Katherine J. Wu is a Digital Editor at NOVA and Story Collider producer. She holds a Ph.D. in Microbiology and Immunobiology from Harvard University. Previously, she served as a AAAS Mass Media Fellow at Smithsonian magazine.

Koch Foundation

NOVA Science Trust Logo

Can you suffer the symptoms of jet lag without even stepping foot outside your current time zone? Much to the chagrin of the reluctant traveler, the answer is apparently yes.And, if you’re reading this, you probably already have. A hefty chunk of the human population—especially those living in particularly plugged-in parts of the world—maintains a different sleep schedule on weekdays than on weekends, staying up later and sleeping in on their days off from work.This phenomenon of so-called <a href="https://www.cell.com/current-biology/fulltext/S0960-9822(12)00325-9?_returnURL=https%3A%2F%2Flinkinghub.elsevier.com%2Fretrieve%2Fpii%2FS0960982212003259%3Fshowall%3Dtrue">social jet lag</a> is so ingrained in us that, according to <a href="https://www.cell.com/current-biology/fulltext/S0960-9822(18)31345-9">a new study</a>, it’s even reflected in how we tweet—and the offset could be taking a serious toll on our health.“This study is indicative of the amazing amount of data that’s out there,” says <a href="https://www.imp.med.uni-muenchen.de/research/molecular_chronobiology-merrow/index.html">Martha Merrow</a>, who studies the biology of sleep at the Ludwig Maximilian University of Munich’s Institute of Medical Physiology, but did not participate in the research. “It shows the power of social media in revealing aspects of human behavior to us.”
 

 
 <div class="ad"></div>
 

 
 When left to their own devices, our bodies’ internal clocks naturally respond to sunlight, the evolutionarily-ancient cue of the day’s beginning and end. But in the last few centuries, humans have retreated indoors for both business and pleasure, diminishing our exposure to the outside environment and gradually uncoupling our physiologies from the Sun. Now, rather than being governed by available daylight, our schedules are shackled to the constantly changing demands of society. And as a result, there’s a serious mismatch between the times we’re asleep on weekdays versus weekends.“There’s this persistent idea that light, because it’s not tangible, doesn’t have an effect on your body,” says study author <a href="http://oliviawalch.com/">Olivia Walch</a>, a mathematician and sleep researcher at the University of Michigan. “And that’s not true. Light is this huge factor in sleep that doesn’t get the credit it deserves.”
 <figure class="richtext-image full-width"><img alt="StockSnap pixabay.jpg" class="richtext-image full-width" height="532" width="800"
 src="https://www.pbs.org/wgbh/nova/media/images/StockSnap_pixabay.width-800.jpg"/><figcaption>For most of human existence, natural sunlight has driven our sleep-wake cycles. But now, most people&#x27;s schedules are severely uncoupled from the Sun. Photo Credit: StockSnap, Pixabay<div class="credit"></div></figcaption>
 </figure>
 Social jet lag tends to go hand-in-hand with chronic sleep-deprivation, which itself is a public health issue that costs the United States <a href="https://www.rand.org/pubs/research_reports/RR1791.html">up to $411 billion</a>—over 2 percent of our gross domestic product—each year. But for most individuals, social jet lag also entails being active when one should be asleep and vice versa, putting people in all-too-frequent states of physiological disorientation. The effects of social jet lag spill over to many other aspects of health and wellbeing: Toiling through this weekly flux has been tied to <a href="https://www.cell.com/current-biology/fulltext/S0960-9822(16)30333-5">heavier drinking</a> and <a href="https://www.dovepress.com/social-jetlag-in-health-and-behavioral-research-a-systematic-review-peer-reviewed-fulltext-article-CPT">smoking</a>, as well as <a href="https://linkinghub.elsevier.com/retrieve/pii/S0960-9822(12)00325-9">obesity</a>, <a href="https://www.dovepress.com/social-jetlag-in-health-and-behavioral-research-a-systematic-review-peer-reviewed-fulltext-article-CPT">depression</a>, and even <a href="https://www.nature.com/articles/s41598-018-23044-8">decreased academic performance</a>.Not everyone experiences social jet lag in the same way. On the one hand, each person’s internal clock is unique. Some people are early risers, while others are night owls, for example—and it’s often the latter group that struggles the most transitioning between the slog of early weekday mornings and their more natural nocturnal pattern on weekends. Even the duration of someone’s morning commute or the latitude of one’s hometown can affect the ways in which an individual experiences, and thus suffers from, sleep schedule variations.Gathering this kind of data can be tough through traditional mail-and-return surveys, especially considering the staggering diversity of culture, climate, and careers in the United States alone. So a team of researchers, led by biophysicist <a href="https://biophysics.uchicago.edu/the-students/eugene_leypunskiy/">Eugene Leypunskiy</a> and biologist <a href="http://rustlab.uchicago.edu/people.html">Michael Rust</a> of the University of Chicago, looked to the modern era’s best proxy for human behavior: the Internet—more specifically, the social media juggernaut of Twitter.
 

 
 <div class="related"></div>
 

 
 The new report, <a href="https://www.cell.com/current-biology/fulltext/S0960-9822(18)31345-9">published today</a> in the journal Current Biology, tracks every publicly available, geographically-tagged tweet from the beginning of 2012 to the end of 2013—two years of data from about 246,000 Twitter users in 1,521 counties across the entire United States. “We weren’t interested in looking at what people were tweeting, but when they were tweeting,” Rust explains.The researchers found that, across the country, tweets—like many Americans—rose slowly in the morning, increased steadily to a peak in the afternoon or evening, then dipped back down at typical bedtimes. Lengthy lulls in Twitter activity indicated when people were probably asleep. And, as expected, these tweeting troughs shifted to later intervals on weekends relative to weekdays. Across counties, the average amount of “Twitter social jet lag”—or the amount of time people were shifting their sleep schedules between work and free days—clocked in at about 75 minutes.When the researchers mapped tweeting patterns across time zones, they discovered striking differences based on geography. For one, latitude and longitude both had an effect on social media habits. In general, tweeters in the American Southwest took longer breaks at bedtime than their Central U.S. and Northeastern counterparts and had less social jet lag overall. The westernmost counties within time zones, however, experienced more pronounced social jetlag—which may have to do with the fact that the Sun rises later in local time in these regions.
 <figure class="richtext-image full-width"><img alt="kansas city traffic Thomanication flickr.jpg" class="richtext-image full-width" height="530" width="800"
 src="https://www.pbs.org/wgbh/nova/media/images/kansas_city_traffic_Thomanication_flickr.width-800.jpg"/><figcaption>Something as subtle as the length of a person’s morning commute can affect the severity of social jet lag. Individuals with long, early commutes may not be sleeping for as long as they need to, or be waking up much earlier than they naturally would. Photo Credit: Thomanication, flickr<div class="credit"></div></figcaption>
 </figure>
 Across counties, certain lifestyle habits also seemed to predict a higher degree of social jet lag, including earlier average start times for commutes and the proportion of the population involved in shift work—both of which may indicate less flexibility in weekday schedules. Pronounced Twitter social jet lag also coincided with higher obesity rates, mirroring <a href="https://linkinghub.elsevier.com/retrieve/pii/S0960-9822(12)00325-9">previous findings</a>.Additionally, Twitter social jet lag ebbed and flowed with the seasons. The sleep schedule mismatch was most severe in early autumn and late winter, but mellowed out during the summer. When the researchers examined the relationship closer, they found that a common contributor to this pattern was the occurrence of summer breaks from school, during which some families seemed shift their weekday sleep cycles later to more closely match their (likely more biologically natural) weekend schedules. The same rules appeared to apply to school holidays spiked intermittently throughout the rest of the year.“The amount of social jet lag can be thought of as a proxy for how on vacation these places are,” Walch explains.Not all counties experienced reprieve from social jet lag in the summer, however. Certain locales felt the same amount of mismatch year-round—and these counties also tended to be more urban, and host residents that attended college, commuted late in the day, and were less likely to smoke or be obese.
 

