Aurelia Butler

Depending on age, humans need 7 to 13 hours of sleep per 24 hours. During this time, a lot happens: Heart rate, breathing and metabolism ebb and flow; hormone levels adjust; the body relaxes. Not so much in the brain.
“The brain is very busy when we sleep, repeating what we have learned during the day,” said Maxim Bazhenov, PhD, professor of medicine and a sleep researcher at University of California San Diego School of Medicine. “Sleep helps reorganize memories and presents them in the most efficient way.”
In previous published work, Bazhenov and colleagues have reported how sleep builds rational memory, the ability to remember arbitrary or indirect associations between objects, people or events, and protects against forgetting old memories.
Artificial neural networks leverage the architecture of the human brain to improve numerous technologies and systems, from basic science and medicine to finance and social media. In some ways, they have achieved superhuman performance, such as computational speed, but they fail in one key aspect: When artificial neural networks learn sequentially, new information overwrites previous information, a phenomenon called catastrophic forgetting.
“In contrast, the human brain learns continuously and incorporates new data into existing knowledge,” said Bazhenov, “and it typically learns best when new training is interleaved with periods of sleep for memory consolidation.”
Writing in the November 18, 2022 issue of PLOS Computational Biology, senior author Bazhenov and colleagues discuss how biological models may help mitigate the threat of catastrophic forgetting in artificial neural networks, boosting their utility across a spectrum of research interests. More

Inspired by the biomechanics of the manta ray, researchers at North Carolina State University have developed an energy-efficient soft robot that can swim more than four times faster than previous swimming soft robots. The robots are called “butterfly bots,” because their swimming motion resembles the way a person’s arms move when they are swimming the butterfly stroke.
“To date, swimming soft robots have not been able to swim faster than one body length per second, but marine animals — such as manta rays — are able to swim much faster, and much more efficiently,” says Jie Yin, corresponding author of a paper on the work and an associate professor of mechanical and aerospace engineering at NC State. “We wanted to draw on the biomechanics of these animals to see if we could develop faster, more energy-efficient soft robots. The prototypes we’ve developed work exceptionally well.”
The researchers developed two types of butterfly bots. One was built specifically for speed, and was able to reach average speeds of 3.74 body lengths per second. A second was designed to be highly maneuverable, capable of making sharp turns to the right or left. This maneuverable prototype was able to reach speeds of 1.7 body lengths per second.
“Researchers who study aerodynamics and biomechanics use something called a Strouhal number to assess the energy efficiency of flying and swimming animals,” says Yinding Chi, first author of the paper and a recent Ph.D. graduate of NC State. “Peak propulsive efficiency occurs when an animal swims or flies with a Strouhal number of between 0.2 and 0.4. Both of our butterfly bots had Strouhal numbers in this range.”
The butterfly bots derive their swimming power from their wings, which are “bistable,” meaning the wings have two stable states. The wing is similar to a snap hair clip. A hair clip is stable until you apply a certain amount of energy (by bending it). When the amount of energy reaches critical point, the hair clip snaps into a different shape — which is also stable.
In the butterfly bots, the hair clip-inspired bistable wings are attached to a soft, silicone body. Users control the switch between the two stable states in the wings by pumping air into chambers inside the soft body. As those chambers inflate and deflate, the body bends up and down — forcing the wings to snap back and forth with it.
“Most previous attempts to develop flapping robots have focused on using motors to provide power directly to the wings,” Yin says. “Our approach uses bistable wings that are passively driven by moving the central body. This is an important distinction, because it allows for a simplified design, which lowers the weight.”
The faster butterfly bot has only one “drive unit” — the soft body — which controls both of its wings. This makes it very fast, but difficult to turn left or right. The maneuverable butterfly bot essentially has two drive units, which are connected side by side. This design allows users to manipulate the wings on both sides, or to “flap” only one wing, which is what enables it to make sharp turns.
“This work is an exciting proof of concept, but it has limitations,” Yin says. “Most obviously, the current prototypes are tethered by slender tubing, which is what we use to pump air into the central bodies. We’re currently working to develop an untethered, autonomous version.”
The paper, “Snapping for high-speed and high-efficient, butterfly stroke-like soft swimmer,” will be published Nov. 18 in the open-access journal Science Advances. The paper was co-authored by Yaoye Hong, a Ph.D. student at NC State; and by Yao Zhao and Yanbin Li, who are postdoctoral researchers at NC State. The work was done with support from the National Science Foundation under grants CMMI-2005374 and CMMI-2126072.
Video of the butterfly bots can be found at https://youtu.be/Pi-2pPDWC1w.
Story Source:
Materials provided by North Carolina State University. Original written by Matt Shipman. Note: Content may be edited for style and length. More

