American kids with food allergies are more than twice as likely to have autism spectrum disorder as kids without, a study of national health data finds. The population-based finding adds to experimental evidence that there may be a connection between false steps or overreactions by the immune system and the neurodevelopmental disorder.

Researchers looked only for an association between allergies and autism spectrum disorder, or ASD, among a total of 199,520 children ages 3 to 17 surveyed from 1997 to 2016 as part of the U.S. National Health Interview Survey. The study was not designed to discover what may be behind the link.
The team found that, out of 1,868 children with autism, 216 had a food allergy — or about 11 percent. By comparison, only about 4 percent of children without autism had a food allergy, the researchers report online June 8 in JAMA Network Open. Kids with autism were also more likely to have respiratory or skin allergies like eczema than kids without autism.

The number of children with autism has more than doubled since 2000, to a prevalence of 16.8 per 1,000 kids. Meanwhile, the number of kids with food allergies rose from 3.4 percent in 1997–1999 to 5.1 percent in 2009–2011.

It is unknown whether developing food allergies may contribute to the development of autism, or vice versa, or if something else is causing both, says study coauthor and epidemiologist Wei Bao of the University of Iowa’s College of Public Health in Iowa City. “The causes of ASD remain unclear,” he says.

Past work in mice and people has pointed to a possible connection between different immune system disorders and autism. There is a higher risk of autism for children with a family history of type 1 diabetes, or with a history on mom’s side of the family of rheumatoid arthritis or celiac disease. Mice that developed a food allergy displayed behaviors characteristic of autism, such as repetitive behaviors and less frequent social interaction, a 2014 study published in Behavioral Brain Research found.
The new finding supports the idea “that different manifestations of immune abnormalities occur in individuals with ASD,” says Christopher McDougle, director of the Lurie Center for Autism at Massachusetts General Hospital in Boston, who wrote a commentary accompanying the study. Food, respiratory and skin allergies are common in the general population, he says, but having these allergies “doesn’t mean your child is going to develop ASD.”

With the aid of a particle accelerator, scientists are bringing back ghosts from the past, revealing portraits hidden underneath the tarnished surface of two roughly 150-year-old silver photographic plates.

Researchers used an accelerator called a synchrotron to produce strong, but nondamaging beams of X-rays to scan the damaged photographs, called daguerreotypes, and map their chemical composition. This allowed chemist Madalena Kozachuk of Western University in London, Canada, and colleagues to trace mercury deposits in the plates and create digital copies of the hidden images, the team reports June 22 in Scientific Reports. One image revealed a woman; the other, a man who had been completely obscured by tarnish.
An early form of photography, daguerreotypes were popular from the 1840s through the 1860s. Photographers crafted the images by making a silver-coated copper plate and treating it with iodine vapor to generate a light-sensitive surface. Subjects sat still for the several minutes required to expose the plate and create an image. Then photographers treated the plate with heated mercury vapor and a gold solution to develop the image, forming tiny silver-mercury-gold particles where light struck the plate during the exposure process. These particles make up the image, reflecting white light. Lighter parts of an image, such as the woman’s hands and collar, have a higher density of these particles.

The researchers used mercury to map the contours of the original images, because that metal remains fixed in place under years of cloudy tarnish. The scans revealed where the original particles were, letting researchers reconstruct the image.

Scanning the roughly 8-by-7-centimeter daguerreotypes, provided by the National Gallery of Canada, was time-consuming, taking about eight hours per square centimeter.
Synchrotrons had never been used to image daguerreotypes before, so Kozachuk didn’t know what to expect. “When the image became apparent, it was jaw-dropping,” she says. “I squealed when the first face popped up.”

The machines are expensive, and getting time to work on them can be difficult. But Kozachuk hopes her research will enable museums with damaged daguerreotypes to reveal more of these faded faces.

Mouse mothers can transmit stress signals to offspring, changing the way the pups’ bodies and brains develop. Some of these stress messages get delivered during birth, scientists suggest July 9 in Nature Neuroscience.

