxoplasma, is a parasite that infects hundreds of millions of people around the world. Although cases are often mild, it can cause severe symptoms in in people with weakened immune systems, and in developing fetuses. It can also persist for years by forming long-lived cysts in tissues, allowing infection to become chronic.

During chronic infection, hundreds of Toxoplasma parasites can pack into a tissue cyst inside a brain or muscle cell. That crowded life carries a cost: Nutrients become harder to obtain, waste accumulates, and energy-producing reactions can become damaging. 

How Toxoplasma reshapes its metabolism to keep growing under such strained conditions has been unclear. But a new study from the lab of MIT Associate Professor Sebastian Lourido, a member of the Whitehead Institute for Biomedical Research, identifies a parasite-specific protein that helps coordinate this response. The protein, named TgPRO, allows Toxoplasma to manage oxidative stress — the buildup of reactive oxygen molecules that can damage cells — by controlling genes involved in energy production and iron use.

The open-access findings, published on Aug. 11 in the journal Cell, reveal the first dedicated regulator of metabolic gene expression identified in apicomplexans, the group of parasites that includes Toxoplasma and the organisms that cause malaria. The study, led by co-first authors and Lourido lab affiliates Christopher Giuliano PhD ’26, a recent graduate student in biology, and Chinmay Kalluraya, a current graduate student in biology, reveals a previously unknown way that parasites regulate metabolism. The findings also point to a possible therapeutic strategy: Inhibiting pathways controlled by TgPRO could make Toxoplasma more vulnerable to antiparasitic drugs that induce oxidative stress, though this approach remains to be tested.

One gene at a time

To discover the genes that support Toxoplasma’s ability to live in crowded cells, the researchers used a genome-wide CRISPR screen to compare Toxoplasma growing at low and high densities. The screen tests the effects of turning off genes one by one at both population densities in order to determine which genes are essential specifically in crowded conditions. It highlighted pathways that make or recycle NAD and NADP, molecules important for energy production and defending against oxidative damage. It also pointed to TgPRO, a previously unstudied protein that was especially important when parasites became crowded.

“A genome-wide screen was a powerful way to ask how crowding affects parasite fitness,” Kalluraya says. “TgPRO emerged as very important at high density. Because almost nothing was known about it, we wanted to understand what it was doing.”

Parasites lacking functional TgPRO accumulated more reactive oxygen molecules and struggled to compete at high density. Experiments showed that the loss of TgPRO disrupted the mitochondrion — the structure that supplies much of a cell’s energy — and changed how parasites processed glucose and other nutrients. Providing additional iron or restoring an important chemical balance inside the mitochondrion improved parasite growth, connecting TgPRO’s effects to iron-dependent energy metabolism.

The team then traced the response to a molecular mechanism. TgPRO is an RNA-binding protein, meaning it attaches to the molecular messages (RNAs) that cells use to make proteins. The researchers found that it binds and stabilizes a select set of messages involved in nutrient use, mitochondrial activity, and the assembly of iron-sulfur clusters, small structures that many enzymes need to function. The experiments connected the original observation — that some parasites faltered only when crowded — to a precise interaction between a regulatory protein and its RNA targets.

“One of the really nice elements of the story is our ability to connect it all the way through — from the original observation and genome-wide screen to the metabolic consequences and the direct interaction between TgPRO and its target RNAs,” Lourido says.

The researchers found that lowering oxygen levels also reduced oxidative stress and partially restored the growth of parasites without TgPRO. Toxoplasma is commonly grown in laboratories at atmospheric oxygen levels, which are considerably higher than those found in most animal tissues. The result suggests that oxygen conditions can strongly shape parasite metabolism, and the researchers caution others studying Toxoplasma to take this into consideration.

Connecting TgPRO to chronic infection

After testing the role of TgPRO in artificially crowded settings, the team also tested whether TgPRO matters during chronic infection, when Toxoplasma forms cysts in the brain. Mice infected with parasites lacking functional TgPRO developed smaller brain cysts, suggesting TgPRO supports parasite growth in the naturally dense environment of a chronic-stage cyst.

“The chronic stage is still somewhat elusive,” Giuliano says. “Showing that TgPRO affects cyst growth suggests that these same metabolic changes are needed in the brain and gives us clues about how the parasites persist there for months or years.”

TgPRO bears little resemblance to the proteins that regulate similar metabolic programs in mammals, yeast, and bacteria, yet it controls many of the same kinds of genes that these organisms adjust when cells face oxidative stress or changing nutrient conditions. This is an example of convergent evolution: Distantly related organisms evolved different molecular machinery to solve a similar biological problem.

