Historically, scientists studying Alzheimer's disease have focused on two proteins that accumulate in patients' brains: amyloid-beta and tau. However, there's a "third thing that people didn't pay attention to, until recently," said study co-author Li-Huei Tsai, a neuroscientist at MIT. That third thing is the accumulation of droplets of fat, or lipids, in the brain, she told Live Science.
The APOE gene makes proteins responsible for transporting fat around cells throughout the body and brain. A version of the gene called APOE4, which increases the risk of late-onset Alzheimer's three- to 12-fold, makes proteins that are less efficient at this function, Tsai said. But how APOE4 affects different brain cells was not clearly understood.
Tsai's team has been working to close that gap for about a decade.
"APOE4 can influence many different cell types," she emphasized. "It's not just one cell type."
Zeroing in on the clues
In past laboratory research, Tsai's team found that APOE4 causes a buildup of cholesterol in astrocytes, star-shaped brain cells that support the function of neurons. Genetic analyses suggested that the gene variant may mess with how these cells process fat.
Another study zoomed in on oligodendrocytes, cells that help insulate neurons' outgoing wires with a fatty material called myelin. This material helps neurons send messages more efficiently. APOE4 caused the oligodendrocytes to accumulate cholesterol. So, in lab mice, the team used a drug to help the cells get rid of that extra cholesterol, and it improved the cells' ability to make myelin and the mice's performance on cognitive tests.
In their latest research, published Thursday (Oct. 15) in the journal Cell Stem Cell, Tsai's team dug into the question of why APOE4 causes fat to accumulate in the first place.
They used stem cells to grow APOE4-carrying oligodendrocytes and then turned specific genes on or off, to see if they increased or decreased the formation of fat droplets. This screening pointed to a major group of genes involved in cell-to-cell communication, called the Wnt signaling pathway. This included a key gene called DKK1, which acts as a brake on this signaling system.
Disabling DKK1 — essentially lifting the brake — reduced the concentration of fat droplets in the cells, the team found.
In brain tissue samples from organ donors with APOE4, the team also found signs of low Wnt signaling. This finding "establishes a very strong link" between the signaling pathway and formation of fat droplets, said Rosie Jackson, a cell biologist at the University of Dundee who was not involved in this study.
Myelin (orange) is a fatty material that coats the outgoing wires of many neurons, helping the cells send signals more efficiently. (Image credit: koto_feja via Getty Images)APOE4's effects across the brain
The researchers then investigated exactly how Wnt signaling was disrupted in oligodendrocytes.
APOE4 seemed to decrease the cells' supply of a protein that helps relay Wnt signals. That helpful protein was being broken down by an enzyme whose levels had been boosted by the APOE4. In mice, the team used a drug to block that extra enzyme, and the treatment ultimately restored normal Wnt signaling, reduced fat buildup, and helped the cells make myelin properly.
This could change how scientists think about treating Alzheimer's disease, Jackson said. It "opens up a new wave of potential therapeutics" that could target the accumulated fats, rather than going after only amyloid-beta and tau, she said.
"I think that this is becoming a really common idea in the field — that we're going to need multiple medications that target slightly different elements," she told Live Science. "And targeting lipids will be a really important part of that."
That said, the findings have not been confirmed in humans yet.
"It's very encouraging that facilitating lipid clearance and lipid movement is a potential therapeutic avenue," said study co-author Joel Blanchard, a neuroscientist at the Icahn School of Medicine at Mount Sinai. The question is how we can do that in a safe and effective manner in humans, he said.
Blanchard has led other research that underscores APOE4's effects beyond fats and myelin. The gene variant may affect cells that support blood vessels in the brain by driving scarring and contributing to amyloid-beta buildup. It may also mess with the "trash collectors" inside astrocytes, again leading to a buildup of harmful proteins.
While cells with other versions of the APOE gene can put a lot of energy toward breaking down the proteins, APOE4-carrying cells put all their energy toward fat storage, Blanchard said.
But more research is needed to understand how to fix the problem, he said. Hopefully, targeting fat storage "could rescue not just one thing but multiple things that go wrong in the course of Alzheimer's disease," he told Live Science.
