More stories

  • in

    Can an image-based electrocardiographic algorithm improve access to care in remote settings?

    Researchers at the Yale Cardiovascular Data Science (CarDS) Lab have developed an artificial intelligence (AI)-based model for clinical diagnosis that can use electrocardiogram (ECG) images, regardless of format or layout, to diagnose multiple heart rhythm and conduction disorders.
    The team led by Dr. Rohan Khera, assistant professor in cardiovascular medicine, developed a novel multilabel automated diagnosis model from ECG images. ECG Dx © is the latest tool from the CarDS Lab designed to make AI-based ECG interpretation accessible in remote settings. They hope the new technology provides an improved method to diagnose key cardiac disorders. The findings were published in Nature Communications on March 24.
    The first author of the study is Veer Sangha, a computer science major at Yale College. “Our study suggests that image and signal models performed comparably for clinical labels on multiple datasets,” said Sangha. “Our approach could expand the applications of artificial intelligence to clinical care targeting increasingly complex challenges.”
    As mobile technology improves, patients increasingly have access to ECG images, which raises new questions about how to incorporate these devices in patient care. Under Khera’s mentorship, Sangha’s research at the CarDS Lab analyzes multi-modal inputs from electronic health records to design potential solutions.
    The model is based on data collected from more than 2 million ECGs from more than 1.5 million patients who received care in Brazil from 2010 to 2017. One in six patients was diagnosed with rhythm disorders. The tool was independently validated through multiple international data sources, with high accuracy for clinical diagnosis from ECGs.
    Machine learning (ML) approaches, specifically those that use deep learning, have transformed automated diagnostic decision-making. For ECGs, they have led to the development of tools that allow clinicians to find hidden or complex patterns. However, deep learning tools use signal-based models, which according to Khera have not been optimized for remote health care settings. Image-based models may offer improvement in the automated diagnosis from ECGs.
    There are a number of clinical and technical challenges when using AI-based applications.
    “Current AI tools rely on raw electrocardiographic signals instead of stored images, which are far more common as ECGs are often printed and scanned as images. Also, many AI-based diagnostic tools are designed for individual clinical disorders, and therefore, may have limited utility in a clinical setting where multiple ECG abnormalities co-occur,” said Khera. “A key advance is that the technology is designed to be smart — it is not dependent on specific ECG layouts and can adapt to existing variations and new layouts. In that respect, it can perform like expert human readers, identifying multiple clinical diagnoses across different formats of printed ECGs that vary across hospitals and countries.”
    This study was supported by research funding from the National Heart, Lung, and Blood Institute of the National Institutes of Health (K23HL153775).
    Story Source:
    Materials provided by Yale University. Original written by Elisabeth Reitman. Note: Content may be edited for style and length. More

  • in

    Physiological signals could be the key to 'emotionally intelligent' AI, scientists say

