Monday, October 1, 2018

Songbird data yields new theory for learning sensorimotor skills

"Our findings suggest that an animal knows that even the perfect neural command is not going to result in the right outcome every time," says Emory biophysicist Ilya Nemenman. (Image courtesy Samuel Sober.)

By Carol Clark

Songbirds learn to sing in a way similar to how humans learn to speak — by listening to their fathers and trying to duplicate the sounds. The bird’s brain sends commands to the vocal muscles to sing what it hears, and then the brain keeps trying to adjust the command until the sound echoes the one made by the parent.

During such trial-and-error processes of sensorimotor learning, a bird remembers not just the best possible command, but a whole suite of possibilities, suggests a study by scientists at Emory University.

The Proceedings of the National Academy of the Sciences (PNAS) published the study results, which include a new mathematical model for the distribution of sensory errors in learning.

“Our findings suggest that an animal knows that even the perfect neural command is not going to result in the right outcome every time,” says Ilya Nemenman, an Emory professor of biophysics and senior author of the paper. “Animals, including humans, want to explore and keep track of a range of possibilities when learning something in order to compensate for variabilities.”

Nemenman uses the example of learning to swing a tennis racket. “You’re only rarely going to hit the ball in the racket’s exact sweet spot,” he says. “And every day when you pick up the racket to play your swing is going to be a little bit different, because your body is different, the racket and the ball are different, and the environmental conditions are different. So your body needs to remember a whole range of commands, in order to adapt to these different situations and get the ball to go where you want.”

First author of the study is Baohua Zhou, a graduate student of physics. Co-authors include David Hofmann and Itai Pinkoviezky (post-doctoral fellows in physics) and Samuel Sober, an associate professor of biology.

Traditional theories of learning propose that animals use sensory error signals to zero in on the optimal motor command, based on a normal distribution of possible errors around it — what is known as a bell curve. Those theories, however, cannot explain the behavioral observations that small sensory errors are more readily corrected, while the larger ones may be ignored by the animal altogether.

For the PNAS paper, the researchers analyzed experimental data on Bengalese finches collected in previous work with the Sober lab. The lab uses finches as a model system for understanding how the brain controls complex vocal behavior and motor behavior in general.

Miniature headphones were custom-fitted to adult birds and used to provide auditory feedback in which the pitch that the bird perceives it vocalizes at could be manipulated, replacing what the bird hears — its natural auditory feedback — with the manipulated version. The birds would try to correct the pitch they were hearing to match the sound they were trying to make. Experiments allowed the researchers to record and measure the relationship between the size of a vocal error the bird perceives, and the probability of the brain making a correction of a specific size.

The researchers analyzed the data and found that the variability of errors in correction did not have the normal distribution of a bell curve, as previously proposed. Instead, the distribution had long tails of variability, indicating that the animal believed that even large fluctuations in the motor commands could sometimes produce a correct pitch. The researchers also found that the birds combined their hypotheses about the relationship between the motor command and the pitch with the new information that their brains received from their ears while singing. In fact, they did this surprisingly accurately.

“The birds are not just trying to sing in the best possible way, but appear to be exploring and trying wide variations,” Nemenman says. “In this way, they learn to correct small errors, but they don’t even try to correct large errors, unless the large error is broken down and built up gradually.”

The researchers created a mathematical model for this process, revealing the pattern of how small errors are corrected quickly and large errors take much longer to correct, and might be neglected altogether, when they contradict the animal’s “beliefs” about the errors that its sensorimotor system can produce.

“Our model provides a new theory for how an animal learns, one that allows us to make predictions for learning that we have tested experimentally,” Nemenman says.

The researchers are now exploring if this model can be used to predict learning in other animals, as well as predicting better rehabilitative protocols for people dealing with major disruptions to their learned behaviors, such as when recovering from a stroke.

The work was funded by the National Institutes of Health BRAIN Initiative, the James S. McDonnell Foundation, and the National Science Foundation. The NVIDIA corporation donated high-performance computing hardware that supported the work.

Related:
BRAIN grant to fund study of how the mind learns
How songbirds learn to sing

Friday, September 21, 2018

Climate change calls for a fresh approach to water woes

An egret spreads its wings above waters of the Everglades. "Climate change is a game changer" when it comes to managing major water basins across the country, says Lance Gunderson, chair of Emory's Department of Environmental Sciences.

By Carol Clark

The Everglades National Park, the largest subtropical wilderness in the United States, is home to 16 different species of wading birds and rare and endangered species like the manatee, the American crocodile and the Florida panther. But the area is also home to humans. The park is a portion of a larger wetland ecosystem, more than half of which has been converted into agricultural production or urban developments. The ecosystem must provide both flood protection and supply water for the park, the agricultural interests and South Forida’s rapidly growing population of nearly eight million people.

Meanwhile, a federal-state initiative to address this challenge, known as the Comprehensive Everglades Restoration Plan, is “sort of stuck in the muddle,” says Lance Gunderson, chair of Emory’s Department of Environmental Sciences. The plan was authorized in 2000 but it hasn’t made much progress.

