Culture

The richer you are, the more likely you'll social distance, study finds

The higher a person's income, the more likely they were to protect themselves at the early stages of the Covid-19 pandemic in the United States, Johns Hopkins University economists find.

When it comes to adopting behaviors including social distancing and mask wearing, the team detected a striking link to their financial well-being. People who made around $230,000 a year were as much as 54% more likely to increase these types of self-protective behaviors compared to people making about $13,000.

"We need to understand these differences because we can wring our hands, and we can blame and shame, but in a way it doesn't matter," said Nick Papageorge, the Broadus Mitchell Associate Professor of Economics. "Policymakers just need to recognize who is going to socially distance, for how long, why and under what circumstances to give us accurate predictions of how the disease will spread and help us establish policies that will be useful."

The findings, that could contribute to more accurate predictions of how the disease will spread, appear in the latest Journal of Population Economics.

As part of a six-country survey, 1,000 people in the United States, from Texas, Florida, California and New York, were asked a series of questions in April 2020 to determine if and how their behavior had changed as Covid-19 cases were beginning to spike across the country. The resulting data includes information on income, gender and race along with unique variables relevant to the pandemic, such as work arrangements and housing quality.

The team, which included economics graduate student Matthew Zahn, found that while almost everyone changed their behavior in some way to try to stay safe, people making the most money made the most changes. The highest earners were 13% more likely to change their behaviors, 32% more likely to increase social distancing and 30% more likely to increase hand washing and mask wearing.

But the team found it was also much easier for people with more money to take extra safety measures.

Higher-income individuals were more likely to report being able to work from home and more likely to have transitioned to telework instead of losing their job. The researchers found the ability to telework emerged as a huge predictor of whether someone would social distance. Compared to somebody who continued to work, people able to telework were 24% more likely to social distance.

"The whole messaging of this pandemic is you're stuck at home teleworking, that must be really tough so here are some recipes for sourdough starter, and here's what you should catch up on Netflix," Papageorge said. "But what about the people who aren't teleworking? What are they going to do?"

The team found lower-income respondents faced increased chances of job and income losses due to the pandemic and limited access to remote work. They were also more likely to live in homes with no access to the outdoors - access to outdoor space was a very strong predictor of social distancing, the researchers found. People with access to open air at home were 20% more likely to social distance.

All of these burdens ensured those earning the least would have a harder time adopting social-distancing behaviors, which could have prolonged the pandemic, the team found. Social distancing was simply more practical, comfortable and feasible for people with more income.

"It's not shocking that if you don't live in a comfortable house you're going to be leaving your house more often. But the point we want to push is that if I'm a policy maker maybe I really need to think about opening city parks in a dense neighborhood during a pandemic. Maybe that's something that's worth the risk. This is why we want to understand these details - they can eventually suggest policies," Papageorge said.

The data showed that women were 23% more likely than men to social distance. Surprisingly, the researchers did not find any meaningful patterns between pre-existing health conditions and people's self-protective actions.

The team is now expanding this research with even more extensive surveys that look into how events such as the Black Lives Matter protests affected behavior during the pandemic, the pandemic's possible effects on addictive behaviors and, more broadly, how to build tools to better understand the unequal burdens the pandemic.

Credit: 
Johns Hopkins University

Deep learning outperforms standard machine learning in biomedical research applications

ATLANTA--Compared to standard machine learning models, deep learning models are largely superior at discerning patterns and discriminative features in brain imaging, despite being more complex in their architecture, according to a new study in Nature Communications led by Georgia State University.

Advanced biomedical technologies such as structural and functional magnetic resonance imaging (MRI and fMRI) or genomic sequencing have produced an enormous volume of data about the human body. By extracting patterns from this information, scientists can glean new insights into health and disease. This is a challenging task, however, given the complexity of the data and the fact that the relationships among types of data are poorly understood.

Deep learning, built on advanced neural networks, can characterize these relationships by combining and analyzing data from many sources. At the Center for Translational Research in Neuroimaging and Data Science (TReNDS), Georgia State researchers are using deep learning to learn more about how mental illness and other disorders affect the brain.

Although deep learning models have been used to solve problems and answer questions in a number of different fields, some experts remain skeptical. Recent critical commentaries have unfavorably compared deep learning with standard machine learning approaches for analyzing brain imaging data.

However, as demonstrated in the study, these conclusions are often based on pre-processed input that deprive deep learning of its main advantage--the ability to learn from the data with little to no preprocessing. Anees Abrol, research scientist at TReNDS and the lead author on the paper, compared representative models from classical machine learning and deep learning, and found that if trained properly, the deep-learning methods have the potential to offer substantially better results, generating superior representations for characterizing the human brain.

"We compared these models side-by-side, observing statistical protocols so everything is apples to apples. And we show that deep learning models perform better, as expected," said co-author Sergey Plis, director of machine learning at TReNDS and associate professor of computer science.

Plis said there are some cases where standard machine learning can outperform deep learning. For example, diagnostic algorithms that plug in single-number measurements such as a patient's body temperature or whether the patient smokes cigarettes would work better using classical machine learning approaches.

"If your application involves analyzing images or if it involves a large array of data that can't really be distilled into a simple measurement without losing information, deep learning can help," Plis said.. "These models are made for really complex problems that require bringing in a lot of experience and intuition."

The downside of deep learning models is they are "data hungry" at the outset and must be trained on lots of information. But once these models are trained, said co-author Vince Calhoun, director of TReNDS and Distinguished University Professor of Psychology, they are just as effective at analyzing reams of complex data as they are at answering simple questions.

"Interestingly, in our study we looked at sample sizes from 100 to 10,000 and in all cases the deep learning approaches were doing better," he said.

Another advantage is that scientists can reverse analyze deep-learning models to understand how they are reaching conclusions about the data. As the published study shows, the trained deep learning models learn to identify meaningful brain biomarkers.

"These models are learning on their own, so we can uncover the defining characteristics that they're looking into that allows them to be accurate," Abrol said. "We can check the data points a model is analyzing and then compare it to the literature to see what the model has found outside of where we told it to look."

