A network of ancient Roman roads converges neatly with satellite images of the Earth at night. A heat map of 15th-century bubonic plague outbreaks bears an eerie resemblance to Europe’s early COVID-19 hot spots. Mapping Past Societies, a free digital atlas hosted by the Initiative for the Science of the Human Past at Harvard, illuminates just a few of these patterns.

“It has a rich dataset of historic, economic, archaeological, environmental, and health information as well as climate data going back much further,” said Santiago Pardo Sánchez ’16, the project’s co-managing editor. “Someone who’s interested in modern transportation could look at how it worked in the past. Someone who’s looking at the plague in Central Asia could also get data from the Middle East.”

Mapping Past Societies, or MAPS, is powered by vast spreadsheets that geo-locate everything from historic rat populations to medieval marketplaces and Roman military structures. Its user-friendly interface, which runs on ArcGIS software, invites discovery by layering multiple phenomena across a single map — or by animating how one dataset plays out over time.

“The shipwreck data have been important to me and other economic historians,” said MAPS founder and general editor Michael McCormick, the Francis Goelet Professor of Medieval History and chair of the Science of the Human Past initiative. “They offer a rather crude but nevertheless rich indicator of economic activity for the period between about 500 B.C.E. and 1500 C.E.”

For much of his career, McCormick focused on the history and archaeology of the fall of the Roman Empire. He was once in the habit of hand-drawing maps for his classes. Then a thought occurred one evening in the 1990s while he was outlining the Roman Empire for an exam: “At this very moment, around the globe, there are probably 100 other professors drawing exactly the same map,” he recalled. “I said, ‘Wait! This is not a good use of our time. There should be one map!’”

Soon he was experimenting with geographic information systems to design his own digital maps, with several appearing in “Origins of the European Economy: Communications and Commerce, A.D. 300-900” (2002). That led to partnering with the Center for Geographic Analysis to launch the free Digital Atlas of Roman Empire and Medieval Civilizations in the mid ’00s. Over the years, DARMC was slowly expanded to incorporate new datasets. Information on the ancient and medieval worlds remains most robust, but more recent additions cover Colonial Latin America, 18th-century France, and more.

The pandemic inspired the team to refresh the project’s branding and interface, relaunching DARMC earlier this year as MAPS. The site’s new dashboard will be familiar to anybody who recalls tracking COVID cases on the Johns Hopkins website. At the same time, recent software updates enabled the addition of that showstopping layering feature.

“Before you could just turn on and off one layer,” Pardo Sánchez noted. “But now, you can do much more. You can change the visualizations with overlays and transparencies. You can share it more easily. You can switch the basemap, or background, to satellite imagery of the Earth at night.”

From the start, students have been key to the project’s success. Undergraduates bring a natural fascination with Roman and medieval history, McCormick said, but many struggle to make meaningful early academic contributions in the field, given the need for proficiency in multiple languages including Greek, Latin, Arabic, and Syriac — not to mention all the must-read secondary literatures in German, Italian, French, and Spanish.

To work on MAPS, however, all they need is curiosity, attention to detail, and facility with spreadsheets. “This is a real intellectual contribution to our understanding of the human past which they can, should, and do cite among their publications,” McCormick said.

“The undergrads working on the project now are younger than the project,” quipped Pardo Sánchez, who made his first MAPS contributions as an undergraduate, cataloguing findings from McCormick’s “Origins of the European Economy.” As an undergraduate concentrating in history, Pardo Sánchez later contributed to one of the site’s biggest datasets, nearly 60,000 records of climate events over the past 2,000 years.

Today, Anika Liv Christensen ’26, a research assistant currently working on MAPS, says, “It was the perfect job for a 19-year-old with no experience. Originally, my job was to check databases for errors. With so many entries, there are bound to be misspellings and formatting problems.”

Christensen, a joint concentrator in music and human evolutionary biology, recently worked on inventorying atypical burial sites in medieval France, currently with about 300 entries (each with up to five individuals per site).

“My litmus test has been: Could a nerdy 12-year-old use it?” Christensen said. “Because if a nerdy 12-year-old can use it, then anybody can.”

The enormous spreadsheets that populate the site’s map are freely available for download to anybody with an internet connection. The information on Roman roads has proved especially popular, McCormick shared. “There was a whole series of economic studies on 21st-century Europe showing that proximity to Roman roads helped predict economic vibrancy today.”

At a recent MAPS kickoff event, co-managing editor Alexander More, M.A ’07, Ph.D. ’14, an associate professor of environmental health at the University of Massachusetts, demonstrated what it looks like to plot the Roman roads alongside information on bubonic plague outbreaks from the 14th century.

“You can really see the data come alive,” he marveled. “For the first time, you can see the progression of this pandemic throughout Europe, with these hotspots emerging at nexuses of Roman roads.”

As bursts of yellow covered Italy and France, yet another historic intersection came into view. “These nexuses are in fact also the same places where COVID emerged in full force in 2020,” More said.

The effortless agility with which humans and animals move is an evolutionary marvel that no robot has yet been able to closely emulate. To help probe the mystery of how brains control and coordinate it all, Harvard neuroscientists have created a virtual rat with an artificial brain that can move around just like a real rodent.

Bence Ölveczky, professor in the Department of Organismic and Evolutionary Biology, led a group of researchers who collaborated with scientists at Google’s DeepMind AI lab to build a biomechanically realistic digital model of a rat. Using high-resolution data recorded from real rats, they trained an artificial neural network — the virtual rat’s “brain” — to control the virtual body in a physics simulator called MuJoco, where gravity and other forces are present. And the results are promising.