Researchers suspected that vaginal microbes from stressed-out moms could affect male pups in ways that leave them vulnerable to stress later in life (SN: 12/14/2013, p. 13). But earlier studies hadn’t demonstrated whether those microbes, picked up during birth, actually caused some of the changes seen in offspring, or if other aspects of life in utero were to blame.
Tracy Bale of the University of Maryland School of Medicine in Baltimore and colleagues subjected pregnant mice to stressful trials that included smelling the scent of a fox for an hour, listening to unusual sounds overnight and being restrained in a tube for 15 minutes. Other pregnant mice didn’t experience these stressors. Then, researchers delivered pups by cesarean section, so that the pups weren’t exposed to their mothers’ community of vaginal microorganisms, or microbiome. After delivery, researchers dosed the pups with vaginal fluid taken from stressed or unstressed mothers.

For male pups not exposed to stress in the womb, vaginal microbes from a stressed mother changed the amount of certain kinds of gut bacteria. (Just as in earlier studies, female pups didn’t show effects of their mothers’ stress.) When those male pups were older, being restrained led them to release more of the stress hormone corticosteroid than mice dosed with microbiota from unstressed moms. And in the brains of adult mice that had experienced chronic stress, genes involved in metabolism and the development of nerve cells behaved differently depending on whether early microbes came from stressed or unstressed mothers.

But some stress effects didn’t seem to depend on the mothers’ microbiomes, results that suggest those effects came from being stressed in utero. For example, compared with pups that weren’t stressed in utero, pups exposed to stress in the womb had higher levels of certain immune cells, as well as key gut genes behaving differently — both possible signs of inflammation.

America is built on lead. Networks of aging pipes made from the bluish-gray metal bring water into millions of U.S. homes. But when lead, a poison to the nervous system, gets into drinking water — as happened in Flint, Mich. — the heavy metal can cause irreparable harm (SN: 3/19/16, p. 8). In The Poisoned City, journalist Anna Clark provides a thorough, nuanced account of the public health disaster in Flint — one that, she argues, was magnified by government malfeasance and decades of systemic racism.
Trouble first began in April 2014. To save the cash-strapped city some money, Flint’s emergency manager switched the city’s source of water from Detroit’s water system, which drew from Lake Huron, to one that tapped the Flint River. But the city’s water treatment program didn’t include corrosion control, which the Michigan Department of Environmental Quality said wasn’t necessary — a violation of federal law. The result: Corroded pipes leached lead into drinking water.

Residents, forced to use the brown, smelly tap water, developed rashes and lost clumps of hair. Twelve people died from Legionella bacteria, which the corrosive water dislodged from pipes, and dozens more were sickened. Despite residents’ complaints, as well as an independent analysis that found higher-than-allowable lead levels, state officials insisted that the water was safe, even when their own internal records showed it was not. “Anyone who is concerned about lead in the drinking water in Flint can relax,” said one spokesperson for the Michigan Department of Environmental Quality.

That’s when one of the book’s heroes, pediatrician Mona Hanna-Attisha, enters Clark’s story. About 18 months after Flint switched to its new water source, the percentage of children under age 5 with high blood-lead levels nearly doubled from 2.1 to 4 percent, Hanna-Attisha discovered after taking a close look at Flint kids’ medical records. (Hanna-Attisha’s own account of her experiences, What the Eyes Don’t See, was published in June.)

Faced with mounting evidence that became hard to ignore, Gov. Rick Snyder negotiated a switch back to Detroit’s water system in October 2015, declaring a state of emergency a few months later. Meanwhile, taps in Flint were retrofitted with filters as the long, slow process of replacing pipes began. The Michigan National Guard trucked in bottled water.
Readers who followed this crisis as it unfolded will still learn plenty in The Poisoned City. Clark goes into exquisite detail explaining not only what happened, but also why it happened. A history of racist housing, education and hiring practices precipitated the city’s “debt, dysfunctional urban policy, disappearing investment, disintegrating infrastructure, and a compromised democratic process,” she writes. The evidence linking these factors to the water crisis is compelling. Anyone wanting to dig deeper can refer to the book’s exhaustive bibliography.

Overall, Clark does a masterful job weaving together history, science and rigorous reporting to tell Flint’s story, which served as a “wake-up call” for cities around the country. A 2016 investigation by the Natural Resources Defense Council found that more than 5,300 water systems across the United States were in violation of federal lead rules. And it’s not just cities that are affected, Clark notes. Rural America is vulnerable, too. But replacing America’s lead pipes is an expensive proposition. By some estimates, removing lead service lines alone would cost somewhere between $30 billion and $1 trillion.

Four years after Flint’s water crisis began, residents are still grappling with lingering effects: potentially lifelong health problems, ruined pipes that will take years to fix and zero trust in government. In April, Michigan declared Flint’s water safe. But people who live in the city are not convinced. And Hanna-Attisha has urged the state to continue Flint’s bottled water program until all of the lead service lines are replaced.