That convergence suggests that coordinating these metabolic pathways may be a fundamental requirement for cells adapting to stress.

Altogether, the study establishes a new paradigm for how apicomplexan parasites regulate their metabolism, and advances the foundation for investigating how Toxoplasma persists inside its hosts.

This work was supported by National Institutes of Health grants and by a Burroughs Wellcome Fund grant awarded to S.L. M.A.S is funded by an Early Career Award from the Wellcome Trust. C.R.H. is funded by a Sir Henry Dale Fellowship from the Wellcome Trust and the Royal Society. J.K. is supported through funding by a generous donor advised by CARIGEST SA and acquired by D.S.-F.

Creating innovations that can succeed in both emerging markets and in developed, higher-income markets requires a targeted, thoughtful design process. In a new book, Amos Winter, the Germeshausen Professor of Mechanical Engineering at MIT and a pioneer in engineering design, and coauthor Vijay Govindarajan, a leading voice in strategy and innovation, present a framework for creating innovations that scale across borders, industries, and income levels.

“Emerging markets are growing at about twice the rate as the U.S. and Europe,” Winter says. “If you’re a multinational company that really wants to maximize your growth rate, it’s not going to be in wealthy markets, it has to be in emerging markets.”

He adds that this is particularly true for legacy companies that have successfully sold products in wealthy markets. 

“You can’t just simply take what you sold in the U.S. and Europe and then try to sell it in a place like India; it’s probably going to be too expensive [or not meet consumer needs],” he says. “We offer guidance on how to leverage what was done before without trying to just copy and paste.”

In their book, “Global by Design: How to Create Innovations That Scale, Travel, and Transform” (Harvard Business Review Press), Winter and Govindarajan present three phases, aiming to guide readers though stages of the innovation process, from identifying opportunities to designing solutions to scaling locally and globally. The authors move beyond theory and provide a practical, disciplined approach to spotting problems, leveraging innovation, and building high-value, low-cost solutions with global appeal.

“The book is aimed at how to approach solving problems that are historically unsolved,” explains Winter. “It talks about how to look at those problems with fresh eyes, how to really distill what are the unique requirements that must be satisfied, which often differ from the requirements in wealthy countries, and figure out how to create not just low-cost solutions, but really high-value solutions.” 

One class of problems Winter points out are issues in the developing world, where people face needs for water, health care, and energy. 

“Even though we have solutions to these problems in wealthy countries, they just haven’t mapped over, often because the solutions are too expensive or they’re not robust enough, leading these problems to persist for generations unsolved,” he says.  

Beyond emerging markets, globally minded solutions can also resonate with more resource-constrained consumers and create upsell opportunities for wealthy consumers. The book also gives guidance about how to not cannibalize existing efforts with product differentiation in different segments.

The authors use five case studies from Winter’s research group, along with industry examples. One case study discusses a product going into production that may reshape agricultural practices in major markets and varying climates throughout the world. 

“We created new drip irrigation emitter technology that cuts pumping power in half, reduces the price of solar-powered irrigation by about 40 percent, requires less than half the plastic to make, and is an industry leader in clog prevention,” Winter explains. 

The emitters will enable the growth of water-saving, renewable-powered irrigation in low-resource, water-stressed markets and offer a valuable alternative to farmers in wealthy countries. The technology is currently being commercialized by Toro and is expected to go on sale in early 2027. 

“It is truly a global product that is a better mousetrap, as it meets or exceeds the performance of any other product on the market in every category,” he says. 

The research behind the drip emitter technology was recently presented in separate journal articles in Nature Communications and Water.

Winter is also director of the K. Lisa Yang Global Engineering and Research (GEAR) Center, where research focuses on solving technical challenges in low-resource communities. Some of his team’s other notable solutions have been in water purification, agricultural equipment, and assistive technology. 

He says he hopes this book broadens design thinking and provides tool sets for innovators to solve meaningful, multifaceted, global problems. 

“We want readers to walk away from 'Global by Design' inspired to create solutions that are financially viable and create positive social impact. Furthermore, we want them to feel empowered, seeing how they can apply their unique experience, perspectives, knowledge, and resources to make meaningful change in the world.”

It turns out the best way to teach students how DNA works may be to start at the end. 

That’s the idea behind a hands-on curriculum from the MIT Edgerton Center: Teach proteins first so that students understand what the DNA will be making, then start at the top. 