 
 <div class="newsletter"></div>
 

 
 Because counties vary in not only their physical location, but also countless cultural characteristics, Rust says, it’s still difficult to disentangle exactly what factors contribute the most to social jet lag place by place. “We were looking at 1500 counties in the US,” Leypunskiy adds, “and every single one tells its own story.”But Merrow suggests it’s likely that culture plays a role. For instance, the tweeters of Orange County, California appear to have longer nighttime lulls in Twitter activity and milder social jet lag than Lafayette County, Louisiana—but the two counties lie at similar latitudes, indicating that the drivers behind behavior may be more complicated than sunlight exposure.Fortunately, there may still be a silver lining to the plight of social jet lag. <a href="https://www.mcn.uni-muenchen.de/members/regular/roenneberg/index.html">Till Roenneberg</a>, a sleep biologist at the Ludwig Maximilian University of Munich’s Institute of Medical Physiology, advises trying to tailor our schedules to our natural rhythms—which inevitably differ from person to person. Night owls, for instance, may benefit from nudging their work schedules later—a strategy that could help close the gap between weekday and weekend sleep habits without compromising total rest. But of course, not everyone has this flexibility. According to Roenneberg, society has yet to wake up to the idea of respecting sleep.An easier fix? Simply being a little more aware of our exposure to light at both appropriate and inappropriate times, Walch says. If your goal is to fall asleep earlier at night, you might try <a href="https://www.scientificamerican.com/article/q-a-why-is-blue-light-before-bedtime-bad-for-sleep/">minimizing evening phone time</a> (or, at the very least, use an app that will filter out the stuff most likely to mess with our internal clocks: <a href="https://www.health.harvard.edu/staying-healthy/blue-light-has-a-dark-side">short wavelength blue light</a>). Or take it a step further and avoid bright electronics entirely in the hours before bedtime—even if it means forsaking that final midnight tweet.
 <figure class="richtext-image full-width"><img alt="Tookapic Pexels.jpg" class="richtext-image full-width" height="531" width="800"
 src="https://www.pbs.org/wgbh/nova/media/images/Tookapic_Pexels.width-800.jpg"/><figcaption>Exposure to artificial light late at night can negatively affect sleep-wake cycles by tricking the body&#x27;s internal clock. As a result, many of us may be active when we should be asleep, and asleep when we should be active. Photo Credit: Tookapic, Pexels<div class="credit"></div></figcaption>
 </figure>
 “Social jet lag is something people experience all the time and don’t even realize it,” Leypunskiy says. “And I think if people want to change it, they have the means to do it… but it’s probably not a one-size-fits-all thing.”In the future, Rust hopes to zoom in on the biology of sleep at an individual level. Aggregations of tweets can only tell us about the average activity of a whole county, he explains. But gathering information from something like a Fitbit or other personal health monitor—with appropriate consent—could shed light on each person’s behaviors, patterns, and preferences.Even a future analysis of the content of tweets could be informative about the culture of social jet lag, says <a href="http://cs.iit.edu/~culotta/">Aron Culotta</a>, a computer scientist who conducts social media analysis at the Illinois Institute of Technology. “That’s one huge advantage of social media,” he explains. “You get to explore the ‘why’ behind a behavior.”However, Culotta cautions that, as with any work that draws from social media, findings from a subset of the population may not always capture the entire picture. Not everyone’s on Twitter—and there could be bias in the way people tweet on weekdays versus weekends. Nonetheless, he praised the current study for its “rigorous approach” to social jet lag research.For Rust, this study has already showcased the astounding diversity of humankind. “Many of us live in bubbles and see people who live similar lives to us,” he says. “But the way people live their lives and structure their time is very different—it isn’t like there’s a generic American.”

Many people maintain a different sleep schedule on weekdays than on weekends, with potentially serious health consequences—but scientists can now track this form of “jet lag” using Twitter. Photo Credit: pxhere

A two-year analysis of nearly a quarter of a million Twitter users reveals how Americans’ social lives are taking a toll on their sleep schedules.

Scientists Track “Social Jet Lag” Sleeping Habits With Twitter Data

It takes a lot to feed a <a href="http://www.pbs.org/big-blue-live/animals/blue-whale/">blue whale</a>. One of these super-sized swimmers can consume as much as 8,000 pounds of food—about the weight of a Hummer H2—on a daily basis. Which means blue whales need to be exquisitely equipped for some serious eating.The secret to this behemoth’s success? Baleen: a hyper-efficient filter-feeding system found exclusively in whales, composed of a comb-like series of overlapping plates with the texture of fingernails. After taking a massive gulp of ocean water, <a href="http://www.pbs.org/wnet/nature/ocean-giants-whale-fact-sheet/7567/">baleen whales</a>—including blue whales and <a href="https://www.pbs.org/video/life-reef-humpback-whales/">humpbacks</a>—expel unwanted liquid through this sieve, leaving behind a nutrient-dense mouthful of tiny, freshly-strained prey like krill and plankton that are easily swallowed whole.But scientists continue to grapple with the evolutionary origins of nature’s biggest oral colander. Baleen whales’ ancestors had teeth—and many of the details behind the transition to baleen, which preserves poorly in the fossil record, remain mysterious.Now, with the re-examination of a 33-million-year-old fossil, researchers are filling in a piece of this long-standing puzzle. According to <a href="https://www.cell.com/current-biology/fulltext/S0960-9822(18)31414-3">a new study</a> published today in the journal Current Biology, some whales jettisoned their teeth entirely before the acquisition of baleen, and in the interim, adopted a third method of feeding: suctioning up their food like giant vacuum cleaners.
 

 
 <div class="ad"></div>
 

 
 “If this is right, this fossil fills in a missing spot on the [whale family] tree,” says <a href="https://www.uprovidence.edu/bio/jessica-martin/">Jessica Martin</a>, a paleontologist at the University of Providence who was not involved in the research. “This might change the way we look at the big picture of this transition.”The fossil in question, newly anointed Maiabalaena nesbittae, was first unearthed in Oregon in the 1970s, but spent decades in the collections of the Smithsonian’s National Museum of Natural History still shrouded in rock. “This fossil’s been languishing for years,” says study author <a href="https://paleobiology.si.edu/staff/individuals/pyenson.html">Nicholas Pyenson</a>, the Museum’s curator of fossil marine mammals. “It’s amazing to finally be able to weigh in on it.”At about 33 million years old, M. nesbittae likely swam the seas of the early Oligocene—a period of rapid and vast geologic change. It was around this time that the southern tip of what’s now South America broke free of Antarctica, rejiggering ocean currents and cooling marine temperatures worldwide. Though it’s still unclear exactly how this transition affected ocean biodiversity, if there was a paleontological moment for a lineage of whales to ditch its dentition, it very well could have been this.
 <figure class="richtext-image full-width"><img alt="Carlos Peredo - NMNH-2018-03422.jpg" class="richtext-image full-width" height="265" width="800"
 src="https://www.pbs.org/wgbh/nova/media/images/Carlos_Peredo_-_NMNH-2018-03422.width-800.jpg"/><figcaption>The upper jaw and skull of Maiabalaena nesbittae, an extinct whale species that lived 33 million years ago. M. nesbittae was toothless, and may have been a suction feeder that gave rise to modern baleen whales. Photo Credit: Smithsonian Institution<div class="credit"></div></figcaption>
 </figure>
 Which makes M. nesbittae a potential watershed whale. Only one precious specimen of the species has been uncovered to date, and the fossil is incomplete. Luckily, most of what’s left involves the skull—which is key to studying how this species managed its massive, yet often microscopic, meals. To interrogate this fossil’s feeding forte, a team of scientists led by Carlos Peredo, a palaeobiologist at George Mason University, decided to reopen the 40-year-old case on M. nesbittae.One thing was quickly obvious to Peredo. “I looked for teeth, and it had no teeth,” he explains with a laugh. “That was the easy one.” Teeth, unlike baleen, are hardened and encased in enamel, and thus far more likely to preserve as fossils. Even if its chompers had somehow been lost to time, the roof of M. nesbittae’s mouth also appeared to be structurally inhospitable to sockets that might otherwise house true dentition. This made it unlikely that this whale was munching on large-bodied prey, à la its toothier ancestors.
 

 
 <div class="related"></div>
 

 
 Next, the team used CT scanning to more closely analyze the skull’s internal architecture without compromising the fossil. As there was little point in searching for ephemeral baleen, Peredo and his colleagues instead searched for attachment sites from which the baleen might have been suspended like vertical window blinds. They discovered the upper jaw was thin and narrow, without enough surface area to accommodate the gums from which baleen plates might grow. To the researchers, this was a clear strike against filter feeding.The researchers believe M. nesbittae utilized another option entirely on the road to baleen: simply slurping prey out of the surrounding waters through the sheer force of suction. Although the fossil’s facial muscles were long gone, the bones of its throat indicated that the whale was capable of retracting its tongue, creating negative pressure that would have drawn in water and small creatures from the surrounding environment. The toothless whale would have had no means to chew its prey, simply swallowing them whole. But the switch to suction was likely a move that would have averted a significant cost: Teeth are energetically expensive to produce. And it was on this blank canvas, the authors propose, that baleen eventually took shape a few million years later.
 <figure class="richtext-image full-width"><img alt="Carlos Peredo - Maiabalaena_Final_Cover.tif" class="richtext-image full-width" height="1035" width="800"
 src="https://www.pbs.org/wgbh/nova/media/images/Carlos_Peredo_-_Maiabalaena_Final_Cover.width-800.jpg"/><figcaption>An artist&#x27;s imagining of a Maiabalaena nesbittae calf nursing from its mother. Artist Credit: Alex Boersma<div class="credit"></div></figcaption>
 </figure>
 As modern humans can attest, teeth come in handy. They allow creatures to both latch onto their prey and grind them down to swallowable portions, significantly increasing the size and breadth of an animal’s viable menu items. Suction feeding, on the other hand, doesn’t pulverize prey before it goes down the gullet, and favors smaller, softer-bodied meals. But suction feeding’s lack of oral ornamentation means it’s a relatively cheap strategy. Even toothy mammals like humans are no stranger to its efficiency: Suckling is the only way babies can eat before their teeth come in.Some scientists, however, are a little more hesitant to embrace the idea that M. nesbittae was a suction feeder. “I’m not convinced it didn’t have teeth,” says <a href="https://www.rachelracicot.org/">Rachel Racicot</a>, a paleobiologist who studies marine mammals at the Keck Science Center. Critical pieces of the fossil’s jaw are fairly worn or incomplete, she points out, muddying the conclusions that can be drawn about the presence or absence of dentition. Until more fossils from this time period reveal themselves, suction feeding will be difficult to confirm, Racicot adds. Alternatively, evidence of <a href="https://www.frontiersin.org/articles/10.3389/fmars.2017.00067/full">a more stepwise transition</a> from teeth to baleen may exist—which could mean there were once intermediate whale species that harbored both in their mouths. The two theories on this interim aren’t necessarily mutually exclusive.
 