Buy The Poisoned City from Amazon.com. Science News is a participant in the Amazon Services LLC Associates Program. Please see our FAQ for more details.

Astronomers have found 12 more moons around Jupiter, and one is really weird. While 11 orbit in the same direction as their nearest neighbors, one doesn’t, potentially putting it on a fatal collision course.

“It’s driving down the highway on the wrong side of the road,” says planetary scientist Scott Sheppard of the Carnegie Institution for Science in Washington, D.C.

Sheppard and colleagues found the moons while looking for something else entirely: a putative planet that could exist beyond the orbit of Neptune, known colloquially as Planet Nine (SN: 7/23/16, p. 7). During a survey in 2017 of the most distant objects in the solar system using the Victor Blanco 4-meter telescope in Chile, Jupiter happened to be visible in the same area of sky that the team was searching during one of its observing runs. “Might as well kill two birds with one stone,” Sheppard thought.
The researchers found a dozen objects moving around the sun at the same rate as Jupiter. Follow-up observations confirmed the moons’ existence and orbits: two inner moons that orbit in the same direction that Jupiter spins, nine outer moons that orbit the planet in the opposite direction and one oddball traveler. The researchers announced two of the moons in 2017 and the remaining 10 on July 16.

The motions of all but the oddball are normal for Jovian moons, which now number a whopping 79. Scientists think that’s because the inner moons formed from a disk of gas and dust that orbited the giant planet in the solar system’s early days, similar to how the planets formed around the sun (SN: 5/12/18, p. 28). The outer moons were probably free-floating space rocks captured when they came too close, and their opposite orbit was set by the direction that they approached Jupiter from.

But one moon broke the mold. This rock, which the team calls Valetudo for the Roman goddess of health and hygiene, is tiny, only about a kilometer across. It orbits in the same direction as Jupiter’s spin, but alongside the farther-out retrograde moons. As a result, Valetudo is probably doomed to collide with one or more of the other moons someday. The researchers are still calculating when, but they expect it to occur sometime between 100 million and a billion years from now.
Valetudo may be the last remnant of a bigger object that has already withstood several collisions, or of a family of moons that has since been smashed to smithereens. “It’s probably the largest surviving member, if not the only one,” Sheppard says.

Such nonconformist satellites are not rare, notes planetary scientist David Jewitt of UCLA, who was not involved in the new work. “But they are very interesting, because we know that they have been captured by their host planets, but we don’t know how, or from where,” he says. Figuring out what oddballs like Valetudo are made of could help nail those details down.

There’s a new clique among quantum particles in a semiconductor.

Electrons and positively charged holes in the material’s atomic lattice band together to create a tight-knit posse dubbed a collexon, researchers report July 26 in Communications Physics. This new class of quasiparticle — a quantum clan that acts like a single subatomic particle — could help researchers better understand semiconductors, which are essential to most modern electronics.

The collexon is similar to a quasiparticle known as an exciton, a pairing of an electron and a hole (SN: 5/17/14, p. 5). While these pairs go it alone in excitons, electron-hole duos in collexons join forces with the surrounding sea of electrons, Christian Nenstiel, a physicist at the Technical University of Berlin, and colleagues report.
The researchers made this discovery when they inserted germanium atoms into a gallium nitride semiconductor, and zapped the material with a laser to see how it emits light. In similar experiments, emissions from excitons faded as the number of impurities, such as the germanium atoms, increased. But this time, at high concentrations of the introduced atoms, light shone at different wavelengths than seen with the excitons. The team deduced that large numbers of wandering electrons, introduced by the germanium, helped stabilize excitons to form the new type of quasiparticle.

It’s too early to predict applications, says study coauthor Gordon Callsen of the École Polytechnique Fédérale de Lausanne in Switzerland. The discovery instead suggests that researchers underestimate interactions among ensembles of particles in semiconductors. “Lots of interesting physics is still waiting for us,” he says.

Scrolling through a news feed often feels like playing Two Truths and a Lie.

Some falsehoods are easy to spot. Like reports that First Lady Melania Trump wanted an exorcist to cleanse the White House of Obama-era demons, or that an Ohio school principal was arrested for defecating in front of a student assembly. In other cases, fiction blends a little too well with fact. Was CNN really raided by the Federal Communications Commission? Did cops actually uncover a meth lab inside an Alabama Walmart? No and no. But anyone scrolling through a slew of stories could easily be fooled.