Manipulable molecular biology sets developed over some 25 years by Kathy Vandiver, Edgerton’s life science project leader as well as director of community outreach education and engagement in the MIT Department of Biological Engineering, have scaled all the way from a gallery-classroom at the MIT Museum to every biology classroom in Worcester Public Schools in Massachusetts.

From one classroom to an entire district

Worcester is the second-largest district in the commonwealth, and its connection to the MIT Edgerton Center traces back to one high school teacher’s transformative learning experience. David Mangus first encountered the hands-on molecular biology sets at a Massachusetts Association of Science Teachers (MAST) conference, where he participated in an MIT workshop and explored the materials firsthand. Later, when he became science curriculum specialist for Worcester Public Schools, he recognized their potential to advance high-quality science education across the district, and he was in a position to implement the curriculum at scale.

Thanks to a grant from the Massachusetts Life Sciences Center, Worcester has been able to procure 55 class sets of MIT-developed DNA, RNA, protein, and tRNA modeling sets, along with instructional booklets and a districtwide professional development program. Every biology teacher participated in a thorough training workshop, so they could experience the joy of hands-on learning themselves and feel fully confident delivering the material in their classrooms.

The commitment in Worcester to training and equipping all of its biology teachers solved the problem that had previously limited the curriculum’s reach. Earlier outreach efforts with other school districts had shown the promise of the sets, but couldn’t scale because teachers were being trained one at a time. Worcester is rewriting the playbook with a cohort of teachers ready to use this hands-on learning approach districtwide, year after year.

The big idea: Teach proteins first

Vandiver is determined to reach more classrooms with her hard-earned insight: flip the script, and teach proteins before DNA.

Every biology student learns the central dogma of biology: DNA makes RNA, RNA makes proteins. But according to Vandiver, who holds a PhD in cellular biology, it’s also one of the most consistently misunderstood sequences in secondary education. It’s not a problem of students’ aptitude, but rather the order in which concepts are introduced. Vandiver likens walking a class through molecular biology without first explaining what a protein is to handing them blueprints for a building they’ve never seen. Most students memorize the steps without ever grasping what proteins are, or why they’re important.

Vandiver’s approach reverses the sequence. Working in pairs, students begin by building and understanding protein structure. They snap together what proteins are made of, learning what they look like and how they function. Only after creating proteins do they revisit the DNA-to-RNA-to-protein pathway, using genes that code for the very protein structures they built earlier. Then, concepts like amino acid order and protein function click into place. Students recognize the endpoint and experience a real “aha” moment, realizing, “Here’s our old friend, the channel protein!”

A growing body of research indicates that active learning experiences are more memorable to students. What the hands build, the mind retains. This curriculum represents a pivot from memorization toward constructing understanding, and toward learning experiences that are more memorable and enjoyable. In other words, these sets create learning that sticks.

Proteins’ Cinderella story

If DNA and its instantly recognizable double helix is a cultural icon, proteins are the stagehands behind the scenes. They build tissue, transport oxygen, fight infection, and carry out nearly everything the body does. Still, in most classrooms, proteins are only introduced after students memorize information about DNA and RNA. Vandiver’s curriculum is a Cinderella story for proteins. It takes the molecule that does all the work but rarely gets the spotlight, and finally makes it the star.

Beyond the fact that proteins come third chronologically in the DNA-to-RNA-to-protein sequence, there was also a practical reason they had been left in the shadows. Walk into any biology classroom, and you’ll find a model of DNA. What you won’t find is a good model of proteins. If there is a good protein model, it’s almost certainly not one that students can manipulate, take apart, and rebuild. That gap is exactly what Vandiver set out to fix.

Over 25 years, beginning when she was still a middle school biology teacher, she prototyped, tested, and ultimately redesigned DNA models from scratch. To create the novel protein models, the pieces that until then simply did not exist, Vandiver collaborated with her husband, Professor J. Kim Vandiver, Forbes Director of the MIT Edgerton Center.

The design of the pieces themselves is part of the learning. Color coding activates schemas of understanding that students already carry. For instance, yellow subunits are hydrophobic, evoking substances like oil or salad dressing. The pieces are tactile and modular, snapping together with a satisfying “click” that reinforces the logic that students will remember. 

Roots at the MIT Museum, branches across America

This approach to biology was refined over years at the MIT Museum. In 2005, Vandiver worked alongside museum staff to create a public space where visitors could learn how cells work, and which could double as a classroom for teaching those same ideas. The result was a gallery, “Learning Lab: The Cell,” funded by the Arthur Vining Davis Foundation, which supported both the space and the workshops held there.