 
 <div class="newsletter"></div>
 

 
 Regardless of how these whales fed in the years after they lost their teeth, the ultimate advent of baleen was revolutionary. According to Peredo, “it’s like growing fingernails out of the roof of your mouth.” There’s nothing else quite like baleen in the animal kingdom, and we probably wouldn’t have the sea’s biggest creatures without it. By dispensing of the sea water “filler” in each gulp, filter-feeding ensures that baleen whales can engulf gargantuan amounts of food in one go: A blue whale can take up <a href="http://blogs.discovermagazine.com/notrocketscience/2010/12/09/blue-whales-can-eat-half-a-million-calories-in-a-single-mouthful/">half a million calories</a> in a single mouthful.“If you drink coffee, you’re taking advantage of filter feeding,” Pyenson explains. “It’s just separating fluid out from a substance.” Though in the case of whales, he adds, they’re hungry for the grinds, not the liquid coffee, so to speak. A good filter means that especially tiny forms of prey—which don’t require chewing—can comprise a viable meal.
 <figure class="richtext-image full-width"><img alt="asf_bestof035.jpg" class="richtext-image full-width" height="531" width="800"
 src="https://www.pbs.org/wgbh/nova/media/images/asf_bestof035.width-800.jpg"/><figcaption>Modern baleen whales, like humpbacks, employ a filter-feeding strategy that enormously ups the nutritional value of each mouthful compared to suction feeding. This has enabled many whale species to grow to gargantuan sizes. Photo Credit: Ari S. Friedlaender/University of California, Santa Cruz under NOAA permit<div class="credit"></div></figcaption>
 </figure>
 Chomping and suction feeding still have their perks, though. All three feeding strategies exist in modern whales, helping many species comfortably share the seas. And adopting one tactic doesn’t preclude a whale from utilizing another. While some toothed whales do chew on their prey, many like orcas or beluga whales simply use their teeth as graspers before swallowing food whole. Others like the beaked whale or narwhal have only rudimentary tooth structures, and must rely on suction to more or less inhale their prey. Though gray whales are equipped with baleen, they happily suction feed as well.This sort of flexibility may become increasingly important as the planet continues to change under the influence of human intervention. “Maiabalaena is like its own little case study,” Peredo says. “The oceans changed dramatically, and this is what happened—whales changed what they ate, the physical structures of their mouths.”But whales aren’t infinitely adaptable; even baleen took millions of years to come into being. It remains unclear how sea life will fare in an age of <a href="https://www.pbs.org/wgbh/nova/article/climate-change-plants-high-temperature/">rising temperatures</a> and increasingly pervasive <a href="https://www.pbs.org/wgbh/nova/article/ships-noises-humpback-whale-song/">acoustic pollution</a>. For now, researchers are keeping an eye on the past. The search for baleen’s origins may yet reveal the climactic changes that triggered its rise—and how whales weathered the storm with their mysterious shuttered smiles.

Humpbacks and many other whales use baleen—the overlapping structures in their mouths where teeth might otherwise be—to filter tiny prey like krill and plankton out of water. But because baleen has the texture of hair or fingernails, it doesn't preserve well in the fossil record. New research now indicates that some whales may have fed by suction in the evolutionary interim between teeth and baleen. Photo Credit: Ari S. Friedlaender/University of California, Santa Cruz under NOAA permit

After discarding their teeth, the ancestors of today’s filter-feeding whales may have fed by suction before evolving baleen.

This 33-Million-Year-Old Whale Sucked Up Food Like A Giant Vacuum

For humans, chess may take a lifetime to master. But Google DeepMind’s new <a href="https://www.pbs.org/wgbh/nova/video/can-machines-think-like-us">artificial intelligence</a> program, AlphaZero, can teach itself to conquer the board in a matter of hours.Building on its past success with the <a href="https://www.pbs.org/wgbh/nova/article/googles-alphago-program-trounces-world-go-champion-in-dramatic-first-match/">AlphaGo</a> suite—a series of computer programs <a href="https://www.pbs.org/wgbh/nova/article/professional-human-go-player-defeated-by-a-computer-for-the-first-time/">designed to play</a> the Chinese board game Go—Google boasts that its new AlphaZero achieves a level of “<a href="https://www.nature.com/articles/nature16961">superhuman</a> performance” at not just one board game, but three: Go, chess, and shogi (essentially, Japanese chess). The team of computer scientists and engineers, led by Google’s <a href="http://www0.cs.ucl.ac.uk/staff/d.silver/web/Home.html">David Silver</a>, reports <a href="http://science.sciencemag.org/cgi/doi/10.1126/science.aar6404">its findings</a> today in the journal Science.“Before this, with machine learning, you could get a machine to do exactly what you want—but only that thing,” says <a href="https://www.ic.gatech.edu/people/ayanna-howard">Ayanna Howard</a>, an expert in interactive computing and artificial intelligence at the Georgia Institute of Technology who did not participate in the research. “But AlphaZero shows that you can have an algorithm that isn’t so [specific], and it can learn within certain parameters.”
 

 
 <div class="ad"></div>
 

 
 AlphaZero’s clever programming certainly ups the ante on gameplay for human and machine alike, but Google has long had its sights set on something bigger: engineering intelligence.The researchers are careful not to claim that AlphaZero is on the verge of world domination (others have been <a href="https://www.telegraph.co.uk/science/2017/12/06/entire-human-chess-knowledge-learned-surpassed-deepminds-alphazero/">a little quicker to jump the gun</a>). Still, Silver and the rest of the DeepMind squad are already hopeful that they’ll someday see a similar system applied to drug design or materials science.So what makes AlphaZero so impressive?
 <figure class="richtext-image full-width"><img alt="Chad Miller flickr 7270219_6d3f41bc71_o.jpg" class="richtext-image full-width" height="600" width="800"
 src="https://www.pbs.org/wgbh/nova/media/images/Chad_Miller_flickr_7270219_6d3f41bc71_o.width-800.jpg"/><figcaption>AlphaZero is based on AlphaGo Zero, part of the AlphaGo suite designed to play the Chinese board game Go, pictured above. Early iterations of the original program were fed data from human-versus-human games; later versions engaged in self-teaching, wherein the software played games against itself to learn its own strategy. Photo Credit: Chad Miller, flickr<div class="credit"></div></figcaption>
 </figure>
 Gameplay has long been revered as a <a href="https://www.pbs.org/wgbh/nova/article/watch-live-as-a-computer-challenges-the-reigning-world-go-champion/">gold standard</a> in artificial intelligence research. Structured, interactive games are simplifications of real-world scenarios: Difficult decisions must be made; wins and losses drive up the stakes; and prediction, critical thinking, and strategy are key.Encoding this kind of skill is tricky. Older game-playing AIs—including the first prototypes of the original AlphaGo—have traditionally been pumped full of codes and data to mimic the experience typically earned through years of natural, human gameplay (essentially, a passive, programmer-derived knowledge dump). With <a href="https://deepmind.com/blog/alphago-zero-learning-scratch">AlphaGo Zero</a> (the most recent version of AlphaGo), and now AlphaZero, the researchers gave the program just one input: the rules of the game in question. Then, the system hunkered down and actively learned the tricks of the trade itself.This strategy, called self-play reinforcement learning, is pretty much exactly what it sounds like: To train for the big leagues, AlphaZero played itself in iteration after iteration, honing its skills by trial and error. And the brute-force approach paid off. Unlike AlphaGo Zero, AlphaZero doesn’t just play Go: It can beat the best AIs in the business at chess and shogi, too. The learning process is also impressively efficient, requiring only two, four, or 30 hours of self-tutelage to outperform programs specifically tailored to master shogi, chess, and Go, respectively. Notably, the study authors didn’t report any instances of AlphaZero going head-to-head with an actual human, Howard says. (The researchers may have assumed that, given that these programs consistently clobber their human counterparts, such a matchup would have been pointless.)
 