    Speech and language recognition technology is a rapidly developing field, which has led to the emergence of novel speech dialog systems, such as Amazon Alexa and Siri. A significant milestone in the development of dialog artificial intelligence (AI) systems is the addition of emotional intelligence. A system able to recognize the emotional states of the user, in addition to understanding language, would generate a more empathetic response, leading to a more immersive experience for the user.
    “Multimodal sentiment analysis” is a group of methods that constitute the gold standard for an AI dialog system with sentiment detection. These methods can automatically analyze a person’s psychological state from their speech, voice color, facial expression, and posture and are crucial for human-centered AI systems. The technique could potentially realize an emotionally intelligent AI with beyond-human capabilities, which understands the user’s sentiment and generates a response accordingly.
    However, current emotion estimation methods focus only on observable information and do not account for the information contained in unobservable signals, such as physiological signals. Such signals are a potential gold mine of emotions that could improve the sentiment estimation performance tremendously.
    In a new study published in the journal IEEE Transactions on Affective Computing, physiological signals were added to multimodal sentiment analysis for the first time by researchers from Japan, a collaborative team comprising Associate Professor Shogo Okada from Japan Advanced Institute of Science and Technology (JAIST) and Prof. Kazunori Komatani from the Institute of Scientific and Industrial Research at Osaka University. “Humans are very good at concealing their feelings. The internal emotional state of a user is not always accurately reflected by the content of the dialog, but since it is difficult for a person to consciously control their biological signals, such as heart rate, it may be useful to use these for estimating their emotional state. This could make for an AI with sentiment estimation capabilities that are beyond human,” explains Dr. Okada.
    The team analyzed 2468 exchanges with a dialog AI obtained from 26 participants to estimate the level of enjoyment experienced by the user during the conversation. The user was then asked to assess how enjoyable or boring they found the conversation to be. The team used the multimodal dialogue data set named “Hazumi1911,” which uniquely combined speech recognition, voice color sensors, facial expression and posture detection with skin potential, a form of physiological response sensing.
    “On comparing all the separate sources of information, the biological signal information proved to be more effective than voice and facial expression. When we combined the language information with biological signal information to estimate the self-assessed internal state while talking with the system, the AI’s performance became comparable to that of a human,” comments an excited Dr. Okada.
    These findings suggest that the detection of physiological signals in humans, which typically remain hidden from our view, could pave the way for highly emotionally intelligent AI-based dialog systems, making for more natural and satisfying human-machine interactions. Moreover, emotionally intelligent AI systems could help identify and monitor mental illness by sensing a change in daily emotional states. They could also come handy in education where the AI could gauge whether the learner is interested and excited over a topic of discussion, or bored, leading to changes in teaching strategy and more efficient educational services.
    Story Source:
    Materials provided by Japan Advanced Institute of Science and Technology. Note: Content may be edited for style and length. More

  • in

    When a band falls flat: Searching for flatness in materials

    Finding the right ingredients to create materials with exotic quantum properties has been a chimera for experimental scientists, due to the endless possible combinations of different elements to be synthesized.
    From now on, the creation of such materials could be less blindfolded thanks to an international collaboration led by Andrei Bernevig, Ikerbasque visiting professor at Donostia International Physics Center (DIPC) and professor at Princeton University, and Nicolas Regnault, from Princeton University and the Ecole Normale Supérieure Paris, CNRS, including the participation of Luis Elcoro from the University of the Basque Country (UPV/EHU).
    The team conducted a systematic search for potential candidates in a massive haystack of 55,000 materials. The elimination process started with the identification of the so-called flat band materials, that is, electronic states with constant kinetic energy. Therefore, in a flat band the behavior of the electrons is governed mostly by the interactions with other electrons. However, researchers realized that flatness is not the only requirement, because when electrons are too tightly bound to the atoms, even in a flat band, they are not able to move around and create interesting states of matter. “You want electrons to see each other, something you can achieve by making sure they are extended in space. That’s exactly what topological bands bring to the table,” says Nicolas Regnault.
    Topology plays a crucial role in modern condensed matter physics as suggested by the three Nobel prizes in 1985, 1997 and 2016. It enforces some quantum wave functions to be extended making them insensitive to local perturbation such as impurities. It might impose some physical properties, such as a resistance, to be quantized or lead to perfectly conducting surface states.
    Fortunately, the team has been at the forefront of characterizing topological properties of bands through their approach known as “topological quantum chemistry,” thereby giving them a large database of materials, as well as the theoretical tools to look for topological flat bands.
    By employing tools ranging from analytical methods to brute-force searches, the team found all the flat band materials currently known in nature. This catalogue of flat band materials is available online https://www.topologicalquantumchemistry.fr/flatbands with its own search engine. “The community can now look for flat topological bands in materials. We have found, out of 55,000 materials, about 700 exhibiting what could potentially be interesting flat bands,” says Yuanfeng Xu, from Princeton University and the Max Planck Institute of Microstructure Physics, one of the two lead authors of the study. “We made sure that the materials we promote are promising candidates for chemical synthesis,” emphasizes Leslie Schoop from the Princeton chemistry department. The team has further classified the topological properties of these bands, uncovering what type of delocalized electrons they host.
    Now that this large catalogue is completed, the team will start growing the predicted materials to experimentally discover the potential myriad of new interacting states. “Now that we know where to look, we need to grow these materials,” says Claudia Felser from the Max Planck Institute for Chemical Physics of Solids. “We have a dream team of experimentalists working with us. They are eager to measure the physical properties of these candidates and see which exciting quantum phenomena will emerge.”
    The catalogue of flat bands, published in Nature on 30 March 2022, represents the end of years of research by the team. “Many people, and many grant institutions and universities to which we presented the project said this was too hard and could never be done. It took us some years, but we did it,” said Andrei Bernevig.
    The publication of this catalogue will not only reduce the serendipity in the search for new materials, but it will allow for large searches of compounds with exotic properties, such as magnetism and superconductivity, with applications in memory devices or in long-range dissipationless transport of power.
    Funding
    Funding for the project was primarily provided by an advanced grant of the European Research Council (ERC) at DIPC (SUPERFLAT, ERC-2020-ADG). More