We live in a golden age of misinformation. On Twitter, falsehoods spread further and faster than the truth (SN: 3/31/18, p. 14). In the run-up to the 2016 U.S. presidential election, the most popular bogus articles got more Facebook shares, reactions and comments than the top real news, according to a BuzzFeed News analysis.

Before the internet, “you could not have a person sitting in an attic and generating conspiracy theories at a mass scale,” says Luca de Alfaro, a computer scientist at the University of California, Santa Cruz. But with today’s social media, peddling lies is all too easy — whether those lies come from outfits like Disinfomedia, a company that has owned several false news websites, or a scrum of teenagers in Macedonia who raked in the cash by writing popular fake news during the 2016 election.
Most internet users probably aren’t intentionally broadcasting bunk. Information overload and the average Web surfer’s limited attention span aren’t exactly conducive to fact-checking vigilance. Confirmation bias feeds in as well. “When you’re dealing with unfiltered information, it’s likely that people will choose something that conforms to their own thinking, even if that information is false,” says Fabiana Zollo, a computer scientist at Ca’ Foscari University of Venice in Italy who studies how information circulates on social networks.

Intentional or not, sharing misinformation can have serious consequences. Fake news doesn’t just threaten the integrity of elections and erode public trust in real news. It threatens lives. False rumors that spread on WhatsApp, a smartphone messaging system, for instance, incited lynchings in India this year that left more than a dozen people dead.

To help sort fake news from truth, programmers are building automated systems that judge the veracity of online stories. A computer program might consider certain characteristics of an article or the reception an article gets on social media. Computers that recognize certain warning signs could alert human fact-checkers, who would do the final verification.

Automatic lie-finding tools are “still in their infancy,” says computer scientist Giovanni Luca Ciampaglia of Indiana University Bloomington. Researchers are exploring which factors most reliably peg fake news. Unfortunately, they have no agreed-upon set of true and false stories to use for testing their tactics. Some programmers rely on established media outlets or state press agencies to determine which stories are true or not, while others draw from lists of reported fake news on social media. So research in this area is something of a free-for-all.

But teams around the world are forging ahead because the internet is a fire hose of information, and asking human fact-checkers to keep up is like aiming that hose at a Brita filter. “It’s sort of mind-numbing,” says Alex Kasprak, a science writer at Snopes, the oldest and largest online fact-checking site, “just the volume of really shoddy stuff that’s out there.”
Substance and style
When it comes to inspecting news content directly, there are two major ways to tell if a story fits the bill for fraudulence: what the author is saying and how the author is saying it.

Ciampaglia and colleagues automated this tedious task with a program that checks how closely related a statement’s subject and object are. To do this, the program uses a vast network of nouns built from facts found in the infobox on the right side of every Wikipedia page — although similar networks have been built from other reservoirs of knowledge, like research databases.
In the Ciampaglia group’s noun network, two nouns are connected if one noun appeared in the infobox of another. The fewer degrees of separation between a statement’s subject and object in this network, and the more specific the intermediate words connecting subject and object, the more likely the computer program is to label a statement as true.

Take the false assertion “Barack Obama is a Muslim.” There are seven degrees of separation between “Obama” and “Islam” in the noun network, including very general nouns, such as “Canada,” that connect to many other words. Given this long, meandering route, the automated fact-checker, described in 2015 in PLOS ONE, deemed Obama unlikely to be Muslim.

But estimating the veracity of statements based on this kind of subject-object separation has limits. For instance, the system deemed it likely that former President George W. Bush is married to Laura Bush. Great. It also decided George W. Bush is probably married to Barbara Bush, his mother. Less great. Ciampaglia and colleagues have been working to give their program a more nuanced view of the relationships between nouns in the network.

Verifying every statement in an article isn’t the only way to see if a story passes the smell test. Writing style may be another giveaway. Benjamin Horne and Sibel Adali, computer scientists at Rensselaer Polytechnic Institute in Troy, N.Y., analyzed 75 true articles from media outlets deemed most trustworthy by Business Insider, as well as 75 false stories from sites on a blacklist of misleading websites. Compared with real news, false articles tended to be shorter and more repetitive with more adverbs. Fake stories also had fewer quotes, technical words and nouns.