 
 <div class="related"></div>
 

 
 AlphaZero was also able to trounce Stockfish (the now unseated AI chess master) and Elmo (the former AI shogi expert) despite evaluating fewer possible next moves on each turn during game play. But because the algorithms in question are inherently different, and may consume different amounts of power, it’s difficult to directly compare AlphaZero to other, older programs, points out <a href="http://www.cs.bath.ac.uk/~jjb/">Joanna Bryson</a>, who studies artificial intelligence at the University of Bath in the United Kingdom and did not contribute to AlphaZero.Google keeps mum about a lot of the fine print on its software, and AlphaZero is no exception. While we don’t know everything about the program’s power consumption, what’s clear is this: AlphaZero has to be packing some serious computational ammo. In those scant hours of training, the program kept itself very busy, engaging in tens or hundreds of thousands of practice rounds to get its board game strategy up to snuff—far more than a human player would need (or, in most cases, could even accomplish) in pursuit of proficiency.This intensive regimen also used 5,000 of Google’s <a href="https://cloud.google.com/tpu/docs/tpus">proprietary machine-learning processor units</a>, or TPUs, which by some estimates <a href="https://www.nextplatform.com/2018/05/10/tearing-apart-googles-tpu-3-0-ai-coprocessor/">consume around 200 watts per chip</a>. No matter how you slice it, AlphaZero requires way more energy than a <a href="https://www.pbs.org/video/nova-wonders-can-we-build-a-brain-j53aqg/">human brain</a>, which runs on <a href="https://www.popsci.com/technology/article/2009-11/neuron-computer-chips-could-overcome-power-limitations-digital">about 20 watts</a>.The absolute energy consumption of AlphaZero must be taken into consideration, adds <a href="https://www.stat.berkeley.edu/~binyu/Site/Welcome.html">Bin Yu</a>, who works at the interface of statistics, machine learning, and artificial intelligence at the University of California, Berkeley. AlphaZero is powerful, but might not be good bang for the buck—especially when adding in the person-hours that went into its creation and execution.
 <figure class="richtext-image full-width"><img alt="Vzsuzsi pixabay chess-2551751_1920.jpg" class="richtext-image full-width" height="555" width="800"
 src="https://www.pbs.org/wgbh/nova/media/images/Vzsuzsi_pixabay_chess-2551751_1920.width-800.jpg"/><figcaption>In chess, AlphaZero defeated the previous chess-playing world champion AI Stockfish, winning 155 games and losing 6 games out of 1000. The rest ended in draws. Photo Credit: Vzsuzsi, Pixabay<div class="credit"></div></figcaption>
 </figure>
 Energetically expensive or not, AlphaZero makes a splash: Most AIs are hyper-specialized on a single task, making this new program—with its triple threat of game play—remarkably flexible. “It’s impressive that AlphaZero was able to use the same architecture for three different games,” Yu says.So, yes. Google’s new AI does set a new mark in several ways. It’s fast. It’s powerful. But does that make it smart?
 

 
 <div class="newsletter"></div>
 

 
 This is where definitions start to get murky. “AlphaZero was able to learn, starting from scratch without any human knowledge, to play each of these games to <a href="https://www.nature.com/articles/nature16961">superhuman</a> level,” DeepMind’s Silver said in a statement to the press.Even if board game expertise requires mental acuity, all proxies for the real world have their limits. In its current iteration, AlphaZero maxes out by winning human-designed games—which may not warrant the potentially alarming label of “superhuman.” Plus, if surprised with a new set of rules mid-game, AlphaZero might get flummoxed. The actual human brain, on the other hand, can store far more than three board games in its repertoire.What’s more, comparing AlphaZero’s baseline to a tabula rasa (blank slate)—as the researchers do—is a stretch, Bryson says. Programmers are still feeding it one crucial morsel of human knowledge: the rules of the game it’s about to play. “It does have far less to go on than anything has before,” Bryson adds, “but the most fundamental thing is, it’s still given rules. Those are explicit.”And those pesky rules could constitute a significant crutch. “Even though these programs learn how to perform, they need the rules of the road,” Howard says. “The world is full of tasks that don’t have these rules.”<div>
 <iframe width="480" height="270" src="https://www.youtube.com/embed/mI3YP4UlPvk?feature=oembed" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>
</div>
When push comes to shove, AlphaZero is an upgrade of an already powerful program—AlphaGo Zero, explains <a href="https://www.faculty.ece.vt.edu/jmpaul/">JoAnn Paul</a>, who studies artificial intelligence and <a href="https://sites.google.com/site/computationaldreaming/home">computational dreaming</a> at the Virginia Polytechnic Institute and State University and was not involved in the new research. AlphaZero uses many of the same building blocks and algorithms as AlphaGo Zero, and still constitutes just a subset of true smarts. “I thought this new development was more evolutionary than revolutionary,” she adds. “None of these algorithms can create. Intelligence is also about storytelling. It’s imagining things that are not yet there. We’re not thinking in those terms in computers.”Part of the problem is, there’s still no consensus on a true definition of “intelligence,” Yu says—and not just in the domain of technology. “It’s still not clear how we are training critically thinking beings, or how we use the unconscious brain,” she adds.To this point, many researchers believe there are likely <a href="https://www.tecweb.org/styles/gardner.html">multiple types of intelligence</a>. And tapping into one far from guarantees the ingredients for another. For instance, some of the smartest people out there are terrible at chess.With these limitations, Yu’s vision of the future of artificial intelligence partners humans and machines in a kind of coevolution. Machines will certainly continue to excel at certain tasks, she explains, but human input and oversight may always be necessary to compensate for the unautomated.Of course, there’s no telling how things will shake out in the AI arena. In the meantime, we have plenty to ponder. “These computers are powerful, and can do certain things better than a human can,” Paul says. “But that still falls short of the mystery of intelligence.”

Google's new artificial intelligence program, AlphaZero, taught itself to play chess, shogi, and Go in a matter of hours, and outperforms the top-ranking AIs in the gameplay arena. Photo Credit: iStock

Yes, Google’s self-teaching artificial intelligence software, AlphaZero, will probably trounce you at chess. But there’s far more to human smarts than a speedy checkmate.

Google says it built a ‘superhuman’ game-playing AI. Is it truly intelligent?

For single-celled organisms, <a href="https://www.pbs.org/video/nova-wonders-whats-living-in-you-fnbfuy/">bacteria</a> are surprisingly chatty.Though they’re often thought of as solitary, bacteria spend a decent bit of time engaging in conversation. But they don’t use words; instead, these microscopic creatures shoot chemical cues into their environment. When many bacteria are in close proximity, these signals can reach a high enough density to prompt complex, cooperative group behavior.Banding together, bacteria can raise collective defenses, enhance access to resources, and even cause serious infections. This collaborative decision-making process is called <a href="https://www.ncbi.nlm.nih.gov/pubmed/11544353">quorum sensing</a>, and it comes in different flavors: Some bacteria collude only with their own kind, while others will speak with other species. But outside the bacterial kingdom, these bugs keep mum. Their chemical code is one that no living creature—<a href="https://www.pbs.org/wgbh/nova/article/bassler-bacteria/">aside from a human researcher</a>—has ever been able to crack.The key word here, however, is “living.” Today, in the journal Cell, a team of researchers unmasks <a href="https://doi.org/10.1016/j.cell.2018.10.059">an unexpected decrypter</a> of the dialect of quorum sensing: A bacteria-hunting virus called a <a href="https://www.pbs.org/wgbh/nova/article/phage-alzheimers-cure/">bacteriophage</a>, or phage for short.
 

 
 <div class="ad"></div>
 

 
 By learning the lexicon of bacteria in the Vibrio genus—including those that cause <a href="https://www.pbs.org/wgbh/nova/bioterror/agen_cholera.html">cholera</a>—a polyglot phage called VP882 effectively wiretaps the conversations of its microbial prey, which might allow the virus to attack when its target population is at its densest. Genetically manipulating the building blocks of this system could someday equip researchers and clinicians with tools to treat bacterial disease—all inspired by the covert ops of these viral eavesdroppers.“This is really beautiful phage biology,” says <a href="https://www.metx.ucsc.edu/research/yildiz.html">Fitnat Yildiz</a>, a microbiologist who studies Vibrio bacteria at the University of California, Santa Cruz but did not participate in the new research. “This is an elegant example of how phage can take a [bacterial] regulatory network and use it to their own advantage.”But phages aren’t the only ones listening in. Study author <a href="https://scholar.princeton.edu/basslerlab/home">Bonnie Bassler</a>, a microbiologist at Princeton University, may be humanity’s most fluent ambassador to the world of <a href="https://www.pbs.org/wgbh/nova/article/bassler-bacteria/">quorum sensing</a>. Years of research from her team have deciphered some of the most important facets of this bacterial vernacular. Just last year, a team from Bassler&#x27;s lab, including graduate student <a href="https://scholar.princeton.edu/basslerlab/people/justin-silpe">Justin Silpe</a>, discovered that bacteria in the Vibrio group are <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5391282/">capable of detecting</a> a molecular signal called DPO. In the world of quorum sensing, when it’s loud, there’s a crowd—and blasting DPO allows microbes to announce themselves to others. A single blip means virtually nothing, but a whole slew of DPO messages indicates the presence of a mob. A swarm of microbes can then take this cue to act in concert.
 <figure class="richtext-image full-width"><img alt="Phage.jpg" class="richtext-image full-width" height="936" width="800"
 src="https://www.pbs.org/wgbh/nova/media/images/Phage.width-800.jpg"/><figcaption>In this image, multiple bacteriophages are attached to the surface of a bacterial cell. Bacteriophage translates to &quot;bacteria eater.&quot; Photo Credit: Graham Beards, Wikimedia Commons<div class="credit"></div></figcaption>
 </figure>
 Through chemicals like DPO, bacteria are constantly chattering in the world around us—but we humans never hear a thing. The system functions like a radio station, in that effective communication requires a signal and a receiver. A tower can ping out all the waves it wants, but only those with the right technology can tune in to the message.That’s no issue for bacteria: All Vibrio species make DPO’s “receiver”—a protein called VqmA. But Silpe was eager to find who else might be privy to the discourse. While searching for versions of VqmA across the kingdoms of life, Silpe noticed an anomaly on his list that wasn’t alive at all: A Vibrio-infecting virus.Both Silpe and Bassler were initially skeptical. “There’s never been a quorum sensing gene on a phage before,” Bassler explains. “Quorum sensing is supposed to be about bacteria communicating.”
 