  • in

    SARS-CoV-2 spike protein more stable, slower changing than earlier version

    New computational simulations of the behavior of SARS-CoV-1 and SARS-CoV-2 spike proteins prior to fusion with human cell receptors show that SARS-CoV-2, the virus that causes COVID-19, is more stable and slower changing than the earlier version that caused the SARS epidemic in 2003.
    Severe acute respiratory syndrome coronaviruses 1 and 2 (SARS-CoV-1 and SARS-CoV-2) have striking similarities, and researchers do not fully understand why the latter has been more infectious.
    The spike proteins of each, which bind to host cell angiotensin converting enzyme 2, otherwise known as the human cell receptor, have been targeted as the potential source of the different transmissibility. Understanding the mechanistic details of the spike proteins prior to binding could lead to the development of better vaccines and medications.
    The new finding does not necessarily mean that SARS-CoV-2 is more likely to bind to cell receptors, but it does mean that its spike protein has a better chance of effective binding.
    “Once it finds the cell receptor and binds to it, the SARS-CoV-2 spike is more likely to stay bound until the rest of the necessary steps are completed for full attachment to the cell and initiation of cell entry,” said Mahmoud Moradi, associate professor of chemistry and biochemistry in the Fulbright College of Arts and Sciences.
    To determine differences in conformational behavior between the two versions of the virus, Moradi’s research team performed an extensive set of equilibrium and nonequilibrium simulations of the molecular dynamics of SARS-CoV-1 and SARS-CoV-2 spike proteins, leading up to binding with cell angiotensin converting enzyme 2. The 3D simulations were done on a microsecond-level, using computational resources provided by the COVID-19 High Performance Computing Consortium.
    Equilibrium simulations allow the models to evolve spontaneously on their own time, while nonequilibrium simulations use external manipulation to induce the desired changes in a system. The former is less biased, but the latter is faster and allows for many more simulations to run. Both methodological approaches provided a consistent picture, independently demonstrating the same conclusion that the SARS-CoV-2 spike proteins were more stable.
    The models revealed other important findings, namely that the energy barrier associated with activation of SARS-CoV-2 was higher, meaning the binding process happened slowly. Slow activation allows the spike protein to evade human immune response more efficiently, because remaining in an inactive state longer means the virus cannot be attacked by antibodies that target the receptor binding domain.
    Researchers understand the importance of the so-called receptor-binding domain, or RBD, which is the critical part of a virus that allows it to dock to human cell receptors and thus gain entry into cells and cause infection. Models produced by Moradi’s team confirm the importance of the receptor-binding domain but also suggest that other domains, such as the N-terminal domain, could play a crucial role in the different binding behavior of SARS-CoV-1 and -2 spike proteins.
    N-terminal domain of a protein is a domain located at the N-terminus or simply the start of the polypeptide chain, as opposed to the C-terminus, which is the end of the chain. Though it is near the receptor-binding domain and is known to be targeted by some antibodies, function of the N-terminal domain in SARS-CoV-1 and -2 spike proteins is not completely understood. Moradi’s team is the first to find evidence for potential interaction of the N-terminal domain and the receptor binding domain.
    “Our study sheds light on the conformational dynamics of the SARS-CoV-1 and SARS-CoV-2 spike proteins,” Moradi said. “Differences in the dynamic behavior of these spike proteins almost certainly contribute to differences in transmissibility and infectivity.”
    The researchers’ study, “Prefusion Spike Protein Conformational Changes Are Slower in SARS-CoV-2 than SARS-Cov-1,” was published in Journal of Biological Chemistry.
    Story Source:
    Materials provided by University of Arkansas. Original written by Matt McGowan. Note: Content may be edited for style and length. More