Based on these results, the researchers created a computer program that used the four strongest distinguishing factors of fake news — number of nouns and number of quotes, redundancy and word counts — to judge article veracity. The program, presented at last year’s International Conference on Web and Social Media in Montreal, correctly sorted fake news from true 71 percent of the time (a program that sorted fake news from true at random would show about 50 percent accuracy). Horne and Adali are looking for additional features to boost accuracy.

Verónica Pérez-Rosas, a computer scientist at the University of Michigan in Ann Arbor, and colleagues compared 240 genuine and 240 made-up articles. Like Horne and Adali, Pérez-Rosas’ team found more adverbs in fake news articles than in real ones. The fake news in this analysis, reported at arXiv.org on August 23, 2017, also tended to use more positive language and express more certainty.
Computers don’t necessarily need humans to tell them which aspects of fake articles give these stories away. Computer scientist and engineer Vagelis Papalexakis of the University of California, Riverside and colleagues built a fake news detector that started by sorting a cache of articles into groups based on how similar the stories were. The researchers didn’t provide explicit instructions on how to assess similarity. Once the program bunched articles according to likeness, the researchers labeled 5 percent of all the articles as factual or false. From this information, the algorithm, described April 24 at arXiv.org, predicted labels for the rest of the unmarked articles. Papalexakis’ team tested this system on almost 32,000 real and 32,000 fake articles shared on Twitter. Fed that little kernel of truth, the program correctly predicted labels for about 69 percent of the other stories.

Adult supervision
Getting it right about 70 percent of the time isn’t nearly accurate enough to trust news-vetting programs on their own. But fake news detectors could offer a proceed-with-caution alert when a user opens a suspicious story in a Web browser, similar to the alert that appears when you’re about to visit a site with no security certificate.

In a similar kind of first step, social media platforms could use misinformation watchdogs to prowl news feeds for questionable stories to then send to human fact-checkers. Today, Facebook considers feedback from users — like those who post disbelieving comments or report that an article is false — when choosing which stories to fact-check. The company then sends these stories to the professional skeptics at FactCheck.org, PolitiFact or Snopes for verification. But Facebook is open to using other signals to find hoaxes more efficiently, says Facebook spokesperson Lauren Svensson.

No matter how good computers get at finding fake news, these systems shouldn’t totally replace human fact-checkers, Horne says. The final call on whether a story is false may require a more nuanced understanding than a computer can provide.

“There’s a huge gray scale” of misinformation, says Julio Amador Diaz Lopez, a computer scientist and economist at Imperial College London. That spectrum — which includes truth taken out of context, propaganda and statements that are virtually impossible to verify, such as religious convictions — may be tough for computers to navigate.

Snopes science writer Kasprak imagines that the future of fact-checking will be like computer-assisted audio transcription. First, the automated system hammers out a rough draft of the transcription. But a human still has to review that text for overlooked details like spelling and punctuation errors, or words that the program just got wrong. Similarly, computers could compile lists of suspect articles for people to check, Kasprak says, emphasizing that humans should still get the final say on what’s labeled as true.

Eyes on the audience
Even as algorithms get more astute at flagging bogus articles, there’s no guarantee that fake news creators won’t step up their game to elude detection. If computer programs are designed to be skeptical of stories that are overly positive or express lots of certainty, then con authors could refine their writing styles accordingly.

“Fake news, like a virus, can evolve and update itself,” says Daqing Li, a network scientist at Beihang University in Beijing who has studied fake news on Twitter. Fortunately, online news stories can be judged on more than the content of their narratives. And other telltale signs of false news might be much harder to manipulate — namely, the kinds of audience engagement these stories attract on social media.
Juan Cao, a computer scientist at the Institute of Computing Technology at the Chinese Academy of Sciences in Beijing, found that on China’s version of Twitter, Sina Weibo, the specific tweets about a certain piece of news are good indicators for whether a particular story is true. Cao’s team built a system that could round up the tweets discussing a particular news event, then sort those posts into two groups: those that expressed support for the story and those that opposed it. The system considered several factors to gauge the credibility of those posts. If, for example, the story centered on a local event that a user was geographically close to, the user’s input was seen as more credible than the input of a user farther away. If a user had been dormant for a long time and started posting about a single story, that abnormal behavior counted against the user’s credibility. By weighing the ethos of the supporting and the skeptical tweets, the program decided whether a particular story was likely to be fake.