The researchers envision that deep learning models are capable of extracting explanations and representations not already known to the field and act as an aid in growing our knowledge of how the human brain functions. They conclude that although more research is needed to find and address weaknesses of deep-learning models, from a mathematical point of view, it's clear these models outperform standard machine learning models in many settings.

"Deep learning's promise perhaps still outweighs its current usefulness to neuroimaging, but we are seeing a lot of real potential for these techniques," Plis said.

Credit: 
Georgia State University

Presidential inaugurations boost tourism, but not this year

image: While new research from West Virginia University economists finds that presidential inaugurations have gained popularity as must-see tourist events in recent years, major security threats will keep visitors away for the inauguration of President-Elect Joe Biden.

Image: 
West Virginia University

While new research from West Virginia University economists finds that presidential inaugurations have gained popularity as must-see tourist events in recent years, major security threats will keep visitors away for the inauguration of President-Elect Joe Biden.

Published in "Tourism Economics," the study, by Joshua Hall, chair and professor of economics, and economic doctoral student Clay Collins, examined the impact of the inaugurations of Barack Obama and Donald Trump on hotel occupancy in the Washington, D.C.-area.

Daily occupancy rates around the inaugurations were four-to-six times higher than the next largest event in the sample. The research team also concluded that inaugurations tend to serve as multiple-day tourist events, with hotel occupancy rates seeing positive leads and lags.

"We got the idea to look at this because of the public debate over the size of the Trump inaugural compared to the Obama one," Hall said. "While we couldn't answer that question given the data, that got us thinking about how important inaugurations are as tourist events."

The available hotel data dated back to 2010, therefore the researchers could only focus on the second Obama inaugural (2012) and the Trump one (2016). One of the complicating factors of accurately comparing the two is that the Trump inaugural also attracted scores of protestors to the area. The Women's March on Washington was also held a day after the Trump event.

Nonetheless, the economists found relatively little difference in the scale of the impact on hotel occupancy overall. One difference is that the Obama inauguration saw greater increases prior to the swearing-in ceremony, while the Trump event saw upticks after inauguration day.

Both inaugurations accounted for large increases in hotel occupancy, considerably more than any tourism-related sporting or cultural event.

Throughout his various studies, Hall has found that unique, one-off events, such as concerts or music festivals, typically serve as greater economic boosters than sporting contests, which tend to attract only local residents.

"Regular events like baseball games may draw a few people from out of the area, but most are locals," Hall said. "You need something that is a once every couple - in this case, four-- years event to get people to consider traveling.

"My sense is that interest in attending inaugurals has come from lower relative costs of attendance over time, both because of improved roads and lower airplane costs, along with rising real incomes. It is far easier to get to get to Washington, D.C. from Morgantown since the completion of Interstate 68, for example."

But, any economic impact from the hotel and tourism industries will be minimal, if any, this time around.

The FBI and Washington, D.C. officials have warned of threats ahead of Inauguration Day, Wednesday (Jan. 20), after the recent deadly riot at the U.S. Capitol. The Presidential Inaugural Committee has asked Americans to stay home and celebrate virtually. In-person attendance is limited to Congress.

As National Guard units took up residence inside the Capitol, Airbnb announced it is blocking and canceling reservations in Washington, D.C. during inauguration.

"This year will definitely be very different," Hall said. "It will not involve the multi-day celebrations of past inaugurals."

Credit: 
West Virginia University

Tetris is no longer just a game, but an algorithm, which ensures maximum room occupancy

image: Traditional allocation of rooms (above), optimal allocation (below). This example shows that two rooms become available. The image has been developed by the research group of the University of Trento

Image: 
©University of Trento

To achieve full occupancy, hotels used to rely exclusively on experience, concentration and human abilities. Then came online booking, which made the reservation collection process faster, but did not solve the risk of turning down long stays because of rooms previously booked for short stays.

To avoid overbooking (accepting more reservations than there is room for) in some cases online sales are blocked before hotels are completely booked. The solution that the University of Trento has just discovered could change the life of hotels by increasing the number of occupied rooms and, therefore, in the revenue of hotel owners.

For an average Italian hotel (50 rooms), an annual increase in turnover of between 5 and 10% is estimated.

The most compact way to combine supply and demand was found by the RoomTetris algorithm, which takes its name from the computer game that inspired it, Tetris.

The software was developed by the Lion Laboratory (Learning and Intelligent OptimizatioN) of the Department of Information Engineering and Computer Science of the University of Trento. The research team, led by Roberto Battiti and Mauro Brunato, collaborated with Filippo Battiti's local start-up Ciaomanager Srl which provided first-hand knowledge of everyday hotel management.

After the patent application, the procedure has also been the subject of an article in the international "Journal of Hospitality and Tourism Technology" (RoomTetris: an optimal procedure for committing rooms to reservations in hotels; Vol. 11 No. 4, 2020 pp. 589-602; DOI 10.1108/JHTT-08-2019-0108).

"It is a new and revolutionary method to manage the accommodation of guests in hotel rooms" comments Roberto Battiti. "We made the unexpected and surprising discovery of an excellent algorithm for assigning guests to rooms: there is no better way to do it and there is a mathematical theorem to prove it."

In short, RoomTetris finds the best solution, the ideal combination between demand and supply, optimizing room occupancy. A tile-matching game that no human mind, no matter how experienced and skilled, could do better, with the seriousness and scientific rigor of a mathematical demonstration. Battiti is proud as a researcher but also satisfied for giving some hope to a sector that, more than others, has been heavily hit by the pandemic's consequences.

"The intuition of the RoomTetris algorithm - he says - derives from the Tetris game, which is well-known among scientists and video game enthusiasts around the world. Coloured tiles of different shapes fall in the playing field and players must place them so that they do not build up, therefore they have to fit the blocks in the best way possible in the free cells".