  • in

    Battery-free MakeCode empowers kids to code sustainably

    Northwestern University engineers have developed the first computer coding platform that enables kids to build and program sustainable, battery-free, energy-harvesting devices.
    Called Battery-free MakeCode, the new tool is based on Microsoft MakeCode, a popular free online learn-to-code platform that introduces beginners to coding basics. The visual platform makes programming easy. Users simply drag and drop blocks of pre-made code to build games like Tetris, program devices that can count steps or make sounds, and create apps that connect sensors, screens, buttons and motors.
    Battery-free MakeCode uses an extension that automatically and invisibly transforms MakeCode into a version that supports programming electronic devices that harvest energy from ambient sources, such as vibrations, movement, radio frequency transmissions and the sun.
    As a part of a pilot program supported by the National Science Foundation, teachers at Pūʻōhala Elementary School in Kāneʻohe, Hawaii, are beginning to implement Battery-free MakeCode into their place-based, sustainability-focused STEM curricula.
    The research behind the new platform was published today (March 30) in the Proceedings of the Association for Computing Machinery on Interactive, Mobile, Wearable and Ubiquitous Technologies. The platform does not require any custom-made hardware and is available free online.
    “Across the nation, coding is becoming a standard part of curricula, and students are learning how to code earlier and earlier,” said Northwestern’s Josiah Hester, the study’s senior author. “Our hope is that as students learn to code, they also learn about concepts around energy and sustainability. With Battery-free MakeCode, we want to enable educators to instruct a new generation of programmers who understand sustainable computing and programming practices.”
    “The tech industry is likely to increase battery-free devices in the next five to 10 years,” added Christopher Kraemer, a Ph.D. candidate in Hester’s laboratory. “So there is a need to improve education around the battery-free programming domain.” More

  • in

    Lottery luck in the light of physics: Researchers present theory on the dynamics of many-particle systems