 
 <div class="related"></div>
 

 
 And when it comes to code breakers, viruses are especially far outside the line of usual suspects. Technically, they’re <a href="https://www.pbs.org/video/where-did-viruses-come-from-rc0f7r/">not even alive</a>. Viral particles are <a href="https://www.pbs.org/wgbh/nova/labs/lab/rna/3/1/">entirely dependent on living cells</a>, like those of humans or bacteria, to replicate and persist. But the viral modus operandi is undoubtedly a successful one. Bacteriophages alone are the <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3109452/">most ubiquitous entities</a> in nature, found wherever their host of choice—bacteria—thrive.When a phage encounters a bacterial cell, it will inject its own genome into its host. Then, the virus co-opts the bacterium’s molecular machinery to manufacture more of itself. The process destroys the host, releasing a fresh batch of phage to begin the infectious cycle anew.But the timing of this exit can vary. After infecting a cell, some phages will nestle their genes in next to their host’s, genetically immortalizing the virus. When the bacterium copies its own genome during reproduction, the phage’s gets taken along for the ride. In these cases, only specific cues from the environment, like damage to the host, can awaken the phage from dormancy.<div>
 <iframe width="480" height="270" src="https://www.youtube.com/embed/HLAIZ5CbjRE?feature=oembed" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>
</div>
For the phage, coming out of hiding is irreversible, as the process destroys the original host. But if there’s a new crop of susceptible bacteria nearby, the benefits far outweigh the risk. So for VP882, monitoring the availability of prey—say, by encoding its own quorum sensing cipher—might be an incredibly effective strategy.At first, the possibility seemed ludicrous to the researchers. But Silpe followed his hunch, and was amazed to find that the phage gene didn’t just look like VqmA—it listened like it, too. When Vibrios in its vicinity sent out DPO dispatches, VP882 took notice. And if the messages amassed, the phage jolted awake and broke loose from the cell.Because more DPO means more bacteria—and, thus, more potential hosts—adding VqmA to its repertoire might be a way for a phage to time a mass exodus, Bassler explains. Though this idea awaits confirmation, Bassler thinks the virus might be using Vibrio’s own secret language against itself. It’s an “insidious and fantastic” tactic, she says.
 

 
 <div class="newsletter"></div>
 

 
 “This is remarkable,” says <a href="https://bsc.ua.edu/profiles/asma-hatoum-aslan/">Asma Hatoum-Aslan</a>, a microbiologist studying bacteria-phage interactions at the University of Alabama who did not contribute to the study. “This shows that this kind of phage, even in the host cell, is clearly not dormant. It’s listening. It’s still interacting with the environment on some level.”The surprising crosstalk between such different entities also illuminates the widespread nature of communication—in even its simplest forms. “Viruses don’t carry much genetics around with them, but they actually have a lot of autonomy,” says <a href="https://turnerlab.yale.edu/">Paul Turner</a>, a phage biologist at Yale University who did not contribute to the study. Phages are dependent on cellular life to propagate, but by turning an ear to their hosts, he says, “they’re still encoded to decide their own fate.”For Silpe and Bassler, this is only the beginning. This may be the first phage that’s officially recognized for learning the quorum sensing lingo, but the researchers think VP882 is unlikely to be alone.
 <figure class="richtext-image full-width"><img alt="187435.jpg" class="richtext-image full-width" height="449" width="800"
 src="https://www.pbs.org/wgbh/nova/media/images/187435.width-800.jpg"/><figcaption>Princeton University microbiologists Justin Silpe and Bonnie Bassler, who discovered a bacteriophage who can eavesdrop on bacterial conversations. Photo Credit: Denise Applewhite, Office of Communications, Princeton University<div class="credit"></div></figcaption>
 </figure>
 “Phages were discovered over a century ago,” says <a href="https://www.weizmann.ac.il/molgen/Sorek/">Rotem Sorek</a>, a microbiologist who studies bacteriophages at the Weizmann Institute of Science in Israel but did not contribute to the study. “The foundations of microbiology came from phages. And still, after all these years, we are still finding very surprising and new features of phage biology.”The researchers have already begun to brainstorm therapeutic applications of this unusual system. One potential avenue might be a riff on <a href="https://www.pbs.org/wgbh/nova/article/phages-for-food-safety/">phage therapy</a>, wherein bacteriophages can be deployed to <a href="https://www.pbs.org/wgbh/nova/article/hacking-a-bacteriophages-genome-may-lead-to-a-cystic-fibrosis-treatment/">selectively kill</a> disease-causing bacteria. VP882 naturally targets bacteria of the Vibrio genus, including cholera-causing Vibrio cholerae. This phage could be engineered to function as something of a Trojan horse, Silpe says, delivering a molecular self-destruct button to dense populations of Vibrio without damaging, say, the <a href="https://www.pbs.org/wgbh/nova/video/nova-wonders-whats-living-in-you/">important gut microbes</a> that live within a human host. Turner compares the technique to “taking a scalpel and cutting something unwanted out of a complex community.”Such a system would involve harnessing the power of a natural system derived from phage-microbe interactions—not unlike <a href="https://www.pbs.org/wgbh/nova/video/how-does-crispr-work/">CRISPR</a>, Turner says. “We’re figuring out what’s evolved in the microbial world and appreciating its elegance,” he explains. “But now we can also think about using these tools to solve human problems.”Editor&#x27;s Note, December 25, 2018: This article has been updated to clarify the role of Justin Silpe in the discovery of the quorum sensing molecule, DPO.

Viruses that infect bacteria, or bacteriophages, are a little bit like infectious syringes. When they burst out of bacterial cells, they need to find new hosts—and some phages may have evolved a way to listen for susceptible prey. Photo Credit: Design Cells, iStock

In a first, scientists have found that viruses that prey on microbes can understand the lingo of bacterial conversations—and may use the intel to time their attacks.

Chatty bacteria may be most vulnerable to viruses

Humans have a knack for <a href="https://www.pbs.org/wgbh/nova/article/the-limits-of-facial-recognition/">remembering</a> <a href="https://www.pbs.org/wgbh/nova/video/where-your-face-comes-from">faces</a>. And when someone forgets yours, it tends to sting.The same is true, it turns out, for the Northern paper wasp (Polistes fuscatus). Though often maligned as painful pests, wasps have more in common with humans than some of us might like to believe—including a particular <a href="http://science.sciencemag.org/content/334/6060/1272">aptitude</a> for recognizing individuals <a href="https://www.pbs.org/wgbh/nova/article/facial-recognition-brain/">by their facial features</a>.While scientists have been aware of this phenomenon for years, what’s less clear is how these wasps accomplish this feat: The ability to distinguish faces could be inherited, or something learned through experience. Now, in <a href="https://www.sciencedirect.com/science/article/pii/S0003347218303178">a new study</a> published in the journal Animal Behaviour, researchers report that part of the process is built into the wasp brain—and species lacking this innate ability may never entirely make up for the difference.
 

 
 <div class="ad"></div>
 

 
 “This work is absolutely groundbreaking,” says <a href="https://www.rmit.edu.au/contact/staff-contacts/academic-staff/d/dyer-associate-professor-adrian">Adrian Dyer</a>, who studies the science of vision at RMIT University in Australia but was not involved in the research. “It shows a bridge between expertise… and a more hardwired system.”Even among paper wasps, P. fuscatus is exceptional: This species seems uniquely equipped to home in on <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1691056/">facial features</a>—and it might have something to do with their social structure. Many P. fuscatus wasp colonies are founded by multiple females who tussle for dominance. Once a pecking order is established, each queen can take on a unique social role, which goes on to dictate how workers integrate into the nest hierarchy. As such, the ability to recognize individuals is essential for these wasps to maintain stable relationships within the colony, affecting everything from displays of <a href="https://onlinelibrary.wiley.com/doi/full/10.1111/j.1558-5646.2009.00833.x">aggression</a> to division of labor and food allocation. Accordingly, the facial features of P. fuscatus wasps are highly variable.In a closely related species, Polistes metricus, or the metricus paper wasp, nests more commonly support single-queen monarchies—a simplified hierarchy that eliminates the need for detailed distinctions between wasps. Unsurprisingly, metricus faces are mostly indistinguishable, <a href="https://onlinelibrary.wiley.com/doi/full/10.1111/j.1420-9101.2009.01923.x">even to their own kind</a>.
 <figure class="richtext-image full-width"><img alt="sankax flickr.jpg" class="richtext-image full-width" height="600" width="800"
 src="https://www.pbs.org/wgbh/nova/media/images/sankax_flickr.width-800.jpg"/><figcaption>The Northern paper wasp (Polistes fuscatus) is capable of distinguishing the facial features of individuals. This comes in handy in their nests, where there a complex hierarchy with multiple queens necessitates that members of different castes keep careful track of each other. Photo Credit: sankax, flickr<div class="credit"></div></figcaption>
 </figure>
 Study author <a href="https://lsa.umich.edu/eeb/people/faculty/tibbetts.html">Elizabeth Tibbetts</a>, an evolutionary biologist at the University of Michigan, realized these P. metricus’ shortcomings might just be the key to unraveling the cognitive basis of facial recognition. Face-savvy P. fuscatus individuals are reared in the presence of facial diversity—which means they’re acquainted with an assortment of colors, shapes, and sizes right from the get-go, whereas metricus wasps don’t enjoy the same privilege. It’s a classic nature versus nurture question: Are P. metricus wasps are inherently ill-suited to the task, or are they simply deprived of the necessary education?To test the influence of experience on face learning, Tibbetts and her colleagues set up a series of hybrid hives. In some, P. fuscatus and P. metricus wasps were reared with their own kind; in others, they were mixed with each other. The researchers also added a third, more distantly-related species, Polistes dominula, or the European paper wasp, to the mix. P. dominula individuals take note of features that indicate physical prowess—the faces of stronger wasps have a more splotchy pattern of black spots—but these motifs are too broad to identify individuals. Tibbetts compares the system to belt colors in a karate studio. “To tell how tough someone is, you can look at their belt color,” she explains. “But you don’t need to memorize individual faces.”
 