Published in Nature, the researchers found that activations in the virtual control network accurately predicted neural activity measured from the brains of real rats producing the same behaviors, said Ölveczky, who is an expert at training (real) rats to learn complex behaviors in order to study their neural circuitry. The feat represents a new approach to studying how the brain controls movement, Ölveczky said, by leveraging advances in deep reinforcement learning and AI, as well as 3D movement-tracking in freely behaving animals.

The collaboration was “fantastic,” Ölveczky said. “DeepMind had developed a pipeline to train biomechanical agents to move around complex environments. We simply didn’t have the resources to run simulations like those, to train these networks.”

Working with the Harvard researchers was, likewise, “a really exciting opportunity for us,” said co-author and Google DeepMind Senior Director of Research Matthew Botvinick. “We’ve learned a huge amount from the challenge of building embodied agents: AI systems that not only have to think intelligently, but also have to translate that thinking into physical action in a complex environment. It seemed plausible that taking this same approach in a neuroscience context might be useful for providing insights in both behavior and brain function.”

Graduate student Diego Aldarondo worked closely with DeepMind researchers to train the artificial neural network to implement what are called inverse dynamics models, which scientists believe our brains use to guide movement. When we reach for a cup of coffee, for example, our brain quickly calculates the trajectory our arm should follow and translates this into motor commands. Similarly, based on data from actual rats, the network was fed a reference trajectory of the desired movement and learned to produce the forces to generate it. This allowed the virtual rat to imitate a diverse range of behaviors, even ones it hadn’t been explicitly trained on.

These simulations may launch an untapped area of virtual neuroscience in which AI-simulated animals, trained to behave like real ones, provide convenient and fully transparent models for studying neural circuits, and even how such circuits are compromised in disease. While Ölveczky’s lab is interested in fundamental questions about how the brain works, the platform could be used, as one example, to engineer better robotic control systems.

A next step might be to give the virtual animal autonomy to solve tasks akin to those encountered by real rats. “From our experiments, we have a lot of ideas about how such tasks are solved, and how the learning algorithms that underlie the acquisition of skilled behaviors are implemented,” Ölveczky continued. “We want to start using the virtual rats to test these ideas and help advance our understanding of how real brains generate complex behavior.”

This research received financial support from the National Institutes of Health.

Using ancient DNA extracted from the toe bone of a museum specimen, Harvard biologists have sequenced the genome of an extinct, flightless bird called the little bush moa, shedding light on an unknown corner of avian genetic history.

Published in Science Advances, the work is the first complete genetic map of the turkey-sized bird whose distant living cousins include the ostrich, emu, and kiwi. It is one of nine known species of moa, all extinct for the last 700 years, which inhabited New Zealand before the late 1200s and the arrival of Polynesian human settlers.

“We’re pulling away the veil across the mystery of this species,” said senior author Scott V. Edwards, professor in the Department of Organismic and Evolutionary Biology and curator of ornithology at the Museum of Comparative Zoology. “We can study modern birds by looking at them and their behavior. With extinct species, we have very little information except what their bones looked like and in some cases what they ate. DNA provides a really exciting window into the natural history of extinct species like the little bush moa.”  

Bush moa were the smallest species of moa, weighing about 60 pounds and distributed in lowland forests across the north and south islands of New Zealand. Genomic analysis has revealed their closest living relatives aren’t kiwis, as was originally speculated, but rather tinamous, a Neotropical bird group from which they diverged genetically about 53 million years ago.

The research offers new genetic evidence for various aspects of bush moa sensory biology. Like many birds, they had four types of cone photoreceptors in their retinas, which gave them not only color but also ultraviolet vision. They had a full set of taste receptors, including bitter and umami.

Perhaps the most remarkable trait of these flightless birds is their complete absence of forelimb skeletal elements that typically comprise birds’ wings, the researchers wrote. Studying the moa genome could offer new clues into how and why some birds evolved to become flightless.

The scientists used high-throughput DNA sequencing, which allows rapid sequencing of short DNA fragments. To produce the bush moa genome, the team sequenced the equivalent of 140 bird genomes, or about 140 billion base pairs of DNA, only about 12 percent of which was actual moa DNA (the rest was bacterial).

They then pieced together the genome, taking each snippet of DNA and mapping it to its correct position. Assembly of extinct species is painstaking work, which has gotten a big boost from technologies like high-throughput sequencing. Other species that have been mapped similarly are the passenger pigeon, the woolly mammoth, and our close relative, the Neanderthal. Using an existing emu genome as a guide, researchers strung together the bush moa’s genetic sequence by finding overlaps between each genetic snippet, essentially reconstructing a long puzzle of 140 billion pieces.

The project originated more than 15 years ago in the lab of the late Allan J. Baker. An expert in ancient bird DNA at the Royal Ontario Museum, Baker was the first to extract and sequence the bird’s DNA from a fossil recovered on the South Island of New Zealand.

Also involved in the initial DNA processing and sequencing was Alison Cloutier, a co-author of the new paper, who formerly worked with Baker and later became a postdoctoral researcher in Edwards’ lab at Harvard, which inherited the data/research.

Reconstructing the genome of a long-extinct bird fills in a new branch of the avian family tree, opening doors to study avian evolution, or even someday, to possibly resurrect these species through de-extinction technologies.

“To me, this work is all about fleshing out the natural history of this amazing species,” Edwards said.