    Physicists at the University of Bayreuth are among the international pioneers of power functional theory. This new approach makes it possible for the first time to precisely describe the dynamics of many-particle systems over time. The particles can be atoms, molecules or larger particles invisible to humans. The new theory generalizes the classical density functional theory, which only applies to many-particle systems in thermal equilibrium. In the Reviews of Modern Physics, a research team led by Prof. Dr. Matthias Schmidt presents the basic features of the theory, which was significantly developed and elaborated in Bayreuth.
    A many-particle system is in thermal equilibrium when the temperature in its interior is balanced and no heat flows take place. This does not necessarily mean that the system is in a rigid state of rest. Some many-particle systems can also be compared to a lottery draw machine, which rotates at a constant speed. The balls have a lot of freedom of movement in it and jump back and forth in a disorderly fashion. In a fluid many-particle system, the particles are packed considerably more densely than in the drum, which is why they constantly collide with each other at short distances and time intervals. Essential properties of such systems can be described completely and precisely with the density functional theory — provided that a thermal equilibrium of the system is given.
    In the case of the lottery draw machine, this equilibrium is lost as soon as the uniform rotation gradually slows down and the chamber goes into reverse. Then the balls with the winning numbers roll onto a rail inside the chamber and are finally ejected. In order to record such processes precisely and without gaps, the power functional theory is needed: it translates the luck of the winners into the language of physics.
    “The classical density functional theory is a very in-depth and at the same time aesthetically appealing theory. It is able to describe and relate the often very complex processes that take place in a system during thermal equilibrium. These processes include, for example, phase transitions, crystallizations, or phenomena such as hydrophobicity, which occurs when surfaces or particles avoid contact with water. Often, such processes are of great technological or biological relevance. The elegance and power of density functional theory has spurred us in Bayreuth for the past ten years to search for ways to make many-particle systems in thermal disequilibrium accessible to an equally precise and elegant physical description. Research partners at the University of Fribourg in Switzerland have contributed to this search with important studies. For example, our joint efforts have resulted in power functional theory, which extends density functional theory to time-dependent processes,” reports Prof. Dr. Matthias Schmidt, who holds a chair in theoretical physics at the University of Bayreuth.
    The presentation of power functional theory (PFT), which has now been published, incorporates research that was primarily located in two focus areas at the University of Bayreuth: Nonlinear Dynamics and Polymer & Colloid Science. The Research Centre for Scientific Computing at the University of Bayreuth has provided substantial support and funding for many of these studies. Here, the power functional theory first proposed in 2013 was tested, further developed and applied to concrete physical problems. Among other things, the studies dealt with active particles that can self-propel, with shear and flow phenomena in colloids and liquids, or with the microscopic structure of liquids. A decisive factor for the successful development of the PFT was that the forces acting in many-body systems and their correlations with observable phenomena could be convincingly derived in this way. Here, methods of computer simulation and applications of statistical mechanics often proved indispensable.
    Story Source:
    Materials provided by Universität Bayreuth. Note: Content may be edited for style and length. More

  • in

    Researchers find topological phenomena at high, technologically relevant frequencies

    New research published in Nature Electronics describes topological control capabilities in an integrated acoustic-electronic system at technologically relevant frequencies. This work paves the way for additional research on topological properties in devices that use high-frequency sound waves, with potential applications including 5G communications and quantum information processing. The study was led by Qicheng (Scott) Zhang, a postdoc in the lab of Charlie Johnson at the University of Pennsylvania, in collaboration with the group of Bo Zhen and colleagues from Beijing University of Posts and Telecommunications and the University of Texas at Austin.
    This research builds on concepts from the field of topological materials, a theoretical framework developed by Penn’s Charlie Kane and Eugene Mele. One example of this type of material is a topological insulator, which acts as an electrical insulator on the inside but has a surface that conducts electricity. Topological phenomena are hypothesized to occur in a wide range of materials, including those that use light or sound waves instead of electricity.
    In this study, Zhang was interested in studying topological phononic crystals, metamaterials that use acoustic waves, or phonons. In these crystals, topological properties are known to exist at low frequencies in the megahertz range, but Zhang wanted to see if topological phenomena might also occur at higher frequencies in the gigahertz range because of the importance of these frequencies for telecommunication applications such as 5G.
    To study this complex system, the researchers combined state-of-the-art methodologies and expertise across theory, simulation, nanofabrication, and experimental measurements. First, researchers in the Zhen lab, who have expertise in studying topological properties in light waves, conducted simulations to determine the best types of devices to fabricate. Then, based on the results of the simulations and using high-precision tools at Penn’s Singh Center for Nanotechnology, the researchers etched nanoscale circuits onto aluminum nitride membranes. These devices were then shipped to the lab of Keji Lai at UT Austin for microwave impedance microscopy, a method that captures high-resolution images of the acoustic waves at incredibly small scales. Lai’s approach uses a commercial atomic force microscope with modifications and additional electronics developed by his lab.
    “Before this, if people want to see what’s going on in these materials, they usually need to go to a national lab and use X-rays,” Lai says. “It’s very tedious, time consuming, and expensive. But in my lab, it’s just a tabletop setup, and we measure a sample in about 10 minutes, and the sensitivity and resolution are better than before.”
    The key finding of this work is the experimental evidence showing that topological phenomena do in fact occur at higher frequency ranges. “This work brings the concept of topology to gigahertz acoustic waves,” says Zhang. “We demonstrated that we can have this interesting physics at a useful range, and now we can build up the platform for more interesting research to come.”
    Another important result is that these properties can be built into the atomic structure of the device so that different areas of the material can propagate signals in unique ways, results that were predicted by theorists but were “amazing” to see experimentally, says Johnson. “That also has its own important implications: When you’re conveying a wave along a sharp trail in ordinary systems that don’t have these topological effect, at every sharp turn you’re going to lose something, like power, but in this system you don’t,” he says.
    Overall, the researchers say that this work provides a critical starting point for progress in both fundamental physics research as well as for developing new devices and technologies. In the near term, the researchers are interested in modifying their device to make it more user-friendly and improving its performance at higher frequencies, including frequencies that are used for applications such as quantum information processing.
    “In terms of technological implications, this is something that could make its way into the toolbox for 5G or beyond,” says Johnson. “The basic technology we’re working on is already in your phone, so the question with topological vibrations is whether we can come up with a way to do something useful at these higher frequency ranges that are characteristic of 5G.”
    Story Source:
    Materials provided by University of Pennsylvania. Original written by Erica K. Brockmeier. Note: Content may be edited for style and length. More