A mind-controlled wheelchair can help a paralyzed person gain new mobility by translating users’ thoughts into mechanical commands. On November 18 in the journal iScience, researchers demonstrate that tetraplegic users can operate mind-controlled wheelchairs in a natural, cluttered environment after training for an extended period.
“We show that mutual learning of both the user and the brain-machine interface algorithm are both important for users to successfully operate such wheelchairs,” says José del R. Millán, the study’s corresponding author at The University of Texas at Austin. “Our research highlights a potential pathway for improved clinical translation of non-invasive brain-machine interface technology.”
Millán and his colleagues recruited three tetraplegic people for the longitudinal study. Each of the participants underwent training sessions three times per week for 2 to 5 months. The participants wore a skullcap that detected their brain activities through electroencephalography (EEG), which would be converted to mechanical commands for the wheelchairs via a brain-machine interface device. The participants were asked to control the direction of the wheelchair by thinking about moving their body parts. Specifically, they needed to think about moving both hands to turn left and both feet to turn right.
In the first training session, three participants had similar levels of accuracy — when the device’s responses aligned with users’ thoughts — of around 43% to 55%. Over the course of training, the brain-machine interface device team saw significant improvement in accuracy in participant 1, who reached an accuracy of over 95% by the end of his training. The team also observed an increase in accuracy in participant 3 to 98% halfway through his training before the team updated his device with a new algorithm.
The improvement seen in participants 1 and 3 is correlated with improvement in feature discriminancy, which is the algorithm’s ability to discriminate the brain activity pattern encoded for “go left” thoughts from that for “go right.” The team found that the better feature discrimnancy is not only a result of machine learning of the device but also learning in the brain of the participants. The EEG of participants 1 and 3 showed clear shifts in brainwave patterns as they improved accuracy in mind-controlling the device.
“We see from the EEG results that the subject has consolidated a skill of modulating different parts of their brains to generate a pattern for ‘go left’ and a different pattern for ‘go right,'” Millán says. “We believe there is a cortical reorganization that happened as a result of the participants’ learning process.”
Compared with participants 1 and 3, participant 2 had no significant changes in brain activity patterns throughout the training. His accuracy increased only slightly during the first few sessions, which remained stable for the rest of the training period. It suggests machine learning alone is insufficient for successfully maneuvering such a mind-controlled device, Millán says
By the end of the training, all participants were asked to drive their wheelchairs across a cluttered hospital room. They had to go around obstacles such as a room divider and hospital beds, which are set up to simulate the real-world environment. Both participants 1 and 3 finished the task while participant 2 failed to complete it.
“It seems that for someone to acquire good brain-machine interface control that allows them to perform relatively complex daily activity like driving the wheelchair in a natural environment, it requires some neuroplastic reorganization in our cortex,” Millán says.
The study also emphasized the role of long-term training in users. Although participant 1 performed exceptionally at the end, he struggled in the first few training sessions as well, Millán says. The longitudinal study is one of the first to evaluate the clinical translation of non-invasive brain-machine interface technology in tetraplegic people.
Next, the team wants to figure out why participant 2 didn’t experience the learning effect. They hope to conduct a more detailed analysis of all participants’ brain signals to understand their differences and possible interventions for people struggling with the learning process in the future.
Story Source:
Materials provided by Cell Press. Note: Content may be edited for style and length. More