He continues: "If the average profit of a hotel is 10-15% of the turnover, the increased room availability generated by the algorithm in the high season can increase it by a further 5-10% (depending on the average occupancy rate and the duration of the stay). With little effort (which is actually made by powerful computers in the cloud) there are certainly cases where the profitability can even double. I bet that within a few years almost all hotels will use our optimal algorithm, and that many hotel management habits will therefore change radically".

In practice, with RoomTetris hotels will no longer allocate rooms at the time of booking, but when guests arrive at the hotel, providing the optimal solution for a higher occupancy rate and increased profitability. Tests to measure the improvements in the occupancy rate compared to the traditional allocation of rooms were carried out through a hotel simulator in different areas, and including real hotels throughout Italy, from Trentino to Sicily, from Sardinia to Puglia.

"The success of RoomTetris, which is the first optimal room allocation algorithm for the hotel industry, suggests that the room allocation process can be managed by this algorithm at check-in, ensuring the best possible performance, at global level", concludes Battiti.

Credit: 
Università di Trento

Population density and virus strains will affect how regions can resume normal life

MADISON, Wis. -- As a new, apparently more transmissible version of the virus that causes COVID-19 has appeared in several countries, new research finds that the transmissibility of viral strains and the population density of a region will play big roles in how vaccination campaigns can help towns and cities return to more normal activities.

The findings suggest that directing vaccines toward densely populated counties would help to interrupt transmission of the disease. Current vaccination distribution plans don't take density into account.

Tony Ives at the University of Wisconsin-Madison and Claudio Bozzuto of the independent data research company Wildlife Analysis GmbH studied the spread of COVID-19 in the U.S. at the start of the pandemic, before people changed their behavior to avoid the disease. This let them uncover factors that may affect the transmission of COVID-19 when masking and physical distancing start to wane and behavior once again resembles the pre-pandemic normal.

"We wanted to get at two things: the first was to try and understand what the dynamics were very early in the pandemic. If we need a vaccination program in place that lets people act normally, then we need to understand the state under those conditions," says Ives, a professor of integrative biology at UW-Madison. "The second was trying to get at a fairly small spatial scale of counties instead of states."

Parsing out county-by-county data from 39 states through May 23, 2020, Ives and Bozzuto found that the higher a county's population density, the more readily SARS-CoV-2, the virus responsible for COVID-19, spread from person to person. This county-level spread is ultimately quantified in the basic reproduction number of the virus, a measure of the average number of people an infected person goes on to infect.

The researchers also found compelling evidence that viral strain matters. Ives and Bozzuto saw that the regions hosting a greater proportion of strains containing a mutation called G614 experienced greater viral spread, a finding supported by other research showing that this strain could be transmitted more readily.

While the G614 mutant is unrelated to B.1.1.7, a strain first identified in the United Kingdom that appears to be spreading more easily right now, the new study reflects the importance that viral strain can play in a local area's overall disease spread.

"We found a clear pattern in the spread rate due to different strains," says Bozzuto. "Our approach was novel because we went directly to the community level to ask, 'Can we see any patterns in the data without making assumptions about individual behavior, including strain-related transmissibility and pathogenicity?'"

To track the rate of viral spread, Ives and Bozzuto worked with the number of people who died of COVID-19 last spring. When testing was limited at the beginning of the pandemic, deaths much more accurately tracked COVID-19 transmission. As long as a relatively constant proportion of infected people die from the disease, data on how deaths increase over time will be directly proportional to the overall rate of spread.

Population density predicted a considerable amount of the difference in the rate of viral spread from county to county during the time period the researchers studied. Counties with low or moderate density did not have high rates of infection spread, though lower transmission rates do not necessarily protect a region from eventually seeing high case counts.

Location explained an even greater fraction of the spread in the researchers' model. Regions within a few hundred miles of one another had similar transmission rates. This regional similarity might have been caused in part by similar public health responses in neighboring counties.

But Ives and Bozzuto also found evidence that regional differences in viral strains explained why neighboring counties looked similar. For example, the low proportion of G614 mutants in the Northwest and Southeast was associated with lower transmission rates.

The researchers investigated several other factors -- such as prevalence of obesity and diabetes, socioeconomic status, and political affiliation -- and found that none of them contributed significantly to the rate of spread of COVID-19 at the very start of the epidemic. Although these factors may affect how susceptible individuals and populations are to complications from the disease, they didn't appear to affect the transmission of the virus from person to person.

The new findings were published in the journal Communications Biology on Jan. 5. Ives and Bozzuto say their work can help public health officials decide where vaccines would do the most good.

"Vaccination programs should consider potential spread rate in different areas. The main driver that will be important is density," says Ives. "From an epidemiological perspective, we would argue that metropolitan areas should be targeted because the level of vaccination or acquired immunity has to be higher than in largely rural areas."

Credit: 
University of Wisconsin-Madison

Stretching more effective than walking to lower high blood pressure: USask study

image: USask kinesiologist researcher Phil Chilibeck

Image: 
USask/Victoria Dinh

A new University of Saskatchewan (USask) study has found that stretching is superior to brisk walking for reducing blood pressure in people with high blood pressure or who are at risk of developing elevated blood pressure levels.

Walking has long been the prescription of choice for physicians trying to help their patients bring down their blood pressure. High blood pressure (hypertension) is a leading risk factor for cardiovascular disease and among the top preventable risk factors affecting overall mortality.

This new finding, published December 18, 2020 in the Journal of Physical Activity and Health, shows that stretching should be part of a well-rounded treatment plan for people wrestling with hypertension.

"Everyone thinks that stretching is just about stretching your muscles," said kinesiology professor Dr. Phil Chilibeck (PhD), a co-author of the study. "But when you stretch your muscles, you're also stretching all the blood vessels that feed into the muscle, including all the arteries. If you reduce the stiffness in your arteries, there's less resistance to blood flow," he said, noting that resistance to blood flow increases blood pressure.

While previous studies have shown stretching can reduce blood pressure, the USask research is the first to pit walking against stretching in a head-to-head comparison in the same group of study participants.