 
 <div class="related"></div>
 

 
 After five days of co-mingling, the researchers enrolled the wasps in a face-learning boot camp. The wasps were trained to make their way through a convoluted portrait gallery of different P. fuscatus faces—a perfectly pleasant setup, save for the fact that the floors of the museum-like maze were electrified. Some of the images, however, were placed in safe, no-shock zones. The researchers tracked whether the wasps were able to use these images to navigate themselves to safety, even when the photos switched locations—a sign that they recognized and remembered these faces’ distinctive features.As expected, P. fuscatus that had pow-wowed with their own species came out ahead in this face-off, quickly learning to discriminate between safe and unsafe images. The same wasn’t true, however, for P. metricus and P. dominula housed with their own kind. This made sense: Only the fuscatus wasps were gleaning instructive information from their peers, and putting it to good use. But Tibbetts and her colleagues found that P. fuscatus also seemed to hold its own when it bunked with P. metricus wasps. Even without direct exposure, P. fuscatus showed a flair for faces—indicating that the P. fuscatus brain might have some unique, face-discriminating hardwiring.
 <figure class="richtext-image full-width"><img alt="P fuscatus individual recognition.jpg" class="richtext-image full-width" height="800" width="800"
 src="https://www.pbs.org/wgbh/nova/media/images/P_fuscatus_individual_recognition.width-800.jpg"/><figcaption>The many faces of Polistes fuscatus. Faces vary in many ways, including coloration, patterning, and the size of individual features. Photo Credit: Elizabeth Tibbetts, University of Michigan<div class="credit"></div></figcaption>
 </figure>
 But this wasn’t the entire story. Tibbetts was next shocked to see that P. metricus wasps—which normally can’t tell faces apart—were also able to successfully maneuver their way through the face maze if they had first fraternized with P. fuscatus wasps. The metricus wasps never did as well as true P. fuscatus wasps, but they outperformed their unexposed peers. Though the metricus wasps had spent only five days with P. fuscatus, this brief stint had apparently been enough for them to learn some of the tricks of the trade. “I thought experience would matter,” Tibbetts explains, “but I didn’t think they would really get better.”When the researchers repeated the experiment with photos of P. fuscatus lacking antennae—a modification that jumbles the ability of these wasps to recognize a face as a face—neither wasp species did as well. This reinforced the idea that face specialization wasn’t just about recognizing general patterns, like karate belt colors; rather, there was something special about the face itself.“This contributes to a growing literature on the idea that things aren’t black and white: It’s not just nature or nurture,” says <a href="https://www.ent.iastate.edu/people/amy-toth">Amy Toth</a>, an entomologist at Iowa State University who did not contribute to the new finding. “It’s the interaction that leads to the development of this complex behavior.”
 

 
 <div class="newsletter"></div>
 

 
 Sadly, P. dominula was not so lucky. Regardless of the company it kept, these wasps didn’t show any particular proficiency for distinguishing faces. For them, experience alone wasn’t enough, indicating that inheritance still plays a big role in face specialization.It’s still not entirely clear what underlies these differences between species, but Tibbetts theorizes that there may be an ancestral influence. P. fuscatus and P. metricus are much more closely related than either species is to P. dominula—and P. metricus’ latent abilities may be a holdover from its past. But until far more wasp species are studied, no conclusions can be definitively drawn, says <a href="http://www.sumnerlab.co.uk/the-group/seirian-sumner/">Seirian Sumner</a>, who studies wasps and other social insects at University College London in the United Kingdom but did not contribute to the new research.
 <figure class="richtext-image full-width"><img alt="Larah McElroy flickr.jpg" class="richtext-image full-width" height="533" width="800"
 src="https://www.pbs.org/wgbh/nova/media/images/Larah_McElroy_flickr.width-800.jpg"/><figcaption>The metricus paper wasp (Polistes metricus) isn&#x27;t capable of distinguishing individual faces. Because each metricus nest tend to be ruled by a single queen, there&#x27;s no need for this level of social intricacy. Photo Credit: Larah McElroy, flickr<div class="credit"></div></figcaption>
 </figure>
 Even more far-flung is the evolutionary distance between humans and wasps. But the fact that the two species have managed to converge on a similar skill fascinates Tibbetts. “Often, when I’m studying wasps, I can’t believe how smart they are,” she says. “They can do so many complicated things. But at the same time, maybe if wasps can do it, maybe you don’t have to be that smart to do these things.”That’s not to knock on wasps, which certainly suffer from an image problem in the eyes of humans, <a href="https://www.smithsonianmag.com/science-nature/why-you-should-let-wasp-be-your-hero-today-180969521/">perhaps unjustly so</a>. Dyer, the vision scientist, thinks studies like these are deeply informative—not just for our understanding of evolution, but also for unexpected applications like <a href="https://www.pbs.org/wgbh/nova/video/can-machines-think-like-us">machine learning</a>. Insect brains have far less computational power than the <a href="https://www.pbs.org/wgbh/nova/video/nova-wonders-can-we-build-a-brain/">human mind</a>, but can still perform exceptionally complex tasks, Dyer explains. “There must be some efficient ways in which miniature brains can do this,” he says. “We can derive real tricks from simple biological systems.”
 <figure class="richtext-image full-width"><img alt="Dann Thombs flickr.jpg" class="richtext-image full-width" height="533" width="800"
 src="https://www.pbs.org/wgbh/nova/media/images/Dann_Thombs_flickr.width-800.jpg"/><figcaption>The European paper wasp (Polistes dominula) recognizes only broad patterns in faces that help wasps size up their competition. More splotchy black spots indicate that an individual is stronger than one with a more uniformly patterned face. Photo Credit: Dann Thombs, flickr<div class="credit"></div></figcaption>
 </figure>
 Tibbetts is also intrigued by this line of questioning. Next, she hopes to uncover what’s actually happening inside the wasp brain during these complicated learning tasks—and if there are further parallels between humans and these ingenious insects.“This work really demonstrates why we should love wasps,” Toth says. “They’re fascinating creatures with complex social behavior. They may be the closest thing you can find to a human in an insect.”

Different species of paper wasps can recognize faces to different extents. Polistes dominula, pictured here, can't tell individuals apart—but it can distinguish broad patterns in faces that denote strength. Photo Credit: Dann Thombs, flickr

Some types of paper wasps can tell individuals apart by facial features alone. Is the ability built in, or can it be learned?

This Wasp Can Recognize Faces. Others Aren’t So Lucky.

As the planet’s thermostat <a href="https://www.smithsonianmag.com/smart-news/world-was-just-issued-12-year-ultimatum-climate-change-180970489/?no-cache">ticks steadily upwards</a>, life forms around the world may be in grave danger—and plants like rice are no exception. Already, global rice yields have <a href="http://www.pnas.org/content/114/35/9326">taken a hit</a> from rising temperatures. If the upward creep in heat continues, declines in this starchy staple could imperil the 3.5 billion people around the world who rely on rice for over 20% of their daily caloric intake.Today, in the journal PNAS, researchers <a href="http://www.pnas.org/lookup/doi/10.1073/pnas.1807988115">report</a> that rice plants are sensitive to the damaging effects of high heat even in the short term. Brief exposures to sweltering temperatures unraveled plant cells’ messenger RNA—a critical molecular envoy that ferries the instructions contained within DNA to the cell’s protein-producing factories—cuing the messages up for destruction. Though the rice eventually recovered, the new finding underscores just how sensitive plants may be to the hazards of a fast-warming world.
 