It was around this time that Amanda Gruhl Mayer ’99, PhD ’08, who had specialized in genetic toxicology, joined the effort, collaborating on the workshops, and leading the graphic design of supplemental learning materials. Now, she contributes to the research behind the curriculum’s newer, more advanced lesson plans.

The space was a hit. Together, Vandiver and Gruhl Mayer taught more than 1,000 students over several years, welcoming them by the busload from across Massachusetts and beyond. The audience spanned middle and high school and included AP biology classes. But it didn’t stop there. Teams of nurses, biotechnology company executives, and even a class of federal judges came through as well. Judges, as it happens, need to understand DNA for forensic reasons.

That wide range of learners taught the team all about how to best convey this information to broad audiences. They discovered that the curriculum works best when students work in pairs with a molecular build set. Vandiver also created a participatory demonstration in which students take on roles and act out some of the more complex cellular processes with their own hands. They animate what the models are doing while the teacher acts as coach and explainer. Afterward, the group discusses the process together, layering in scientific language. Vandiver emphasizes how these resources free the teacher “to uncouple the overwhelming biology vocabulary from the conceptual story and provide an overview of the process.”

More than 1,000 students, one busload at a time, is a success story, but also hints at a constraint. The learning was impactful and memorable, but the outreach couldn’t grow fast enough. That’s what makes Worcester, and its entire district of teachers trained at once, such a significant step.

A Teacher of the Year reacts

The MIT team has traveled across the country delivering training on these materials, including a recent trip to Rapid City, South Dakota. In attendance was the district’s Teacher of the Year for 2025, Ross Hunter. He was the kind of participant whose enthusiasm drew in those around him, and his reflection captured a larger trend across education:

“This MIT hands-on solution does an amazing job of providing an understanding of protein synthesis … An additional value of these models is that they are not technology-based. In the next several years in education, we are certainly going to see the pendulum swing back toward a classroom with less technology … The tide is certainly turning, and MIT’s DNA and protein modeling kits can help.”

It’s a philosophy of learning older than any screen, and one Vandiver likes to express with the words of Confucius: “I hear and I forget. I see and I remember. I do and I understand.” 

What's next: Going national

This philosophy, as applied to molecular biology, is about to reach its widest audience yet. This October, the Edgerton Center will feature its molecular biology sets at the National Association of Biology Teachers Conference in Dallas. One of the premier gatherings for biology educators in the country, it’s a chance to put the proteins-first approach directly into the hands of teachers who could use it most.

In the end, the curriculum’s strength comes down to three key reversals. It makes proteins, not DNA, the star of the story: a Cinderella molecule finally stepping into the light. It trades screens and memorization for models students can hold, because hands-on learning sticks. And it flips the sequence of teaching itself, starting at the finish line so that the journey there makes sense. Teach it in that order, put it in their hands, and molecular biology transforms from something that students memorize into something that they understand.

MIT researchers have overcome a major challenge holding back the real-world deployment of microwave quantum technologies for advanced signal processing and secure communications.

The team developed a scalable platform that generates pairs of highly correlated radio frequency waves, without the need for bulky and expensive cooling equipment. In quantum technologies, these linked radio waves can be used for noise-resilient communication or high-precision radar and sensing. However, they’re usually only generated in research labs, under extremely cold conditions.

The MIT researchers fabricated a small, electronic device that can generate the same type of highly correlated signals at room temperature. 

The device incorporates a magnetic film, which interacts with microwave energy inside a metal cavity to split an incoming signal into two linked output signals. The researchers used the device to demonstrate secure communications by encoding information in a signal that could only be recovered using its partner signal.

“We’ve shown how the quantum properties of magnets can be leveraged to realize new communication and detection technologies. I hope our demonstration of this platform will enable further development of room-temperature quantum simulators, which have huge potential to enable many future discoveries,” says Qiuyuan Wang, an electrical engineering and computer science (EECS) graduate student and lead author of a paper on this technique.

Wang is joined on the paper by Aravind Karthigeyan, a graduate student at the University of Illinois at Urbana-Champaign; Chung-Tao Chou, an MIT postdoc; and senior author Luqiao Liu, an associate professor in EECS and a member of the Research Laboratory of Electronics. The research appears today in Nature Electronics.

Synchronized signals

Microwave photons are fundamental particles that form the signals used for wireless communication and sensing. 

Scientists can split one microwave photon into two tightly correlated photons using a device called a Josephson junction, which is an element of a superconducting circuit. These linked microwave photons can be used in applications like secure communications or high-performance radar systems that can detect extremely faint signals.