Chilibeck and colleagues randomly assigned 40 older men and women (mean age 61) to two groups for the eight-week study period: one did a whole-body stretching routine for 30 minutes a day, five days a week, and the other group walked briskly for the same amount of time and frequency. All participants had elevated blood pressure, or stage 1 hypertension, at the start of the study.

Before and after the study, Chilibeck and colleagues measured participants' blood pressure while they were sitting, lying down, and over 24 hours using a portable monitor--widely considered the gold standard for accurate blood pressure measurement. Stretching resulted in bigger reductions in blood pressure across all three types of measurement. The walkers did, however, lose more body fat off their waist in the eight-week study.

People who are walking to reduce their high blood pressure should continue to do so, but also add in some stretching sessions, according to Chilibeck.

"I don't want people to come away from our research thinking they shouldn't be doing some form of aerobic activity. Things like walking, biking, or cross- country skiing all have a positive effect on body fat, cholesterol levels, and blood sugar."

While the study protocol had participants stretching for 30 minutes at a time, Chilibeck suspects the same benefits can be achieved by doing a shorter routine that emphasizes the larger muscle groups in the legs, particularly the quadriceps and hamstrings. Yoga produces similar reductions in blood pressure, he said.

The beauty of stretching, said Chilibeck, is that it's so easy to incorporate into a person's daily routine. You're not at the mercy of the weather and it's easy on your joints--a big plus for people with osteoarthritis. And it doesn't require a big commitment of time, another barrier to exercise for many people. "When you're relaxing in the evening, instead of just sitting on the couch, you can get down on the floor and stretch while you're watching TV," he said.

Chilibeck and colleagues are now seeking funding to do a larger study involving more participants. They'd like to expand the scope beyond blood pressure measurement to explore some of the physiological reasons behind why stretching reduces blood pressure--such as arterial stiffness and changes in the body's nervous system resulting from stretching.

Credit: 
University of Saskatchewan

Geologic history written in garnet sand

image: Suzanne Baldwin, Thonis Family Professor, examining a gneiss, a type of metamorphic rock on a field expedition to Goodenough Island, Milne Bay Province, Papua New Guinea.

Image: 
Photo Professor Paul Fitzgerald

On a beach on a remote island in eastern Papua New Guinea, a country located in the southwestern Pacific to the north of Australia, garnet sand reveals an important geologic discovery. Similar to messages in bottles that have traveled across the oceans, sediments derived from the erosion of rocks carry information from another time and place. In this case the grains of garnet sand reveal a story of traveling from the surface to deep into the Earth (~75 miles), and then returning to the surface before ending up on a beach as sand grains. Over the course of this geologic journey, the rock type changed as some minerals were changed, and other materials were included (trapped) within the newly formed garnets. The story is preserved in garnet compositions, as well as in their trapped inclusions: solids (e.g., very rare minerals such as coesite - a high pressure form of quartz), liquids (e.g., water) and gases (e.g., CO2).

Suzanne Baldwin, Thonis Family Professor in the Department of Earth and Environmental Sciences at Syracuse University's College of Arts and Sciences, has led many field expeditions to Papua New Guinea. Her team's latest results on this tectonically active region have just been published in the prestigious journal Proceedings of the National Academy of Sciences (PNAS).

By reading the rock record researchers revealed the recycling pathway from the surface to deep within the upper mantle and then back to the surface as a result of tectonic and sedimentary processes. The compositions of that sand also hold various key components that reveal how quickly this recycling happened. In this case, transit through the rock cycle happened in less than ~10 million years. This may seem like a long time, but for these geologic processes, it is actually remarkably short.

The garnet sands are just the latest piece of the puzzle to understand the geologic evolution of this region. It is the only location on Earth where active exhumation of high- and ultrahigh- pressure metamorphic rocks is occurring during the same rock cycle that produced these metamorphic rocks. The international group of researchers, including Joseph Gonzalez '19 Ph.D. from Syracuse University (now a European Research Council postdoctoral researcher at the University of Pavia, Italy), Ph.D. student Jan Schönig and Professor Hilmar von Eynatten from the University Göttingen in Germany, and Professor Hugh Davies (formerly from the University of Papua New Guinea, now at The Australian National University), revealed how the trapped inclusions in garnet sand can be used to determine rock recycling processes within active plate boundary zones.

At active plate boundaries, like the one the team studied in eastern Papua New Guinea, converging tectonic plates slide toward each other with one plate moving underneath the other to form a subduction zone. During this process, rocks are subducted deep into the Earth. Over time, forces on the plate boundaries may change and rocks can be exhumed to the surface through a process known as lithospheric deformation. The trapped inclusions preserve a record of crustal subduction and rapid exhumation linking upper mantle and surface processes on these short geologic timescales. By applying their approach to both modern sediments and sedimentary rocks, researchers can now reveal the tempo of rock recycling processes throughout Earth's history.

Credit: 
Syracuse University

Toadlet peptide transforms into a deadly weapon against bacteria

image: The peptide uperin 3.5 is secreted by the Australian toadlet's skin. When exposed to bacterial membranes, it rapidly changes its structure and transforms into a deadly antimicrobial weapon. The pictures were taken using a transmission electron microscope (TEM) in the Electron Microscopy Centers in the Technion Department of Materials Science and Engineering and in the Department of Chemical Engineering. The cross-α atomic structure was determined by data collected at the ESRF synchrotron.

Image: 
Nir Salinas/Technion

An antibacterial peptide that turns on and off

The researchers solved the 3D molecular structure of an antibacterial peptide named uperin 3.5, which is secreted on the skin of the Australian toadlet (Uperoleia mjobergii) as part of its immune system. They found that the peptide self-assembles into a unique fibrous structure, which via a sophisticated structural adaptation mechanism can change its form in the presence of bacteria to protect the toadlet from infections. This provides unique atomic-level evidence explaining a regulation mechanism of an antimicrobial peptide.