 
 <div class="ad"></div>
 

 
 With each degree-Celsius increase in average global temperature, rice crop yields fall by <a href="http://www.pnas.org/content/114/35/9326">about 3%</a>. And as the planet warms and <a href="http://www.un.org/en/development/desa/news/population/2015-report.html">the human population expands</a>, the gap between supply and demand for this staple crop continues to widen. But while these trends are clear at a macroscopic level, the inner workings of plant cells coping with toasty temperatures remain far less clear.To better understand heat tolerance in crops, a team of scientists led by <a href="http://www.personal.psu.edu/sma3/people.html">Zhao Su</a>, a plant biologist at Pennsylvania State University, heated cuttings of young rice shoots (Oryza sativa) to a temperature of 42º Celsius (nearly 108º Fahrenheit) for 10 minutes. Though temporary, the mini heat wave constituted a rapid 20º-Celsius shift upwards from room temperature—an escalation that’s “more than a human could ever withstand,” explains study author <a href="http://chem.psu.edu/directory/pcb5">Philip Bevilacqua</a>, an RNA chemist also at Pennsylvania State University.When the researchers analyzed the shoots’ cellular contents, they found that the temporary trauma had compromised the rice’s ability to relay genetic information through their messenger RNA, or mRNA.
 <figure class="richtext-image full-width"><img alt="PIXNIO-15202-2700x1767.jpeg" class="richtext-image full-width" height="523" width="800"
 src="https://www.pbs.org/wgbh/nova/media/images/PIXNIO-15202-2700x1767.width-800.jpg"/><figcaption>With every degree-Celsius increase in temperature, rice crop yields drop by over 3%. Photo Credit: Pixnio<div class="credit"></div></figcaption>
 </figure>
 Plants, like humans, store their genetic information in the form of DNA. Contained within this molecular manuscript are all the necessary instructions for growth and survival—including all the cell’s contingency plans for emergencies or unanticipated changes in the surrounding environments. The DNA manual is extremely comprehensive, so cells access excerpts from it on an as-needed basis, making short-lived photocopies of the original text in the form of mRNA. Without mRNA, cells can’t receive the intel necessary to manufacture the building blocks necessary for life.But for such a critical piece of the cellular puzzle, mRNA is frustratingly fickle. Unlike DNA, which is stored in two strands that lock into each other like the halves of a zipper, the synthesis of mRNA creates an unwieldy single-stranded molecule. These open-faced strands can loop around and bind to themselves in a semblance of secure structure, but this makeshift jumble is inherently far less durable than that of double-stranded DNA.Because of mRNA’s infamous instability, the scientists reasoned it might be particularly susceptible to suffering at high temperatures. When they dosed the rice cells with a chemical marker that tags single-stranded mRNA molecules, Su and his colleagues discovered that heated plants contained higher proportions of unfurled mRNA compared to shoots lounging at room temperature. The head and tail ends of many mRNA strands appeared particularly fragile, accruing more of the incriminating chemical tags. In the heat of the moment, it seemed, these cellular messengers were melting.
 

 
 <div class="related"></div>
 

 
 When the researchers probed the plants further, they discovered that several of the mRNAs slumping into single strands were being chewed up by a destructive machine called the exosome. The looser the mRNA, the more likely it was to be ghosted out of existence—implying that mRNA unfolding made the messages <a href="http://rstb.royalsocietypublishing.org/content/373/1762/20180160">more prone to annihilation</a>. As the rice shoots attempted to weather the warmth, their exosomes were disposing of cells’ hottest messes, along with the genetic information they carried.Luckily, the short spikes in temperature didn’t appear to put any of the rice plants permanently out of commission. Because the cells’ DNA remained intact, new mRNAs were eventually copied and redistributed as the rice recovered from the sizzling onslaught. But it took about 10 hours for the shoots’ mRNAs to fully reset—quite the price to pay for a 10-minute hot flash.Over a thousand mRNA transcripts, all carrying instructions from different genes, became more or less abundant after the rise in temperature. Some of them even appeared to code for proteins that control how certain stretches of DNA are photocopied into mRNA—which means the effects of heat stress could be even more widespread than they already appear.
 <figure class="richtext-image full-width"><img alt="rice-1736909.jpg" class="richtext-image full-width" height="607" width="800"
 src="https://www.pbs.org/wgbh/nova/media/images/rice-1736909.width-800.jpg"/><figcaption>When subjected to just 10 minutes at a high temperature, messenger RNA in rice plants becomes unstable and is degraded by a molecular cleanup crew. Photo Credit: einszweifrei, Pixabay<div class="credit"></div></figcaption>
 </figure>
 Until the researchers learn more, it’s unclear what exactly these molecular fluxes signify, explains <a href="https://www.danforthcenter.org/scientists-research/principal-investigators/ru-zhang">Ru Zhang</a>, a plant biologist at the Donald Danforth Plant Science Center who was not involved in the study. They could be adaptive, allowing the rice to acclimate to the warmth. Or they could be indicative of just how vulnerable plants’ internal infrastructure is to temperature changes of this magnitude. Either way, resources were likely being diverted away from normal growth.“The paper is an elegant piece of science,” says <a href="https://agronomy.unl.edu/cassman">Kenneth Cassman</a>, an expert in food security and agriculture at the University of Nebraska, Lincoln who did not participate in the new research. “It’s a significant advance in helping us understand how short term exposure to heat affects rice… and how a plant grows and yields.”However, 10 minutes under high heat in a laboratory environment isn’t the same as chronic exposure to, say, the gradual onset of global warming. According to Cassman, there may not be a ton of overlap in the processes that govern plants’ responses to long-term changes in temperature and to short exposures like the one in this study—but there are still several key instances in which the new findings may translate directly.
 

 
 <div class="newsletter"></div>
 

 
 For one, the changes associated with climate change affect <a href="http://science.sciencemag.org/content/305/5686/994">both averages and extremes</a> in temperature, so daily spikes in heat could involve a 108º-Fahrenheit stint in the sun. Additionally, crops like rice and maize that are grown in high densities often experience temporary spikes in sun exposure when winds disrupt the foliage, says study author <a href="http://bio.psu.edu/directory/sma3">Sarah Assmann</a>, a plant biologist at Pennsylvania State University. Given the vast changes that occurred in the rice plants after just a few minutes of heat, even the ephemeral passing of a cloud overhead could disrupt some of these cells’ internal functions.Additionally, the researchers’ findings will likely translate over to other vital crops, Cassman says. Ironically, he points out, because most (<a href="http://ricepedia.org/rice-as-a-crop/where-is-rice-grown">but not all</a>) rice is grown in standing water, it’s actually one of the few crops that are a bit buffered against rising temperatures. But much of the plant, including its shoots, is warmed by direct sunlight, and rice yields have still been <a href="http://www.pnas.org/content/114/35/9326">curbed</a> by <a href="http://www.pnas.org/content/101/27/9971">global warming</a>. This could prove even more ominous for the vast majority of cultivated plants that don&#x27;t enjoy such liquid luxury, and remain especially vulnerable to surges of heat on dry land.Moving forward, Su, Bevilacqua, and Assmann hope to repeat their work in other strains of rice—including cultivars that have proven <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4052674/">hardy</a> in climes that tend to pack the heat. Comparing the responses of these different varieties to changes in temperature, along with the current findings, may help researchers target certain regions of plant genomes for manipulation in the future, Assmann explains, informing the production of more heat- and drought-resistant crops.
 <figure class="richtext-image full-width"><img alt="Corncobs.jpg" class="richtext-image full-width" height="533" width="800"
 src="https://www.pbs.org/wgbh/nova/media/images/Corncobs.width-800.jpg"/><figcaption>Maize is another staple crop that will continue to suffer from rising temperatures. The molecular changes that occur in rice plants are likely to translate over to other plants. Photo Credit: Sam Fentress, Wikimedia Commons<div class="credit"></div></figcaption>
 </figure>
 One possible intervention may already be apparent. It’s the sloppiest mRNAs (that is, those with particularly frayed heads and tails) that are especially susceptible to destruction, Bevilacqua explains. So genetic modifications that secure these ends—the molecular equivalent of clipping an aglet onto the end of a frazzled shoelace—might just increase their longevity.Even with exquisite knowledge of the coping mechanisms of plants in peril, however, there’s no guarantee a magic bullet will unveil itself, Cassman says. Plants have spent millennia adapting to heat, and researchers may not get much bang for their buck by trying to optimize their resilience further. If that’s the case, the ceiling on genetically engineering heat-resistant plants could be far lower than expected.But tinkering with the genes of crops is far from the only option, Bevilacqua says. “You can only go so far [with genetic engineering],” he adds. “Yes, it’s possible for us to intervene and help with plant production and response to these stresses as they increase. The bigger answer is to have policies to address and try to reverse climate change.”

India_-_Sights_&_Culture_-_Planting_Rice_Paddy_5_(3245008474).jpg

Women planting in a rice paddy in the Kancheepuram district of Tamil Nadu, India. Photo Credit: McKay Savage, Wikimedia Commons

Brief exposures to high temperatures compromise rice plants’ ability to relay genetic information.

The Planet’s Getting Warmer. Can Plants Take the Heat?

Somewhere along our evolutionary trajectory, we humans lost our regenerative powers. Sure, we can manage <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2701258/">a hunk of liver here</a>, <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5454966/">the tip of a finger there</a>—but in contrast to the <a href="https://www.pbs.org/wgbh/nova/article/saving-axolotls/">salamanders</a>, flatworms, and <a href="https://www.popularmechanics.com/science/animals/a27571/starfish-disease-science/">sea stars</a> of the world, the spontaneous remaking of entire organs and limbs eludes humankind.For biologist <a href="http://ase.tufts.edu/biology/labs/levin/">Michael Levin</a> of Tufts University, however, “lost” isn’t the same as “nonexistent.”Only a few million years separate humans from our closest regenerating relatives. If our animal ancestors built bone and tissue from scratch throughout adulthood, these superpowers may still lie buried within our own tissues, waiting to be reawakened. All it would take, says Levin, is the right signal.Now, Levin and his colleagues may have taken the next steps in pinpointing that very trigger by reviving these clandestine capabilities in African clawed frogs. Their work, <a href="https://www.cell.com/cell-reports/fulltext/S2211-1247(18)31573-0">published today</a> in the journal Cell Reports, kick-started the growth of partial but high-functioning hind legs in amphibian amputees after 24 hours of hormonal treatment. Though the work is still preliminary, such findings may someday inform researchers on the path to finding—and flipping—a similar switch in humans.
 