Cao’s group tested this technique on 73 real and 73 fake stories, labeled as such by organizations like China’s state-run Xinhua News Agency. The algorithm examined about 50,000 tweets about these stories on Sina Weibo, and recognized fake news correctly about 84 percent of the time. Cao’s team described the findings in 2016 in Phoenix at an Association for the Advancement of Artificial Intelligence conference. UC Santa Cruz’s de Alfaro and colleagues similarly reported in Macedonia at last year’s European Conference on Machine Learning and Principles and Practices of Knowledge Discovery in Databases that hoaxes can be distinguished from real news circulating on Facebook based on which users like these stories.

Rather than looking at who’s reacting to an article, a computer could look at how the story is getting passed around on social media. Li and colleagues studied the shapes of repost networks that branched out from news stories on social media. The researchers analyzed repost networks of about 1,700 fake and 500 true news stories on Weibo, as well as about 30 fake and 30 true news networks on Twitter. On both social media sites, Li’s team found, most people tended to repost real news straight from a single source, whereas fake news tended to spread more through people reposting from other reposters.

A typical network of real news reposts “looks much more like a star, but the fake news spreads more like a tree,” Li says. This held true even when Li’s team ignored news originally posted by well-known, official sources, like news outlets themselves. Reported March 9 at arXiv.org, these findings suggest that computers could use social media engagement as a litmus test for truthfulness, even without putting individual posts under the microscope.Juan Cao, a computer scientist at the Institute of Computing Technology at the Chinese Academy of Sciences in Beijing, found that on China’s version of Twitter, Sina Weibo, the specific tweets about a certain piece of news are good indicators for whether a particular story is true. Cao’s team built a system that could round up the tweets discussing a particular news event, then sort those posts into two groups: those that expressed support for the story and those that opposed it. The system considered several factors to gauge the credibility of those posts. If, for example, the story centered on a local event that a user was geographically close to, the user’s input was seen as more credible than the input of a user farther away. If a user had been dormant for a long time and started posting about a single story, that abnormal behavior counted against the user’s credibility. By weighing the ethos of the supporting and the skeptical tweets, the program decided whether a particular story was likely to be fake.

Cao’s group tested this technique on 73 real and 73 fake stories, labeled as such by organizations like China’s state-run Xinhua News Agency. The algorithm examined about 50,000 tweets about these stories on Sina Weibo, and recognized fake news correctly about 84 percent of the time. Cao’s team described the findings in 2016 in Phoenix at an Association for the Advancement of Artificial Intelligence conference. UC Santa Cruz’s de Alfaro and colleagues similarly reported in Macedonia at last year’s European Conference on Machine Learning and Principles and Practices of Knowledge Discovery in Databases that hoaxes can be distinguished from real news circulating on Facebook based on which users like these stories.

Rather than looking at who’s reacting to an article, a computer could look at how the story is getting passed around on social media. Li and colleagues studied the shapes of repost networks that branched out from news stories on social media. The researchers analyzed repost networks of about 1,700 fake and 500 true news stories on Weibo, as well as about 30 fake and 30 true news networks on Twitter. On both social media sites, Li’s team found, most people tended to repost real news straight from a single source, whereas fake news tended to spread more through people reposting from other reposters.

A typical network of real news reposts “looks much more like a star, but the fake news spreads more like a tree,” Li says. This held true even when Li’s team ignored news originally posted by well-known, official sources, like news outlets themselves. Reported March 9 at arXiv.org, these findings suggest that computers could use social media engagement as a litmus test for truthfulness, even without putting individual posts under the microscope.
Truth to the people
When misinformation is caught circulating on social networks, how best to deal with it is still an open question. Simply scrubbing bogus articles from news feeds is probably not the way to go. Social media platforms exerting that level of control over what visitors can see “would be like a totalitarian state,” says Murphy Choy, a data analyst at SSON Analytics in Singapore. “It’s going to become very uncomfortable for all parties involved.”

Platforms could put warning signs on misinformation. But labeling stories that have been verified as false may have an unfortunate “implied truth effect.” People might put more trust in any stories that aren’t explicitly flagged as false, whether they’ve been checked or not, according to research posted last September on the Social Science Research Network by human behavior researchers Gordon Pennycook, of the University of Regina in Canada, and David Rand at Yale University.