The antibacterial fibrils on the toadlet's skin have a structure that is reminiscent of amyloid fibrils, which are a hallmark of neurodegenerative diseases, such as Alzheimer's and Parkinson's. Although amyloid fibrils have been considered pathogenic for decades, it has recently been discovered that certain amyloid fibrils can benefit the organisms that produce them, from human to microbes. For example, certain bacteria produce such fibrils to fight human immune cells.

The findings suggest that the antibacterial peptide secreted on the toadlet's skin self-assembles into a "dormant" configuration in the form of highly stable amyloid fibrils, which scientists describe as a cross-β conformation. These fibrils serve as a reservoir of potential attacker molecules that can be activated when bacteria are present. Once the peptide encounters the bacterial membrane, it changes its molecular configuration to a less compact cross-α form, and transforms into a deadly weapon. "This is a sophisticated protective mechanism of the toadlet, induced by the attacking bacteria themselves," says structural biologist Meytal Landau, the lead author of this study. "This is a unique example of an evolutionary design of switchable supramolecular structures to control activity."

Potential for future medical applications

Antimicrobial peptides are found in all kingdoms of life, and thus are hypothesised to be commonly used as weapons in nature, occasionally effective in killing not only bacteria, but also cancer cells. Moreover, the unique amyloid-like properties of the toadlet's antibacterial peptide, discovered in this study, shed light on potential physiological properties of amyloid fibrils associated with neurodegenerative and systemic disorders.

The researchers hope that their discovery will lead to medical and technological applications, e.g. development of synthetic antimicrobial peptides that would be activated only in the presence of bacteria. Synthetic peptides of this kind could also serve as a stable coating for medical devices or implants, or even in industrial equipment that requires sterile conditions.

Credit: 
European Molecular Biology Laboratory

Turn off that camera during virtual meetings, environmental study says

image: A new study estimates the approximate carbon, water and land footprints associated with each hour of data spent in popular internet apps.

Image: 
Purdue University/Kayla Wiles

WEST LAFAYETTE, Ind. -- It's not just to hide clutter anymore - add "saving the planet" to the reasons you leave the camera off during your next virtual meeting.

A new study says that despite a record drop in global carbon emissions in 2020, a pandemic-driven shift to remote work and more at-home entertainment still presents significant environmental impact due to how internet data is stored and transferred around the world.

Just one hour of videoconferencing or streaming, for example, emits 150-1,000 grams of carbon dioxide (a gallon of gasoline burned from a car emits about 8,887 grams), requires 2-12 liters of water and demands a land area adding up to about the size of an iPad Mini.

But leaving your camera off during a web call can reduce these footprints by 96%. Streaming content in standard definition rather than in high definition while using apps such as Netflix or Hulu also could bring an 86% reduction, the researchers estimated.

The study, conducted by researchers from Purdue University, Yale University and the Massachusetts Institute of Technology, is the first to analyze the water and land footprints associated with internet infrastructure in addition to carbon footprints. The findings are published in the journal Resources, Conservation & Recycling.

"If you just focus on one type of footprint, you miss out on others that can provide a more holistic look at environmental impact," said Roshanak "Roshi" Nateghi, a Purdue professor of industrial engineering, whose work looks to uncover gaps and assumptions in energy research that have led to underestimating the effects of climate change.

A number of countries have reported at least a 20% increase in internet traffic since March. If the trend continues through the end of 2021, this increased internet use alone would require a forest of about 71,600 square miles - twice the land area of Indiana - to sequester the emitted carbon, the study found.

The additional water needed in the processing and transmission of data would also be enough to fill more than 300,000 Olympic-size swimming pools, while the resulting land footprint would be about equal to the size of Los Angeles.

The team estimated the carbon, water and land footprints associated with each gigabyte of data used in YouTube, Zoom, Facebook, Instagram, Twitter, TikTok and 12 other platforms, as well as in online gaming and miscellaneous web surfing. As expected, the more video used in an application, the larger the footprints.

Because data processing uses a lot of electricity, and any production of electricity has carbon, water and land footprints, reducing data download reduces environmental damage.

"Banking systems tell you the positive environmental impact of going paperless, but no one tells you the benefit of turning off your camera or reducing your streaming quality. So without your consent, these platforms are increasing your environmental footprint," said Kaveh Madani, who led and directed this study as a visiting fellow at the Yale MacMillan Center.

The internet's carbon footprint had already been increasing before COVID-19 lockdowns, accounting for about 3.7% of global greenhouse gas emissions. But the water and land footprints of internet infrastructure have largely been overlooked in studies of how internet use impacts the environment, Madani said.

Madani teamed up with Nateghi's research group to investigate these footprints and how they might be affected by increased internet traffic, finding that the footprints not only vary by web platform, but also by the country. The team gathered data for Brazil, China, France, Germany, India, Iran, Japan, Mexico, Pakistan, Russia, South Africa, the U.K. and the U.S.

Processing and transmitting internet data in the U.S., the researchers found, has a carbon footprint that is 9% higher than the world median, but water and land footprints that are 45% and 58% lower, respectively.

Incorporating the water and land footprints of internet infrastructure painted a surprising picture for a few countries. Even though Germany, a world renewable energy leader, has a carbon footprint well below the world median, its water and land footprints are much higher. The country's energy production land footprint, for example, is 204% above the median, the researchers calculated.

Purdue graduate students Renee Obringer, Benjamin Rachunok and Debora Maia-Silva performed the calculations and data analysis in collaboration with Maryam Arbabzadeh, a postdoctoral research associate at MIT. The estimates are based on publicly available data for each platform and country, models developed by Madani's research group and known values of energy use per gigabyte of fixed-line internet use.

The estimates are rough, the researchers say, since they're only as good as the data made available by service providers and third parties. But the team believes that the estimates still help to document a trend and bring a more comprehensive understanding of environmental footprints associated with internet use.

"These are the best estimates given the available data. In view of these reported surges, there is a hope now for higher transparency to guide policy," Nateghi said.