To enable secure communications using these correlated signals, engineers could design electronic devices that encode data in one signal by altering the signal’s properties, such that the information could only be decoded at the other end of the transmission using the matching signal. But to operate effectively, superconducting circuits must be kept at temperatures below 273 degrees Celsius, usually inside a bulky, expensive, and energy-intensive cryostat machine.

While pursuing a different line of research, the scientists in Liu’s group realized they could generate the same highly correlated microwave signals using magnets instead of cryogenically cooled superconducting circuits.

By putting a magnetic film into a microwave resonator, which is a metal cavity that traps electromagnetic energy, they could split one incoming microwave photon into a pair of perfectly synchronized signals with distinct frequencies, at room temperature.

“On its own, each signal looks random, but their phase relationship remains strongly correlated,” Wang explains.

Their device relies on magnons, which are tiny packets of magnetic energy. Typically, pumping microwave photons into a magnetic system generates a pair of correlated magnons with the same frequency. 

Even though both magnons are correlated, because they have the same frequency, scientists can’t separate them. They would need to separate the magnons to use one signal for transmission and the other for detection in secure communications.

A hybrid system

By coupling a magnetic film with a microwave resonator and carefully controlling the energy they pump into the device, the researchers could form hybrid magnon-photon waves. These hybrid waves output a pair of synchronized signals with distinct microwave frequencies.

The signals remain strongly correlated, but since the frequencies are always different and random, an attacker can’t recover the information encoded in one signal without having the matching one to use as a key.

The researchers demonstrated this by encoding a small image in the frequency of one microwave signal. They successfully decoded the signal and extracted the image using its partner.

“Magnonic systems exhibit a remarkably rich range of nonlinear dynamics, but these nonlinearities have not yet been harnessed for practical applications as extensively as those in nonlinear optics and other dynamical systems. In this work, we address one important challenge: the spectral overlap between a pair of ‘twin’ magnons generated by the same pump photon. By using the level repulsion arising from coupling between magnons and microwave photons, we were able to separate the two magnons in frequency,” says Liu. “We believe this demonstration could provide a foundation for technologies such as quantum radar, secure communications, and quantum-limited sensing, all of which rely on correlated — and ultimately entangled — microwave sources.”

This hybrid magnon-microwave system could also be used in noise-resilient communication by enabling the receiver to decode a message that has been garbled by random data that interfere with the transmission.

Correlated microwave signals are also a key element of a quantum simulator, which is a device that can emulate the complex behavior and interactions of subatomic particles that classical computers can’t handle. Scientists are developing quantum simulators to discover new drugs and materials. 

By generating correlated signals at room temperature, this new technique can improve the scalability and reduce the costs of quantum simulation. In the future, the researchers want to develop a scalable architecture for their platform, moving it one step closer to real-world deployment. They also want to explore additional applications for the process and use their platform to study the underlying physics of correlated microwave signals. 

“The creation of a non-degenerate parametric magnon-polariton platform marks an important milestone for cavity magnonics, extending the field beyond coherent microwave generation to the production of multichannel correlated microwave photons,” says Can-Ming Hu, a distinguished profess or physics and astronomy at the University of Manitoba in Canada, who was not involved with this paper. “This breakthrough will broadly impact secure microwave communications, hardware random number generation, correlation-based signal processing, and intelligent microwave sensing — all operating within the classical regime at room temperature. Looking ahead, this platform could well be remembered as the starting point for realizing quantum-inspired microwave sensing and communication technologies based on nonlinear cavity magnonics.”

This research was supported, in part, by the National Science Foundation and the U.S. Department of Energy.

Immune cells that are engineered to attack cancer cells, known as CAR-T cells, are used to treat some types of blood cancer. However, only about 5 percent of hospitals in the United States have the ability to generate and deliver CAR-T cells to patients. For many patients, this means the cells need to be frozen and shipped long-distance.

To help make this type of therapy accessible to more people, researchers at MIT have developed a new way to protect the cells from damage that can occur when they are frozen for storage and shipment. Their technique significantly reduces the use of a chemical preservative that is now used to protect the cells, which should make it easier for more hospitals to provide this treatment option to patients.

Instead of treating the cells with a cryoprotective chemical that has to be removed before treatment, the researchers were able to preserve them using a nontoxic antifreeze sugar.

“With this approach, you could theoretically just thaw the cells and then inject them, without any extra processing steps. We think that could allow a lot more cancer treatment centers to be able to give CAR-T cell therapy,” says Ana Jaklenec, a principal investigator in MIT’s Koch Institute for Integrative Cancer Research and one of the lead authors of the study, which appears this week in Trends in Biotechnology.