Rather than remove stories, Facebook shows debunked stories lower in users’ news feeds, which can cut a false article’s future views by 80 percent, company spokesperson Svensson says. Facebook also displays articles that debunk false stories whenever users encounter the related stories — though that technique may backfire. In a study of Facebook users who like and share conspiracy news, Zollo and colleague Walter Quattrociocchi found that after conspiracists interacted with debunking articles, these users actually increased their activity on Facebook conspiracy pages. The researchers reported this finding in June in Complex Spreading Phenomena in Social Systems.

There’s still a lot of work to be done in teaching computers — and people — to recognize fake news. As the old saying goes: A lie can get halfway around the world before the truth has put on its shoes. But keen-eyed computer algorithms may at least slow down fake stories with some new ankle weights.

Humans’ gift of gab probably wasn’t the evolutionary boon that scientists once thought.

There’s no evidence that FOXP2, sometimes called “the language gene,” gave humans such a big evolutionary advantage that it was quickly adopted across the species, what scientists call a selective sweep. That finding, reported online August 2 in Cell, follows years of debate about the role of FOXP2 in human evolution.

In 2002, the gene became famous when researchers thought they had found evidence that a tweak in FOXP2 spread quickly to all humans — and only humans — about 200,000 years ago. That tweak swapped two amino acids in the human version of the gene for ones different than in other animals’ versions of the gene. FOXP2 is involved in vocal learning in songbirds, and people with mutations in the gene have speech and language problems. Many researchers initially thought that the amino acid swap was what enabled humans to speak. Speech would have given humans a leg up on competition from Neandertals and other ancient hominids.
That view helped make FOXP2 a textbook example of selective sweeps. Some researchers even suggested that FOXP2 was the gene that defines humans, until it became clear that the gene did not allow humans to settle the world and replace other hominids, says archeaogeneticist Johannes Krause at the Max Planck Institute for the Science of Human History in Jena, Germany, who was not involved in the study. “It was not the one gene to rule them all.”

The FOXP2 sweep theory first ran into trouble in 2008, when researchers discovered that Neandertals also had the two amino acid tweaks (SN Online: 11/14/08). That meant the change happened at least 700,000 years ago, before humans and Neandertal became separate branches of the hominid family tree. Then in 2009, some members of the 2002 team that originally reported the sweep presented new evidence showing that the two amino acid change wasn’t what swept to evolutionary prominence after all.

“That was sad, but that’s how it is,” says Wolfgang Enard, an evolutionary geneticist at Ludwig-Maximilians-Universität in Munich, who was involved in both the 2002 and 2009 studies. Still, there were hints that other genetic variants in and around FOXP2 might have been involved in a sweep, so the debate continued.
Evolutionary and population geneticist Elizabeth Atkinson of Massachusetts General Hospital in Boston and colleagues decided to revisit the gene’s evolution “to see if FOXP2 ’s story held up using modern techniques,” she says. The researchers conducted a similar statistical analysis of patterns of genetic variation in FOXP2 as was done in the 2002 study. But this time the team studied more people, especially more people of African descent, and used data from the entire genome.
In a selective sweep, one pattern of genetic variants around a gene becomes much more common than other versions of the gene until nearly everyone has the popular version. When considering all the people in their study together, the researchers picked up the same statistical signal for a selective sweep as Enard’s group had. But when Atkinson’s team examined Africans separately from Europeans and Asians, the signs of a sweep were erased.

That result reflects what happened in human history, Krause says.

When humans migrated out of Africa, certain versions of genes were carried with the migrants and other forms were left behind in Africa. The version of FOXP2 that left with the migrants became more common as the migrant population grew. Atkinson’s team identified the statistical signal as being from population growth, rather than a selective sweep, by looking at changes elsewhere in the genome. If FOXP2 were getting swept, it would be the only gene sending the statistical signal. Instead, other parts of the genome scored similarly to FOXP2 on the statistical test.

The finding doesn’t mean that changes in FOXP2 weren’t important for language evolution, says Kirk Lohmueller, a population geneticist at UCLA. But geneticists may have to rethink some assumptions about how the evolution of species works. Selective sweeps were thought to be a major way that natural selection — the process that drives evolution — altered species. But these and other results suggest that selective sweeps were not very common in human evolution.

Many of the traits associated with being human, including speech and language, are controlled by multiple genes, so no one gene may have given a sweep-worthy boost. Or perhaps a speech and language sweep happened, but so long ago that its signal is too weak to pick up now, Lohmueller says.