Credit: 
Purdue University

Neither 'meniscal' nor 'mechanical' symptoms predict findings on knee arthroscopy

January 14, 2021 - Orthopaedic surgeons have traditionally been taught that certain types of knee symptoms indicate damage to specialized structures called the menisci. But these "meniscal" and "mechanical" symptoms do not reflect what surgeons will find at knee arthroscopy, reports a study in The Journal of Bone & Joint Surgery. The journal is published in the Lippincott portfolio in partnership with Wolters Kluwer.

Both types of symptoms are strongly related to the overall amount of cartilage damage in the knee joint - but not to the presence of meniscal tears, according to the new research by Elizabeth G. Matzkin, MD, and colleagues of Harvard Medical School. They write, "Our study adds strong evidence to refute the current dogma that meniscal pathology is the primary driver of traditionally defined 'meniscal' and 'mechanical' knee symptoms."

New evidence 'questions diagnostic utility' of knee symptoms
Sometimes called the "shock absorbers" of the knee, the menisci are a pair of horseshoe-shaped pieces of cartilage that help to cushion the knee joint. Among orthopaedic surgeons, conventional wisdom is that "mechanical" symptoms like knee locking or catching are caused by meniscal tears, which can be treated with an arthroscopic surgical procedure. However, a growing body of evidence suggests that other types of injuries can also cause these types of knee symptoms.

Other recent studies have suggested that "meniscal" symptoms like knee grinding, popping, or clicking are better indicators of meniscal tears. Dr. Matzkin and colleagues questioned whether either of these two types of symptoms are truly indicators of meniscus tear within the knee joint - and whether they are of value in surgical decision-making.

The researchers analyzed the relationship between these knee symptoms and the findings of knee arthroscopy in 565 patients. All patients underwent knee arthroscopy performed by Dr. Matzkin between 2012 and 2019. The goal was to determine whether "mechanical" or "meniscal" symptoms were related to the presence of meniscal tears, cartilage damage, or other types of joint disease identified via arthroscopy.

The most frequently reported symptoms included worse pain with knee pivoting/twisting (85 percent), knee grinding (74 percent), and knee catching (63 percent). Arthroscopy revealed unstable meniscal tears in 12 percent of patients, stable tears in 66 percent, and no tear in 21 percent.

Meniscal tear patterns and meniscal damage were found to be unrelated to individual symptoms of knee grinding or pain with pivoting.

In contrast, the overall severity of cartilage damage was significantly associated with knee symptoms - especially catching and grinding. Knee catching was present in 49 percent of patients with no cartilage damage versus 78 percent of those with damage in all three compartments of the knee.

Because meniscal tears and cartilage damage often occur together, the study suggests that structural cartilage damage is the "primary driver" of knee symptoms. "Our data align with the observation that cartilage damage and meniscal tears are both manifestations of overall degenerative knee disease," Dr. Matzkin and coauthors write.

The findings "further question the diagnostic utility" of mechanical and meniscal symptoms and their use "as the indications for knee arthroscopy," the researchers add. Dr. Matzkin and colleagues believe that future studies should avoid these misleading terms, and instead use specific descriptors of knee symptoms. The researchers plan further studies to assess how specific types of symptoms are related to outcomes after arthroscopic surgery.

Click here to read "Meniscal and Mechanical Symptoms Are Associated with Cartilage Damage, Not Meniscal Pathology."
DOI: 10.2106/JBJS.20.01193

Credit: 
Wolters Kluwer Health

Lack of managers keeps India's businesses small

In today's economy, American businesses often tap into professional management to grow, but most firms in India and other developing countries are family owned and often shun outside managers. A new study co-authored by Yale economist Michael Peters explores the effects that the absence of outside professional management has on India's businesses and the country's economy.

The study, published in the American Economic Review, uses a novel model to compare the relationship between the efficiency of outside managers and firm growth in the United States and India. It shows that the lack of managerial delegation factors significantly into why businesses in India tend to stay small and has wider implications on the country's economy, constraining innovation, economic growth, and per capita income.

"There's been growing evidence that managerial services might be the key missing input for many firms in poor countries," said Peters, associate professor of economics in Yale's Faculty of Arts and Sciences. "Our analysis confirms that the absence of managerial delegation is a significant factor in why successful Indian businesses fail to grow, which reduces the overall productivity of the country's economy."

By growing, companies generate employment, promote innovation, and contribute to a countries' economic productivity, the researchers say.

In developed countries, family-owned firms such as Walmart, Ford Motor Co., and the Lego Group, which all emerged from humble beginnings, grew into corporate behemoths, with hundreds of thousands of employees, by delegating key operations to outside managers. But businesses in developing countries rarely hire managers outside the owners' families, the researchers note.

Peters and his co-authors, Ufuk Akcigit of the University of Chicago and Harun Alp of the University of Pennsylvania, focused their analysis on this disparity between developed and developing countries' firm sizes. They created a quantitative model that centers the role of managerial delegation in firm growth. It incorporated plant-level data from the United States and India and was calibrated to recognize that business in India might face higher barriers to growth, such as having less access to start-up capital, than U.S. businesses.

In India, more than 9 out of 10 of manufacturing businesses have fewer than four employees, and those small firms account for more than half of total employment. In contrast, two-thirds of U.S. manufacturing employment is concentrated in establishments with at least 100 employees, and only one-third of firms have fewer than four employees, according to the study.

The researchers found that India's economy suffers from "a lack of selection" -- the process of creative destruction through which successful businesses expand while unproductive firms close or are swallowed up by competitors -- allowing unproductive businesses to survive because successful businesses do not expand. Their analysis showed that the low productivity of outside managers in India, which they estimate to be substantially lower than in the United States, is one cause of the lack of selection and has negative consequences for India's economy.

If Indian businesses used outside managers as efficiently as U.S. businesses, it would boost economic productivity in India and increase the country's per capita income by about 11%, according to the study.