In the study, the researchers showed that cells preserved using this process had higher survival rates and could be successfully used to treat lymphoma and glioblastoma in mice.

Robert Langer, the David H. Koch Institute Professor at MIT, is also a senior author of the paper. MIT postdocs Amy Lee and Khanh Tran are the paper’s lead authors.

Preserving cells

To make CAR-T cells, doctors isolate T cells from patient blood samples. These cells are then engineered to express a protein called chimeric antigen receptor (CAR), which can be designed to target specific proteins found on cancer cells.

Then, the cells spend several weeks proliferating until there are enough to transfuse back into the patient. A small number of hospitals are equipped to generate and administer these cells, but most CAR-T cells are generated at centralized lab facilities. Once ready, these cells are frozen and shipped to a hospital or cancer treatment center.

To protect the cells from ice crystals that can damage their membranes, the cells are treated with a chemical called dimethyl sulfoxide (DMSO), which prevents ice crystal formation. This compound must be removed before the cells are transfused, but most hospitals don’t have the expertise to do this, which limits their ability to provide CAR-T cell treatment.

The process of removing DMSO can also harm cells, reducing the number of CAR-T cells that are viable and effective. In the new study, the MIT team wanted to find a way to reduce or eliminate DMSO from the process, which could make it easier for these cells to reach more patients.

“We looked at this cell-manufacturing process to see if there are ways to improve it, to increase the efficacy and hopefully eventually get to the point where these cells can be easily distributed to treatment centers,” Jaklenec says. “Our goal was to eliminate adding this chemical and really focus on safe excipients like sugars.”

The researchers employed two sugars that scientists have previously used to help cells survive cold temperatures. These sugars — trehalose and sucrose — help cells to naturally combat cold by protecting proteins from denaturation and preventing the formation of ice crystals. This antifreeze mechanism is found in many Arctic organisms, such as North American wood frogs, and helps them to survive extreme subzero temperatures. 

To get sugar molecules into the cells, the researchers used a technique called electroporation. By applying a small electrical current to the cells, they can briefly create holes in the cell membrane, allowing large molecules such as sugars to pass through. They found that they still needed to add a small amount of DMSO, but not enough that it had to be removed later. 

“We believe that our cryopreservation strategy can truly improve the cell therapeutic accessibility because with our strategy, you don’t need to remove the cryoprotectants. You could use the cells upon thawing,” Lee says.

More effective therapy

The researchers tested this technique on CAR-T cells as well as mesenchymal stem cells, which can differentiate into many other cell types and hold potential for use in regenerative medicine. For both types of cells, a higher percentage of the cells survived the freezing and thawing process when sugars were used as the main cryoprotectant instead of DMSO. 

They also used thawed CAR-T cells to treat non-Hodgkin’s lymphoma and glioblastoma, in mouse models. Mice treated with CAR-T cells preserved using the new strategy had higher survival rates than mice treated with cells preserved using the conventional DMSO approach.

“Preservation methods for living biotherapeutics have seen limited innovation, remain poorly characterized at scale, and often compromise cell viability and function after thawing,” Tran says. “We believe that our findings underscore the importance of thorough characterization and optimization of every stage of cell therapy manufacturing, which could have dramatic impacts on treatment efficacy.”

The researchers now hope to work with hospitals to explore whether their new technique could be easily integrated into the process of producing and thawing CAR-T cells.

“If that’s successful from a cell viability and functionality standpoint, perhaps we will do a small trial with patients,” Jaklenec says.

Vijay G. Sankaran, a professor of pediatrics at Boston Children’s Hospital and Harvard Medical School and a Howard Hughes Medical Institute Investigator, who was not involved in the study, says he is excited by the potential applications of the research. 

“As a pediatric hematologist and oncologist, many of the cell therapies we use, including CAR-T cells and blood stem cells, require us to collect and freeze a substantial number of cells, so that enough healthy cells are available after thawing for when patients need treatment. This work suggests an innovative approach that could help more cells survive the freezing and thawing process, potentially making these powerful therapies more reliable and effective. Of course, further work will be needed to validate these results in settings where this approach can be clinically applied,” Sankaran says.

This work was supported by postdoctoral fellowships from the Ludwig Center at MIT’s Koch Institute and the Convergence Scholars Program at the MIT Marble Center for Cancer Nanomedicine.

Concrete has served as the foundation of empires for thousands of years. Today, it’s one of the most common materials in the world. But one look at the ancient Roman concrete structures still standing suggests that ancient builders knew something about durability that we don’t.