  • in

    An automatic information extraction system for scientific articles on COVID-19

    The global bio-health research community is making a tremendous effort to generate knowledge relating to COVID-19 and SARS-CoV-2. In practice, this effort means a huge, very rapid production of scientific publications, which makes it difficult to consult and analyse all the information. That is why experts and decision-making bodies need to be provided with information systems to enable them to acquire the knowledge they need.
    This is precisely what has been explored in the VIGICOVID researchers project run by the UPV/EHU’s HiTZ Centre, the UNED’s NLP & IR group, and Elhuyar’s Artificial Intelligence and Language Technologies Unit, thanks to Fondo Supera COVID-19 funding awarded by the CRUE. In the study, under the coordination of the UNED research group they have created a prototype to extract information through questions and answers in natural language from an updated set of scientific articles on COVID-19 and SARS-CoV-2 published by the global research community.
    “The information search paradigm is changing thanks to artificial intelligence,” said Eneko Agirre, head of the UPV/EHU’s HiTZ Centre. “Until now, when searching for information on the internet, a question is entered, and the answer has to be sought in the documents displayed by the system. However, in line with the new paradigm, systems that provide the answer directly without any need to read the whole document are becoming more and more widespread.”
    In this system, “the user does not request information using keywords, but asks a question directly,” explained Elhuyar researcher Xabier Saralegi. The system searches for answers to this question in two steps: “Firstly, it retrieves documents that may contain the answer to the question asked by using a technology that combines keywords with direct questions. That is why we have explored neural architectures,” added Dr Saralegi. Deep neural architectures fed with examples were used: “That means that search models and question answering models are trained by means of deep machine learning.”
    Once the set of documents has been extracted, they are reprocessed through a question and answer system in order to obtain specific answers: “We have built the engine that answers the questions; when the engine is given a question and a document, it is able to detect whether or not the answer is in the document, and if it is, it tells us exactly where it is,” explained Dr Agirre.
    A readily marketable prototype
    The researchers are satisfied with the results of their research: “From the techniques and evaluations we analysed in our experiments, we took those that give the prototype the best results,” said the Elhuyar researcher. A solid technological base has been established, and several scientific papers on the subject have been published. “We have come up with another way of running searches for whenever information is urgently needed, and this facilitates the information use process. On the research level, we have shown that the proposed technology works, and that the system provides good results,” Agirre pointed out.
    “Our result is a prototype of a basic research project. It is not a commercial product,” stressed Saralegi. But such prototypes can be modelled easily within a short time, which means they can be marketed and made available to society. These researchers stress that artificial intelligence enables increasingly powerful tools to be made available for working with large document bases. “We are making very rapid progress in this area. And what is more, everything that is investigated can readily reach the market,” concluded the UPV/EHU researcher.
    Story Source:
    Materials provided by University of the Basque Country. Note: Content may be edited for style and length. More