The study found a strong complementary relationship between the quality of outside managers and other factors affecting firm growth, such as access to capital or credit. Increasing the efficiency of managerial delegation to U.S. standards would increase average firm size by 3%, the researchers said. In contrast, if U.S. businesses had to operate with management practices common in India, the country's average firm would shrink by about 15%. This disparity between the U.S. and India shows that the productivity of outside managers is not the only determinant of firm growth and that other forces prevent successful Indian businesses from expanding, Peters explained.

"The complementarity between the efficiency of delegating managerial tasks and other aspects affecting firm growth, such as access to credit, is one of our key results," said Peters, who is affiliated with Yale's Economic Growth Center. "For improvements to managerial quality to have a large effect, other factors hindering growth must be addressed. If you repair a punctured tire, you still can't drive if your other tires are flat."

Credit: 
Yale University

Reverse engineering 3D chromosome models for individual cells

image: A new computational technique that uses heat map data to reverse engineer highly detailed models of chromosomes and researchers have uncovered new information about the close spatial relationships that chromatin folding creates between genes.

Image: 
UIC

Genome analysis can provide information on genes and their location on a strand of DNA, but such analysis reveals little about their spatial location in relation to one another within chromosomes -- the highly complex, three-dimensional structures that hold genetic information.

Chromosomes resemble a fuzzy "X" in microscopy images and can carry thousands of genes. They are formed when DNA winds around proteins -- called histones -- which are further folded into complexes called chromatin, which make up individual chromosomes.

Knowing which genes are located in spatial proximity within the chromatin is important because genes that are near each other generally work together.

Now, researchers at the University of Illinois Chicago report on a computational technique that uses heat map data to reverse engineer highly detailed models of chromosomes. Through this work, the researchers have uncovered new information about the close spatial relationships that chromatin folding creates between genes that can be highly distant from one another along DNA strands.

Their findings are published in the journal Nature Communications.

"Folding of the chromatin brings genes that are far away from each other into close proximity. If we know that certain groups of genes are spatial neighbors because of this folding, that tells us they most likely work together to drive processes such as the development of immunity, or even more fundamental processes like development or cell differentiation," said Jie Liang, UIC Richard and Loan Hill Professor of Bioengineering and a corresponding author on the paper. "This is important for better understanding these processes or development of new therapeutics to prevent or treat cancer and other diseases."

Liang and his colleagues developed a way to reverse engineer the complex structures of individual chromosomes using information from a process called Hi-C. Hi-C generates heat maps based on probabilities reflecting which genes are most likely to be spatially close to one another. These heat maps can provide approximate three-dimensional information on how chromosomes are organized, but because they are based on genetic material from multiple cells, the maps represent average likelihoods of proximity between genes, not exact locations.

Liang and colleagues looked at Hi-C heat maps of chromosomes from cells of fruit fly embryos, which have only eight chromosomes. They used these heat maps together with new advanced computational methods to generate extremely detailed three-dimensional maps of the chromosomes of individual cells.

"For the first time, we are able to produce single-cell models that accurately represent genetic spatial relationships within chromosomes," Liang said. "With these models, we can uncover rich biological patterns and answer basic biological questions about three-dimensional structural changes chromosomes undergo to cause stem cells to develop into different tissues, and how malfunctions in these processes lead to diseases such as cancer."

Credit: 
University of Illinois Chicago

Borrowing from birds, experts reduce search times for novel high-entropy alloys to seconds

image: Computational materials scientists at Ames Laboratory developed an evolutionary algorithm, using a hybrid version of a computer program called Cuckoo Search (CS), which mimics the brood parasite behavior of cuckoo birds.

Image: 
U.S. Department of Energy, Ames Laboratory

Computational materials science experts at the U.S. Department of Energy's Ames Laboratory enhanced an algorithm that borrows its approach from the nesting habits of cuckoo birds, reducing the search time for new high-tech alloys from weeks to mere seconds.

The scientists are investigating a type of alloys called high-entropy alloys, a novel class of materials that are highly sought after for a host of unusual and potentially beneficial properties. They are lightweight in relation to their strength, fracture-resistant, highly corrosion and oxidation resistant, and stand up well in high-temperature and high-pressure environments -- making them attractive materials for aerospace industry, space exploration, nuclear energy, and defense applications.

While the promise of these materials is great, they present major difficulties to scientists attempting to search for and customize them for use in technologies. Because these alloys are constructed of five or more different elements, they are expensive and difficult to develop and search experimentally, making an Edison-like approach a nonstarter. With so many ingredients, and so many different ways to construct them, there are nearly endless permutations of recipes for their design. Among literally billions of options, how do researchers narrow their search to a few excellent potential candidates for an application?

The answer in this case is an evolutionary algorithm, using a hybrid version of a computer program developed ten years ago, called Cuckoo Search (CS). Cuckoo birds are brood parasites, laying their eggs in the nest of a host bird such that they end up rearing the bigger, stronger cuckoo chick as one of its own.

"This 'survival of the fittest' strategy from the behavior of birds is the idea behind Cuckoo Search," said Duane Johnson, a computational materials scientist at Ames Laboratory. Each egg represents a possible solution, competing to be the best solution in any given nest in a fixed number of possible nests. The best solution of each nest competes against other nests, until the best solution is found.

The Ames Laboratory team put a twist on the Cuckoo Search, which greatly speeded up the process of locating ideal alloys or the best "egg" within a huge number of possibilities. The original CS takes advantage of a mathematical concept called Lévy flight, which computational theorists use to their advantage in searching extremely large data sets. But, while this method works for large data sets, the Ames Lab team found that pairing another mathematical concept, a Monte Carlo algorithm, with Lévy flight, greatly reduced the time to achieving optimal candidates for high-entropy alloys, providing optimal models almost on the fly.

"With the model-building bottleneck eliminated, computational design can be performed that is currently impractical, said Johnson, "As our hybrid CS is problem-agnostic, it offers application in optimization in many diverse fields."