MIT Associate Professor Admir Masic has spent his career studying ancient Roman concrete. His work has uncovered details about what gave Roman concrete its legendary durability, including the manufacturing process that endowed it with self-healing properties.

In 2021, Masic decided to apply those findings to improve the durability of modern concrete by co-founding DMAT. Today, the company has developed additional technology to create a concrete additive that increases the lifespan of concrete structures by 50 percent and reduces CO2 emissions to 40% of traditional concrete.

The company’s concrete has been used to make complex infrastructure across Europe including underground water tanks, road barriers, and pavement in Italy and Switzerland. The company plans to expand to the U.S. soon.

“We can now offer an extremely competitively priced, self-healing product that is easy to implement and available worldwide,” Masic says. “What’s exciting to me is that this material could become the industry standard without requiring companies to change how they operate. It doesn’t introduce any uncertainty, because it’s based on ancient Roman technology that has been tested for thousands of years. By applying lessons from the past, we’re enabling a better future for the modern concrete industry.”

Applying ancient insights

Masic’s research at MIT has involved using new characterization techniques to probe the chemical makeup of ancient concrete. It has also brought him to well-preserved ancient construction sites in Pompeii, where historical practices could be deconstructed.

In a 2023 study funded, in part, by the Concrete Sustainability Hub, Masic and collaborators showed that when ancient Roman concrete cracks, reservoirs of calcium inside it desolve and recrystallize to fill in the new openings. Using that insight, the team developed new concrete formulations based on the Ancient Roman technique that deliberately retain calcium-rich lime clasts throughout the mix. The researchers spent a year testing samples to show the technique improved the mechanical performance and durability of different forms of concrete.

Those findings served as the foundation of DMAT. Masic partnered with Italian entrepreneur Paolo Sabatini to commercialize the technology shortly after the paper was published.

DMAT has since developed a large portfolio of proprietary technology on top of what was licensed from MIT. As it developed its solution, DMAT worked with company laboratories to secure safety and performance certifications in the European Union and ensure it fit modern concrete-making practices.

“At DMAT, we like to view concrete as an ecosystem,” says Sabatini, who serves as DMAT’s CEO and co-founder. “How does a material become the biggest industry in the world? There are considerations around not just materials but also transportation, price, and certifications. In order to get adoption, you need to design something that fits within the current industry’s ecosystem.”

Today DMAT supplies additives that can be mixed with concrete and mortar to extend the lifespan and performance of the materials. DMAT sells its additives to developers as well as concrete manufacturers to incorporate when mixing the concrete. More recently, the company has also introduced a line of ready-mix bagged mortars for structural restoration.

“When we work with clients, we can customize the concrete mix for their project and then supply filler using our recipe,” Sabatini explains. “We provide recipes to concrete manufacturers and engineers that improve the performance of concrete. But we also work across the supply chain with developers, architects, construction companies, and others.”

The first few years of the company were spent developing the technology and establishing relationships with the industry while attaining the necessary certifications to deploy in Europe.

“What’s good about DMAT is that the company is truly embedded into the concrete industry,” Masic says. “The company isn’t selling an idea. They have gone slow and carefully chosen projects to ensure they are successful in providing self-healing concrete without significant added cost.”

Built for scale

Other self-healing concretes use bacteria or polymer substances as additives, which can be more expensive, not to mention less familiar to people in the industry. DMAT’s founders have spent years honing their recipes to achieve self-healing properties with materials more familiar to the industry.

As a result, they believe the company is now in a strong position to scale. And scalability is crucial to make an impact in the industry: Concrete today is the most produced material in the world. It’s responsible for approximately 5-8 percent of global CO2 emissions.

“There’s something profound about how ancient builders, without our modern chemistry, engineered self-healing material that still stands today,” Masic says. “My group research and work with DMAT is to make the modern built environment better by applying the best lessons from the past to today’s challenges.”

When an artificial intelligence image generator produces a portrait, whose work went into it? The question sits at the center of lawsuits, licensing deals, and proposed regulations worldwide. Artists want credit. Companies want clarity. Policymakers want a way to assign responsibility.

New work from a team of researchers at MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) suggests that for models trained on large datasets, the question may often have no answer. It's not that the tools for finding it are inadequate. The connection itself has disappeared.

The scientists identified a phenomenon they call attribution decay, where the more data a generative model is trained on, the less any individual training example matters to any particular output. It feels counterintuitive, but at sufficiently large scales, they find, you can often remove any single image from the training data, or every image by a given artist, or every photograph of a given person, and the generated sample doesn't change.