Researchers at the Hong Kong University of Science and Technology (HKUST) have recently developed a novel integration scheme for efficient coupling between III-V compound semiconductor devices and silicon components on silicon photonics (Si-photonics) platform by selective direct epitaxy, unlocking the potential of integrating energy-efficient photonics with cost-effective electronics, as well as enabling the next generation telecommunications with low cost, high speed and large capacity.
Over the past few years, data traffic has been growing exponentially driven by various applications and emerging techniques such as big data, automobiles, cloud applications and sensors. To address the issues, Si-photonics has been widely investigated as a core technology to enable, extend, and increase data transmission through energy-efficient, high-capacity and low-cost optical interconnects. While silicon-based passive components have been well established on Si-photonics platform, the lasers and photodetectors can’t be realized by silicon and necessitate the integration of other materials such as III-V compound semiconductors on silicon.
III-V lasers and photodetectors on silicon has been investigated by two main methods. The first one is the bonding-based method which has yielded devices with impressive performance. However, it requires complicated manufacturing technique that is low yield and high-cost, making mass production very challenging. The other way is direct epitaxy method by growing multiple layers of III-V on silicon. While it provides a solution with lower cost, larger scalability and higher integration density, the micrometers thick III-V buffer layers which are crucial for this method hinders efficient light coupling between III-V and silicon — the key for integrated Si-photonics.
To address these issues, the team led by Prof. Kei-May LAU, Professor Emeritus of the Department of Electronic and Computer Engineering at Hong Kong University of Science and Technology (HKUST), developed lateral aspect ratio trapping (LART) — a novel selective direct epitaxy method that can selectively grow III-V materials on silicon-on-insulator (SOI) in a lateral direction without the need of thick buffers. Furthermore, based on this novel technology, the team devised and demonstrated unique in-plane integration of III-V photodetectors and silicon elements with high coupling efficiency between III-V and silicon. Compared to the commercial ones, the performance of photodetectors by such approach is less noisy, more sensitive, and has wider operation range, with record-high speed of over 112 Gb/s — way faster than existing products. For the first time, the III-V devices can be efficiently coupled with Si elements by direct epitaxy. The integration strategy can be easily applied to the integration of various III-V devices and Si-based components, thereby enabling the ultimate goal of integrating photonics with electronics on silicon photonics platform for data communications.
“This was made possible by our latest development of a novel growth technique named lateral aspect ratio trapping (LART) and our unique design of coupling strategy on the SOI platform. Our team’s combined expertise and insights into both device physics and growth mechanisms allow us to accomplish the challenging task of efficient coupling between III-V and Si and cross-correlated analysis of epitaxial growth and device performance,” said Prof. Lau. “This work will provide practical solutions for photonic integrated circuits and fully integrated Si-photonics, light coupling between III-V lasers and Si components can be realized through this method” said Dr. Ying Xue, first author of the study.
This is a collaborative work with a research team led by Prof. Hon Ki Tsang of Department of Electronic Engineering at Chinese University of Hong Kong (CUHK) and a research team led by Prof. Xinlun Cai of School of Electronics and Information Technology at Sun Yat-sen University (SYSU). The device fabrication technology in the work was developed at HKUST’s Nanosystem Fabrication Facility (NFF) on Clear Water Bay campus. The work is supported by Research Grants Council of Hong Kong and Innovation Technology Fund of Hong Kong. This work has recently been published in Optica.
Story Source:
Materials provided by Hong Kong University of Science and Technology. Note: Content may be edited for style and length. More

Selfless behaviour and cooperation cannot be taken for granted. Mohammad Salahshour of the Max Planck Institute for Mathematics in the Sciences (now at Max Planck Institute of Animal Behavior), has used a game theory-based approach to show why it can be worthwhile for individuals to set self-interests aside.
One of the most fundamental questions facing humanity is: why do we behave morally? Because it is by no means self-evident that under certain circumstances we set our self-interest aside and put ourselves in the service of a group — sometimes to the point of self-sacrifice. Many theories have been developed to get to the bottom of this moral conundrum. There are two well-known proposed solutions: that individuals help their relatives so that the common genes survive (kin selection), and that the principle of “you scratch my back and I’ll scratch yours” applies. If people help each other, everyone benefits in the end (principle of reciprocity).
Prisoner’s dilemma combined with a coordination game
Mathematician Mohammad Salahshour of the Max Planck Institute for Mathematics in the Sciences in Leipzig, Germany, has used the tools of game theory to explain the emergence of moral norms — because game theory studies how people make rational decisions in conflict situations. For Salahshour, the question at the outset was: why do moral norms exist in the first place? And why do we have different, or even contrasting moral norms? For example, while some norms such as “help others,” promote self-sacrificing behaviour, others, such as dress codes, appear not to have much to do with curbing selfishness. To answer these questions, Salahshour coupled two games: first, the classic prisoner’s dilemma, in which two players must decide whether to cooperate for a small reward or betray themselves for a much larger reward (social dilemma). This game can be a typical example of a social dilemma, where success of a group as a whole requires individuals to behave selflessly. In this game everybody loses out if too many members of a group behave selfishly, compared to a scenario in which everybody acts altruistically. However, if only a few individuals behave selfishly, they can receive a better outcome than their altruistic team members. .Second, a game that focuses on typical decisions within groups, such as a coordination task, distribution of resources, choice of a leader, or conflict resolution. Many of these problems can be ultimately categorized as coordination or anticoordination problems.
Without coupling the two games, it is clear that in the Prisoner’s Dilemma, cooperation does not pay off, and self-interested behaviour is the best choice from the individual’s perspective if there are enough people who act selflessly. But individuals who act selfishly are not able to solve coordination problems efficiently and lose a lot of resources due to failing to coordinate their activity. The situation can be completely different when the results of the two games are considered as a whole and there are moral norms at work which favour cooperation: now cooperation in the prisoner’s dilemma can suddenly pay off because the gain in the second game more than compensates for the loss in the first game.
Out of self-interest to coordination and cooperation
As a result of this process, not only cooperative behaviour emerges, but also a social order. All individuals benefit from it — and for this reason, moral behaviour pay off for them. “In my evolutionary model, there were no selfless behaviours at the beginning, but more and more moral norms emerged as a result of the coupling of the two games,” Salahshour reports. “Then I observed a sudden transition to a system where there is a lot of cooperation.” In this “moral state,” a set of norms of coordination evolve which help individuals to better coordinate their activity, and it is precisely through this that social norms and moral standards can emerge. However, coordination norms favour cooperation: cooperation turns out to be a rewarding behaviour for the individual as well. Mahammad Salahshour: “A moral system behaves like a Trojan horse: once established out of the individuals’ self-interest to promote order and organization, it also brings self-sacrificing cooperation.”
Through his work, Salahshour hopes to better understand social systems. “This can help improve people’s lives in the future,” he explains. “But you can also use my game-theoretic approach to explain the emergence of social norms in social media. There, people exchange information and make strategic decisions at the same time — for example, who to support or what cause to support.” Again, he said, two dynamics are at work at once: the exchange of information and the emergence of cooperative strategies. Their interplay is not yet well understood — but perhaps game theory will soon shed new light on this topical issue as well.
Story Source:
Materials provided by Max-Planck-Gesellschaft. Note: Content may be edited for style and length. More