This article is for informational purposes only and is not meant to offer medical advice.
'> A gene variant triples the risk of getting Alzheimer's disease — fat buildup in brain cells may be to blame, study finds A drone powered by artificial intelligence (AI) has completed a championship race course faster than some of the world's best human drone pilots.
In a study published Oct. 14 in the journal Science Robotics, scientists created an AI system capable of pushing a quadcopter through a demanding course at over 60 mph (100 km/h), using little more than a single forward-facing camera, an inertial measurement unit, and a tiny onboard computer.
The system, called MonoRace, even won the 2025 Abu Dhabi Autonomous Racing League drone racing competition, outperforming every rival autonomous team and posting faster times than three former human world-champion drone pilots.
MonoRace accomplished this without the aid of GPS, external motion-capture systems, stereo cameras, or a preloaded map of the course. The system is an onboard autonomous flight model created by researchers at the Delft University of Technology's Micro Air Vehicles Lab in the Netherlands. Its name reflects its central design feature; "mono" refers to its monocular vision. The drone relies on a single rolling shutter camera and monitors its own motion with an internal measurement unit (IMU) that gauges acceleration and rotation.
Flying at breakneck speeds
The limited sensor package feeds a neural network, a machine-learning model that uses interconnected artificial neurons to learn patterns from data. This neural network identifies racing gates and estimates the drone's position and motion to generate motor commands. Then, it sends the commands to the four propellers at a rate of 500 times per second through the drone's flight controller.
The system fits a drone that weighs a mere 966 grams (2.13 pounds) and measures approximately 11.8 x 11 x 3.9 inches (30 x 28 x 10 centimeters). Because the MonoRace hardware was so compact, lightweight and computationally efficient, the team was able to operate it entirely aboard the drone.
On the Abu Dhabi course, MonoRace completed two laps around 11 gates in 16.56 seconds, hitting a top speed of approximately 63 mph (102 km/h).
A recently released video tracks MonoRace's development and performance in the competition, with real flight sequences demonstrating how the drone banks sharply, accelerates through gates, and navigates a course without human assistance.
However, some of the sequences are computer-generated. Study co-author Guido de Croon, a professor at the Delft University of Technology, said those are intentional visual effects. The only fully simulated imagery appears at about the two-minute mark, during the segment showing the training process.
"Here we show the gates in a visual simulator for the video but the real training happened in a stripped-down, pure physics simulator that ran really fast," de Croon told Live Science in an email. "Simulating vision was not done during training, as the control neural network receives an estimated state as input, which comes from a different perception and state estimation process."
The rest of the footage is real, although some shots include overlays showing how the drone interprets its camera feed. A convolutional neural network segments the white gates, sometimes displaying them as red in the normal camera image. Small circles mark detected gate corners, red lines indicate gate edges, and a blue overlay shows the portion of the image where the drone expects to find the next gate.
How the AI learned to race
Scientists trained the MonoRace model with reinforcement learning, a machine learning method in which an agent repeatedly attempts a task and receives feedback based on its success or failure.
"The reinforcement learning ran in simulation, because it needs lots of trial and error — like millions of tries," de Croon said. "It actually learns to fly from scratch."
The network learned to translate the drone's position on the track, velocity and orientation directly into motor commands. According to de Croon, early training focused on the basics, like avoiding crashes and reaching the first gate. As the training progressed, the system learned to advance farther along the course and fly at higher speeds.
The researchers didn't train the system on a single aircraft. Instead, they varied dimensions such as weight, voltage, drag, sensor noise and propeller condition so the AI model could cope with real-world variations and imperfections.
"That ensures that the trained network can control different drones," de Croon said.
MonoRace could become a training tool for human drone pilots, "a bit like powerful chess programs have also inspired humans playing the game," de Croon said.
Its broader significance lies in demonstrating advancements in physical AI, especially how a robot can perceive, decide and act at high speeds despite limited sensing and computing power. The control network runs on a microcontroller-class chip, meaning that sophisticated control doesn't require a large computer.
De Croon noted that a similarly sized network could optimize for minimal energy use rather than minimum lap time. Because most machines rely on small processors and conventional controllers, energy-optimal neural control could eventually reduce power consumption in everyday robotics, he said.