And if removing something changes nothing, the researchers argue, it can't be said to be responsible for anything. 

"If you take away a piece of data and the output of the model doesn't change, then that piece of data didn't affect the output," says Zheng Dai SM ’21, PhD ’24, former MIT CSAIL researcher and lead author on the work. "So it doesn't make much sense to attribute the output to that piece of data. And if you then do this one at a time for every other piece of data and find that the output doesn’t change for any of them either, then it doesn't make much sense to attribute the output to any one of them."

"All previous methods were approximate," says MIT Professor David Gifford, who is an MIT CSAIL principal investigator. "They really could not absolutely show that deleting individual things did not change the output. This paper introduces the first method that is absolute. You're actually deleting the inputs and deleting all influences of the inputs. This is the first exact method for doing large-scale deletion efficiently and showing that the results don't change."

Dai and Gifford's project is described in an open-access paper published today in Nature Communications.

The retraining problem

Testing this idea directly meant answering a what-if question. What would this model have produced if it had never seen this particular image? Answering it honestly means retraining the model from scratch without that image, then doing it again for the next image, and the next. With millions of training examples, the math quickly becomes prohibitive, which is why prior work in the attribution field has relied on approximations that estimate a training example's influence, rather than actually removing it.

Their workaround is an architecture they built themselves, called a "diffusion ensemble." Instead of one monolithic model, it's made up of many smaller components, each trained on a different slice of the data. Want to know what the model would do without a particular image? Just switch off the parts that saw it. No retraining, no approximation. What's left is a true counterfactual model, not an estimate of one.

Of course, a clever architecture only matters if it still works as a generator. So the team put the ensembles head to head with 24 conventional diffusion models trained on the exact same data. The images came out looking about as good by standard measures. 

One nice surprise in the numbers: The more training data, the better the ensembles held up against their single-model counterparts, a hint that they may actually be more data-efficient.

"When you have low amounts of data, they do very poorly," says Dai. "But if you have more data, it actually scales better compared to the vanilla diffusion model." 

Exploring a counterfactual universe

With ablation working, the researchers could finally ask their question at scale. Take one generated image, then imagine every alternate version of it, each produced by removing a different piece of the training data. The team calls this the image's counterfactual universe. The distance between the original and its most different alternate, the counterfactual radius, captures the most that any single piece of training data could have mattered.

They trained 24 ensembles on datasets from 256 images to more than 160,000, pulled from seven public collections including CIFAR-10, CelebA, MetFaces, and ArtBench. The pattern was consistent: The bigger the training set, the smaller the radius, shrinking along an inverse power law. It held whether differences were measured pixel by pixel or by semantic meaning, with statistical significance both ways.

The team also stress-tested their own result. Maybe ablation itself was the culprit? They redid it the brute-force way at small scale, training 1,282 separate models, and the decay showed up anyway. Maybe bigger datasets just make each removal proportionally smaller? They pinned the removed fraction in place, and it persisted. Fixed epochs, text-prompted models, class-conditioned models, four similarity metrics — the finding survived everything.

The privacy paradox

The implications run in a direction that surprised the researchers themselves.

Gifford sees the finding as bearing directly on the legal question of whether model outputs are derivative works. 

"One way to think about this is that these models are creative. They are not simply copying what they are fed, but creating brand new outputs. If those outputs have nothing to do with any individual piece of training data, that raises questions about fair use, about whether the outputs are themselves copyrightable as novel works, and about how authors get compensated when what comes out of a model isn't attributable to anything on the internet." 

Gifford also notes that the work shows how to produce outputs that are guaranteed to be unattributable, a capability he frames as an obligation for the industry, rather than a loophole. 

"In order for these companies to claim their outputs aren't derivative of the internet in a copyright-infringing way, they need to revise their models to take advantage of the advances in this work, so they can show they're not creating derivatives of individual people or items."

The work looks at diffusion models, now dominant in generating audiovisual media and prevalent in scientific applications including protein structure modeling and therapeutic discovery. Whether the same decay holds for the large language models at the center of the highest-profile copyright litigation is still an open question.

"If attribution worked, it would reliably tell us whether similarities between a model's output and a copyright-protected work are due to copying or coincidence," says James Grimmelmann, a law professor at Cornell Law School and Cornell Tech. "But this paper provides reason to think that attribution will fail for interesting models. Instead, technologists and courts will need to resort to other methods for assessing copying."

Dai and Gifford's work was supported by Schmidt Futures.