 
 <div class="ad"></div>
 

 
 Cut the tail off a tadpole, and <a href="https://www.nationalgeographic.com/science/phenomena/2013/01/13/tadpole-heal-thyself-vilified-chemicals-spur-regeneration/">another will take its place</a>. But even in the amphibian world, there are sobering consequences to growing up: If a mature frog loses a leg, the best it can hope for is a frail cartilaginous spike to take its place. This ability is technically regenerative—certainly far more than what any human could accomplish—but the resulting appendage does little for a creature that relies heavily on its hind limbs to navigate its aquatic environment.As the frog ages, its regenerative skills seem to go into hibernation—making it “a good stepping stone between animals that [fully] regenerate, like salamanders, and animals that don’t, like mammals,” Levin explains.To rouse regeneration from its slumber, the researchers first needed to identify the right molecular alarm. The ideal candidate, they reasoned, would be a powerful chemical cue that is both naturally produced by animal bodies and capable of directing a vast array of cellular functions—a winning combination that, for the researchers, pointed to progesterone.<div>
 <iframe width="480" height="270" src="https://www.youtube.com/embed/LfcJ13YYbTA?feature=oembed" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>
</div>
Every month, as a natural part of the female menstrual cycle, progesterone regenerates the entire endometrium, or lining of the uterus. “We thought, ‘It’s impossible that progesterone is only acting at this level,’” explains lead author <a href="http://sites.tufts.edu/celiaherrerarincon/">Celia Herrera-Rincon</a>, a neuroscientist at Tufts University.Progesterone wields far more power than simply preparing the uterus for pregnancy. Receptors for this hormone exist all over the bodies of both male and female mammals, and in the past several decades, researchers have shown it to be a potent catalyst for wound healing, promoting <a href="https://docplayer.net/74761056-Role-of-progesterone-in-peripheral-nerve-repair.html">nerve repair</a>, <a href="https://www.cell.com/developmental-cell/fulltext/S1534-5807(01)00064-8?_returnURL=https%3A%2F%2Flinkinghub.elsevier.com%2Fretrieve%2Fpii%2FS1534580701000648%3Fshowall%3Dtrue">bone remodeling</a>, and <a href="https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1524-475X.2008.00440.x">blood vessel formation</a>. If applied in the right place at the right time, progesterone could very well coax an injured limb onto a regenerative cascade.“Progesterone is very useful and safe,” explains <a href="https://www.popcouncil.org/research/expert/regine-sitruk-ware">Régine Sitruk-Ware</a>, a reproductive endocrinologist at the Population Council’s Center for Biomedical Research who did not participate in the new research. “It’s helpful for the uterus, but in many other layers of the body—in vessels, in bone, in the brain, and in the [nervous system] in general—progesterone has a positive, regenerative aspect.”
 

 
 <div class="related"></div>
 

 
 To harness the power of progesterone, the researchers designed a silicon apparatus filled with a hydrating gel and stitched it onto <a href="https://www.sciencedaily.com/releases/2016/10/161019155533.htm">African clawed frogs</a> (Xenopus laevis) whose right hind legs had been amputated three hours prior. Some of the devices also contained a dose of progesterone, which was steadily fed to the wound site for 24 hours. When the day was up, the team carefully detached the device—and waited.Within a matter of weeks, the progesterone-treated frogs had a serious leg up on their peers. In the absence of the hormone, frog amputees produced the typical featureless rods, regardless of whether or not they’d been fitted with the apparatus. But the frogs hit with progesterone began to sprout bigger, more structured appendages. And nearly ten months out from the initial spark of progesterone, partial, paddle-like limbs had pushed their way out of the wound site.When the researchers inspected the burgeoning limbs more closely, they found that, unlike the simple, cartilaginous spikes of typical amputees, they seemed to have the form and patterning of true legs. The limbs were thick and full of new blood vessels, nerves, and bone, and were forming what looked like tentative knees: the beginnings of actual joints. These frogs’ legs weren’t just regenerating more than their untreated counterparts—they were regenerating better. It’s a phenomenon that regenerative biologist <a href="http://firm.uni-frankfurt.de/john-barker">John Barker</a> of the Frankfurt Initiative for Regenerative Medicine praises as “smart proliferation”: the ability to form order out of the chaos of growing, dividing cells.
 <figure class="richtext-image full-width"><img alt="Herrera-Rincon Figures for Katie Wu.jpg" class="richtext-image full-width" height="247" width="800"
 src="https://www.pbs.org/wgbh/nova/media/images/Herrera-Rincon_Figures_for_Katie_Wu_F8vWy65.width-800.jpg"/><figcaption>Amputation of adult frogs&#x27; hind limbs usually results in the formation of a cartilaginous &quot;spike&quot; (left). But attaching a device that feeds progesterone to the amputation site for just 24 hours stimulates the growth of a paddle-like structure (right). The dashed line indicates the plane of amputation. Photo Credit: Celia Herrera-Rincon, Tufts University<div class="credit"></div></figcaption>
 </figure>
 “It was amazing that we were creating structures with only 24 hours of treatment,” Herrera-Rincon says. “I remember showing Mike [Levin] and being fascinated.”When the frogs got the chance to stretch their legs, the untreated, spike-toting regenerates proved to be feeble swimmers and spent more time loitering in place. Meanwhile, the animals that had worn the progesterone-dosing device glided merrily by, almost indistinguishable from their unamputated kin.Though the legs could function, the regeneration was far from complete. While the limb that appeared had clear features, the growth stopped abruptly before the frog’s hind digits would otherwise appear. And, importantly, not all the treated frogs responded to the same extent, indicating that the approach is unlikely to be a one-size-fits-all treatment. But for a first pass, Levin says, the amount of regeneration achieved was “really striking.”“If the person grew a limb like the frogs did, no one would say it was perfect,” says <a href="https://www.whitedlab.com/">Jessica Whited</a>, a regenerative biologist at Harvard University’s Brigham Regenerative Medicine Center who frequently collaborates with Levin’s lab, but was not involved in the newest finding. “But it’s important to demonstrate something that would be better than what grows naturally.”
 

 
 <div class="newsletter"></div>
 

 
 Levin and his team are currently tinkering with their cocktail of apparatus-administered drugs, as well as with the duration for which the device is worn after amputation. With just a few additions and adjustments, he says, the preliminary results are already looking good.“To rebuild any structure, you need to give the cells some signal about what they’re rebuilding,” Levin explains. “And progesterone [alone] gives an incomplete message.”The rest of this message, Whited adds, will involve a lot of fine-tuning. When limbs grow (and regrow) naturally, they rely on a suite of complex anatomical signals that both drive the production of new tissue and bone and tell it when and where to stop. But the domino effect likely still applies: The key will be finding the right recipe of chemicals to jumpstart the process, then sit back and let nature take its course.
 <figure class="richtext-image full-width"><img alt="4819444864_4a99e4cc5b_o.jpg" class="richtext-image full-width" height="545" width="800"
 src="https://www.pbs.org/wgbh/nova/media/images/4819444864_4a99e4cc5b_o.width-800.jpg"/><figcaption>Deer antlers are mammalian appendages capable of repeated rounds of regeneration. Male deer (and female reindeer) shed and sprout their antlers anew each year. Credit: USFWS Mountain-Prairie, flickr<div class="credit"></div></figcaption>
 </figure>
 “This isn’t ready for human trials… but it’s also not science fiction,” Levin says of the device. “I don’t think humans are barred from regeneration in any fundamental way.”We’re not exactly a hop, skip, and a jump away from regenerating limbs in people—even partial ones. Though frogs and humans share quite a few genes and basic anatomical structures, Barker explains, we’re still fundamentally very different creatures. But, even if the packaging varies from species to species, the basic toolkit of regeneration might translate across even fairly far-flung species. There are even <a href="/pages/videopages/extreme-animal-weapons/">mammals that naturally regenerate</a>: Deer regularly <a href="https://www.sciencedirect.com/science/article/pii/S135727251400226X">regrow their antlers</a> well into adulthood—a process that, not-so-coincidentally, is also driven by sex hormones.Even as technology advances, it’s unclear if humans will ever regenerate as slickly as salamanders. But maybe these able-bodied amphibians are inching us just a bit closer: One small leap for a frog may someday mean one giant step for humankind.

Xenopus_laevis_TimVickers_WikimediaCommons.JPG

Adult African clawed frogs (Xenopus laevis) usually develop spike-like structures at sites of limb amputation. But researchers can now equip these frogs with bioreactors that help them partially regenerate their hind legs. Photo Credit: TimVickers, Wikimedia Commons

Twenty-four hours of the hormone progesterone can kickstart over nine months of limb regeneration in amphibian amputees.

Katherine J. Wu

Articles