Health officials have confirmed 12 cases of rat lungworm disease in the continental United States since January 2011 — including six patients who had not traveled abroad but still contracted the illness caused by a parasite endemic to tropical regions in Asia and Hawaii.

While the disease can be mild, it can become extreme and cause severe neurological problems. In most of the new cases, patients complained of headache, fever, weakness and symptoms consistent with meningitis, the U.S. Centers for Disease Control and Prevention reports in the Aug. 3 Morbidity and Mortality Weekly Report.
The disease is also known as angiostrongyliasis, after the parasitic roundworm Angiostrongylus cantonensis whose larvae hatch in the lungs of rats and then are expelled in the rodents’ excrement. At that point, the larvae can be picked up by snails and slugs, and then passed along to humans if the snails and slugs are eaten raw. On July 30, researchers added centipedes to the list of creatures that can transmit the disease to humans, after a Chinese woman and her son contracted the disease in 2012 after eating raw centipedes bought at a market (SN Online: 7/30/18).

More than half of the recent U.S. cases involved patients who had eaten raw vegetables, likely inadvertently consuming a snail or slug, and at least one case involved a toddler who ate slugs while playing. Of the six cases confirmed as originating within the country, four were reported from Texas, one from Tennessee and one from Alabama.

“We don’t know exactly the source of the infection,” CDC epidemiologist Susan Montgomery says. “Fresh produce really should be washed thoroughly and carefully.”

The CDC also confirmed 18 cases and reported three more probable cases in 2017 alone in Hawaii.
Montgomery and her team only tracked cases sent to CDC labs for testing, meaning there may be more undiagnosed cases.

Dim light emanating from the purgatory between galaxies could illuminate the most shadowy constituents of the cosmos.

Dark matter, an unidentified type of particle that interacts gravitationally but otherwise shuns normal matter, lurks throughout clusters of galaxies. Because the elusive substance emits no light, it’s difficult to pin down how it is distributed, even though it makes up the majority of a cluster’s mass. But a feeble glow known as intracluster light could reveal dark matter’s whereabouts, researchers suggest July 30 at arXiv.org. The intermediary could eventually help scientists get a better handle on what dark matter is and how it behaves.
Galaxy clusters grow by swallowing up additional galaxies. As galaxies are assimilated, they can be torn apart and their stars scattered. It’s those stars that produce intracluster light. And where there’s intracluster light, there’s dark matter, the team found. “The shape of this very diffuse light traces very nicely the shape of the total mass of the cluster,” says study coauthor Mireia Montes, an astrophysicist at the University of New South Wales in Sydney. Once stripped from their galaxies, the stars are tugged by the dark matter’s gravity and thereby end up concentrated in the same regions as it resides.

Typically, scientists use an effect called gravitational lensing to map dark matter (SN: 10/17/15, p. 24). A galaxy cluster’s mass acts like a lens, bending light from more distant objects. By measuring that bending, scientists can see how the dark matter’s mass is distributed within the cluster. However, “that is an incredibly hard measurement to make,” says astrophysicist Stacy Kim of Ohio State University, who was not involved with the research. Measuring intracluster light is easier, Kim says, but teasing out the faint light is still challenging, requiring extended observations with a powerful telescope.
Scientists sometimes use another proxy for dark matter: X-rays emitted by hot gas within a cluster. But if a galaxy cluster has recently merged with another, collisions between gas clouds mean that the X-rays will be displaced from the dark matter. So a map of the matter made using X-rays might be skewed. The stars that produce intracluster light don’t have that problem, because they don’t get knocked off course in cluster mergers the way colliding gas clouds do.

In a study of six galaxy clusters, each observed with NASA’s Hubble Space Telescope, the researchers found that the distribution of intracluster light matched up well with the dark matter mass distribution as determined by gravitational lensing. The X-ray distribution didn’t match, because the six clusters had each been roiled by a recent smashup with another cluster. The team hopes to study more clusters to see if the match between dark matter and intracluster light holds up.

By measuring intracluster light, scientists could “perhaps learn something about the nature of dark matter,” says astrophysicist James Bullock of the University of California, Irvine, who was not involved with the research.

If the distribution of dark matter in galaxy clusters doesn’t agree with standard theoretical predictions, that could reveal new properties of the unidentified particles. For example, dark matter might be interacting with itself (SN: 7/7/18, p. 9). So having a new method to trace out dark matter is great, Bullock says. “This is definitely promising.”