Your entire medical journey lives in digital health records, but how do you know if those records are wrong, incomplete or missing important information? That’s the focus of research done by Varadraj Gurupur, associate professor in UCF’s School of Global Health Management and Informatics.
His latest project created an algorithm that can predict and measure the incompleteness of electronic health records — in everything from your lab results to disease diagnoses, medical history to prescription records.
Missing information in the electronic health records (EHR) that hospitals and doctor’s offices keep is like a leaking pipe, he says. If you don’t know where the leak is, you can’t fix it, and soon the house can flood. The same dangers can happen in healthcare. A recent study by Gurupur revealed that a critical percent of digital health records contained missing information.
His algorithm uses mathematics and computer science to answer, “Where is the water leaking?” he says. The analysis performed by Gurupur and his team found that the level of incompleteness per year varies and there is not a pattern of where missing data happens. His algorithm helps identify attributes that have a higher tendency to be incomplete — the areas of the water pipe that are more vulnerable and can break more frequently.
His previous studies have documented that the biggest reasons for missing health information are communication and education. Communication between patients and their providers isn’t always clear — especially if the patient is interacting with a healthcare professional who does not speak their native language. Cultural barriers may keep patients from sharing important information with their providers. Digital technology also creates its own challenges. Providers may not fill out electronic records until day’s end — and forget what the patient said or not have it accurately in their notes. Hospitals and clinics switch electronic health record systems, requiring extensive new retraining which results in a learning curve for providers. Some healthcare workers, especially those who did not grow up with technology, may not be adept at using EHRs.
“Missing health information can sometimes be as simple as a person who isn’t sure what button to push in the new system,” Gurupur says. More

Researchers at the National Institute of Standards and Technology (NIST) have created grids of tiny clumps of atoms known as quantum dots and studied what happens when electrons dive into these archipelagos of atomic islands. Measuring the behavior of electrons in these relatively simple setups promises deep insights into how electrons behave in complex real-world materials and could help researchers engineer devices that make possible powerful quantum computers and other innovative technologies.
In work published in Nature Communications, the researchers made multiple 3-by-3 grids of precisely spaced quantum dots, each comprising one to three phosphorus atoms. Attached to the grids were electrical leads and other components that enabled electrons to flow through them. The grids provided playing fields in which electrons could behave in nearly ideal, textbook-like conditions, free of the confounding effects of real-world materials.
The researchers injected electrons into the grids and observed how they behaved as the researchers varied conditions such as the spacing between the dots. For grids in which the dots were close, the electrons tended to spread out and act like waves, essentially existing in several places at one time. When the dots were far apart, they would sometimes get trapped in individual dots, like electrons in materials with insulating properties.
Advanced versions of the grid would allow researchers to study the behavior of electrons in controllable environments with a level of detail that would be impossible for the world’s most powerful conventional computers to simulate accurately. It would open the door to full-fledged “analog quantum simulators” that unlock the secrets of exotic materials such as high-temperature superconductors. It could also provide hints about how to create materials, such as topological insulators, by controlling the geometry of the quantum dot array.
In related work just published in ACS Nano, the same NIST researchers improved their fabrication method so they can now reliably create an array of identical, equally spaced dots with exactly one atom each, leading to even more ideal environments necessary for a fully accurate quantum simulator. The researchers have set their sights on making such a simulator with a larger grid of quantum dots: A 5×5 array of dots can produce rich electron behavior that is impossible to simulate in even the most advanced supercomputers.
Story Source:
Materials provided by National Institute of Standards and Technology (NIST). Note: Content may be edited for style and length. More