'> World's fastest AI-powered racing drone outflies the best human pilots . These thrusters release bursts of high-pressure gas to generate thrust, allowing the vehicle to accelerate rapidly and, according to Tesla, become airborne for short periods.
The new Roadster will come in three variants: Performance, Thrust and Rocket. Only the Rocket is capable of flight. The Performance model is a two-seater equipped with a suction system that's designed to improve traction and handling (or drive up inverted ramps, as in Tesla's demo), while the Thrust model adds rear-mounted thrusters that Tesla claims can propel the vehicle from 0 to 60 mph (97 km/h) in under a second.
The highest-end model, Rocket, features SpaceX cold-gas thrusters positioned around the vehicle to enable short bursts of flight and in-air control. This single-seat model will be available by invitation only, according to Tesla.
Tesla hasn't disclosed how far or for how long its EV can remain airborne, nor is it the first company to develop a hybrid road-air vehicle. There is also skepticism from outside viewers on whether the demonstration could be replicated outside of controlled scenarios.
"Even assuming that 50% of the mass of the car is the compressed air, that gives you less than 10s of useful thrust (with thrust-to-weight ratio larger than 1.0). Probably more like 5 seconds, actually," he posted on the social media platform. "So the car must be an empty, lightweight shell containing big compressed air tanks. Cute demo, but totally impractical. But of course, you knew that."
He likened it to SpaceShipOne, which debuted in 2004 as the first privately funded spacecraft and made three record-setting flights before being retired.
"Welcome Tesla to the flying car race, we've been here alone for eleven years," Alef CEO Jim Dukhovny said in a statement emailed to Live Science. "Having competition is a good thing, especially from such a major company as Tesla."
Production of the new Roadster is scheduled for 2028, but Tesla doesn't have the best track record of meeting its deadlines. The company has delayed production of its Optimus humanoid robot a few times since first unveiling the machine in 2021, for example, with an initial start in 2023, before being pushed back. Only in recent months has production begun to scale to a few hundred units per week.
'> 'Cute demo, but totally impractical': Experts are skeptical that Tesla's 'Rocket' Roadster could actually work in the real world
in the first century B.C.
The fortified military camp is made up of an earthen platform roughly 39 feet (12 meters) wide, which is surrounded by a stone perimeter wall about 5 feet (1.5 m) thick, according to a translated statement. Uncovered during excavations earlier this year, the camp is situated about one-third of a mile (500 m) from where an Iron Age settlement was once inhabited by the Cantabri people, a group of Celtic tribes that lived among the Cantabrian Mountains in modern-day northern Spain.
The Cantabri fiercely resisted Roman rule for a decade, but in 19 B.C., they were conquered by the military commanded by Emperor Augustus. The Cantabrian Wars marked the end of local control over the territory, as the Iberian Peninsula then came under control of the Roman Empire.
Among the well-preserved remains at the camp were metal weapon fragments and small iron tacks known as hobnails, which were hammered into the bottom of Roman soldiers' leather military sandals to help provide traction on rocky and muddy terrain. These objects are evidence that the Roman military was prepared to attack and subdue the Iron Age community, according to the statement.
The archaeologists also uncovered evidence of a major fire that destroyed most of the pre-Roman settlement, including two temples. González Gómez de Agüero suggested that the Romans may have deliberately set the fire as part of their conquest.
"It's not just about defeating the enemy, but about humiliating them, erasing their identity, and breaking their connection to the land," he said in the statement.
Researchers are still working on establishing the chronology of the Romans' final attack on the Cantabri and want to understand if the settlement was destroyed in a single event or over time. Previously, another research team found 66 Roman camps in northern Spain, showing the immensity of the empire's attacks against the Cantabri.
The latest find is helping researchers understand the scale and location of Rome's fight for control of the Iberian Peninsula.
"It is an important step in understanding how the conquest was carried out on the Cantabrian slope of León, an area where it was necessary to link the Roman camps with Iron Age settlements," Carlos Fernández Rodríguez, an archaeologist at the University of León, said in the statement.
'> Romans may have set fire to Celtic town during conquest to 'humiliate' the enemy, archaeologists find