Credit: 
DOE/Ames National Laboratory

New PTSD biotypes enables improved tests, sheds light on divergent treatments efficacy

Researchers from the PTSD Systems Biology Consortium, led by scientists from the Walter Reed Army Institute of Research, have identified distinct biotypes for post-traumatic stress disorder, the first of their kind for any psychological disorder. "These biotypes can refine the development of screening tools and may explain the varying efficacy of PTSD treatments", said Dr. Marti Jett, leader of the consortium and WRAIR chief scientist.

Publishing their work in Molecular Psychiatry in a manuscript first authored by WRAIR's Dr. Ruoting Yang, researchers used blood tests from male, combat-exposed veterans across a three year period to identify two PTSD biotypes, G1--characterized by mild, inherent co-morbidities typical of PTSD--and G2--which includes more severe symptoms typical of PTSD and report more physical distress--with differing genetic markers and underlying mechanisms of disease. Building on previously published work using machine learning, led by Dr. Francis J. Doyle III, dean of Harvard University's School of Engineering and Applied Sciences and computational lead of the PTSD Systems Biology Consortium, findings were expanded and validated with two additional veteran cohorts and an active-duty cohort.

PTSD diagnosis has long been complicated by a reliance on self-reporting of patient symptoms, particularly the underreporting of signs of distress due to perceived stigma. "These findings help overcome that gap, using data that link objective molecular and physiological measures with PTSD biotypes as a screening tool for early indicators of distress and to avert full, chronic PTSD," explained Dr. Charles Marmar, chair of New York University Langone Health's Department of Psychiatry and clinical lead of the PTSD Systems Biology Consortium.

Additionally, one PTSD medication is currently FDA-approved for use in military personnel and is thought to be approximately 50% effective; clinical trials for other medications are further limited in efficacy. "These data set the stage for physicians to link treatments to specific biotypes, providing a blueprint for targeted therapeutics and better patient outcomes," said Dr. Kerry Ressler, consortium member and chief scientific officer of McLean Hospital.

Researchers with the PTSD Systems Biology Consortium, a network of government and academic laboratories, plan to continue their research to further identify and validate PTSD biotypes to develop better screening tools, including a test to biotype military personnel with probable PTSD symptoms in field settings away from clinicians. Additionally, future studies are planned to incorporate biotyping into clinical trials for PTSD therapeutics currently in development.

Credit: 
Walter Reed Army Institute of Research

Cancer models created by mechanical engineers offer new insight into tumor growth

TROY, N.Y. -- As cancer and tumor cells move inside the human body, they impart and are subject to mechanical forces. In order to understand how these actions affect cancer cell growth, spread, and invasion, a team of engineers at Rensselaer Polytechnic Institute is developing new models that mimic aspects of the mechanical environment within the body, providing new insight into how and why tumors develop in certain ways.

In research published today in Integrative Biology, a team of engineers from Rensselaer developed an in vitro -- in the lab -- lymphatic vessel model to study the growth of tumor emboli, collections of tumor cells within vessels that are often associated with increased metastasis and tumor recurrence.

"The growth of tumor emboli is important for the spread and metastasis of certain types of cancers. For example, inflammatory breast cancer has this growth pattern of just massive spread of emboli growing within dermal lymphatic vessels, in the breast, and it's very aggressive in that way," said Kristen Mills, an assistant professor of mechanical engineering at Rensselaer, who led this research. "This growth of tumor emboli hasn't really been studied very much."

Previous tumor models in the lab have used unconstrained spheroids of tumor cells that don't allow researchers to study the mechanical interactions that truly happen within the physical boundaries of a tubular-shaped vessel. To more closely mimic what happens in the body, Mills and her team modeled a lymphatic vessel in the lab using a small channel in a gel. They placed these channels within either stiff or soft bioinert gel, in order to mimic the constraints that tumor emboli may encounter in stiff, diseased tissue, or soft, healthy tissue.

To incorporate the varying growth behaviors of cells, the team used breast and colon cancer cell aggregates to model the emboli, one which is more aggressive, the other less.

Researchers found that the model of a stiff tumor environment constrained both types of tumor emboli to the cylindrical channel, causing rapid growth of the emboli along the vessel. But, they found that the growth of the cancer cells was different, based on their type, in the softer matrix model, which mimics healthy colon or breast tissue.

The aggressive cells were not affected by the presence of the open channel at all and grew as a sphere, bulging into the matrix, while the less aggressive cells first grew along the channel for several days before bulging into the matrix. The researchers linked the differences in growth to the force generation capability of the cancer cells. Independent measurements indicated that the aggressive cancer cells were capable of exerting significantly more force than the less aggressive cells.

Growth along a vessel or channel is problematic, Mills said, because the cells within the tumor have constant access to life-sustaining nutrients through the vessel wall. When the tumor cells grow in a sphere and begin to bulge, Mills said, the cells in the center of that mass get farther and farther away from nutrients until they are cut off and eventually die. This information could be critical for therapeutic design and prescription.

"We think this is important because it somehow suggests that, if the tumor tissue surrounding these vessels has been abnormally stiffened through disease, or if -- for example -- the breast tissue is already denser or stiffer, you might really see much more growth within the vessels than with a softer tissue," said Mills, who is also a member of the Center for Biotechnology and Interdisciplinary Studies (CBIS) at Rensselaer.

Jonathan Kulwatno, a graduate student in biomedical engineering at Rensselaer, was first author on this paper. Mills and Kulwatno collaborated with researchers from Mount Sinai.

This unique, mechanical engineering approach to studying cancer is a hallmark of the transformative work being done within CBIS, where researchers like Mills can collaborate across disciplines with other engineers and scientists working in similar areas.

"Professor Mills' engineering approach to studying cancer is laying a critical foundation, where interdisciplinary approaches are used to understand the mechanobiology of cancer with an eventual therapeutic goal," said Deepak Vashishth, the director of CBIS.

As Mills and her team continue their research, she said, they look for new ways to make these models more complex without losing the ability to isolate critical mechanical mechanisms.

Credit: 
Rensselaer Polytechnic Institute