Climate change throws another wrench in the works, affecting the Everglades and other large watersheds across the United States in new and unpredictable ways. Extreme weather events and rising sea levels, combined with a growing population, will lead to “more intense arguments” about already contested issues of water quality and water usage, Gunderson says.

Gunderson, a wetlands ecologist, recently partnered with Barbara Cosens, a legal scholar at the University of Idaho, to lead an interdisciplinary team of researchers in a project to assess the adaptive capacity of six major U.S. water basins to changing climates. In addition to the Everglades, the basins include the Anacostia, the Columbia, the Klamath, the Platte and the Rio Grande rivers. The project was funded by NSF Social-Ecological Synthesis Center at the University of Maryland. (Watch the video below to learn more.)



“Climate change is a game changer when it comes to the management of these regional-scale water systems across the country,” Gunderson says. “These systems are managed through assumptions about climate and models that are based on averages. Now, managers are struggling to adapt to more extremes — like earlier snow melts, more floods and droughts, and more intense storms.”

Even without extreme events, water management is complex. The Everglades, for example, is not just an issue of restoring biological diversity. It’s an economic problem that often puts government agencies, agriculture, developers, residents, and environmental groups at loggerheads.

“These are complex problems and we can’t plan or analyze our way out of them,” Gunderson says. “We have to learn our way out of them.”

Instead of relying on the court system or government policies, he says people need to come together in organic, self-organized ways for “adaptive governance.” Such approaches can forge new paths through a problem by trying small experiments to see if they work.

The Klamath River basin, for instance, benefited by farmers and Native Americans coming together informally, instead of going to court, to talk about possible ways to reallocate water to satisfy both sides.

“Informal, adaptive management lets you learn while you’re doing,” Gunderson says. “It allows people without resources to be engaged in the process. Change happens when little groups of people work together collectively on wicked problems that have no easy solutions or easy answers.”

As chair of Environmental Sciences Gunderson is also confronted with the problem of how to train students to deal with the issues that will face them when they graduate. The department is blending facets of political science, ecology, sociology, biology, geology and health into its curriculum.

“These specialties are at the intersection of major environmental problems and we are trying to build some integrated understanding around them,” Gunderson says. “Our world is becoming more complex and we want students to have the skills to confront that complexity.”

Related:
Students develop device to help cope with climate change
Responding to climate change
The growing role of farming and nitrous oxide in climate change
Putting people into the climate change picture

Tuesday, August 28, 2018

The math of malaria: Drug resistance 'a numbers game' of competing parasites

"Computer models can sometimes give you insights that would be too difficult to get in a real-world setting," says Mary Bushman. She developed a malaria model for her PhD thesis, advised by Emory evolutionary biologist Jaap de Roode. (Ann Watson, Emory Photo/Video)

By Carol Clark

A new mathematical model for malaria shows how competition between parasite strains within a human host reduces the odds of drug resistance developing in a high-transmission setting. But if a drug-resistant strain does become established, that same competition drives the spread of resistance faster, under strong selection from antimalarial drug use.

“It’s basically a numbers game,” says Mary Bushman, who developed the model for her PhD thesis in Emory University’s Population Biology, Ecology and Evolution Graduate Program. “When you already have multiple strains of malaria within a population, and a drug-resistant strain comes along, it will usually go extinct simply because it’s a late-comer. Whichever strain is there first has the advantage.”

PLOS Biology published the findings, a computational framework that modeled a malaria epidemic across multiple scales: Transmission of parasites from mosquitos to humans, and the dynamics of parasites competing to infect blood cells while they also battle the immune system of a human host.

After creating the model, Bushman ran simulations tracking malaria in a population for roughly 14 years. The simulations included 400 theoretical people who were randomly bitten by 12,000 mosquitos that were infected with malaria parasites classified as either drug resistant or drug susceptible. Various levels of treatment with antimalarial drugs were also part of the simulations.

“Our model holds strong relevance for infectious diseases beyond malaria,” says Jaap de Roode, an evolutionary biologist at Emory and senior author of the paper. “We hope this research gives others a method to look at disease dynamics across scales of biological organisms to learn how drug resistance develops in a range of pathogens.”

The study’s authors also include Emory biologist Rustom Antia (a specialist in infectious disease modeling) and Venkatachalam Udhayakumar, a malaria expert from the Centers of Disease Control and Prevention’s Division of Parasitic Diseases and Malaria.

The researchers are now working to develop their specific model for malaria into a generalized software tool for infectious diseases. “Computer models can sometimes give you insights that would be too difficult to get in a real-world setting,” says Bushman, who is now a post-doctoral fellow in the Antia lab.

"The distinction between establishment and spread just jumped out of the data," Bushman says.

Malaria occurs in poor, tropical and subtropical areas of the world, although most of the global death toll consists of children from sub-Saharan Africa. People infected in this high-transmission area often have multiple strains of the parasite and, by the time they have reached adulthood, they have usually developed partial immunity.

“It’s a baffling disease,” Bushman says. “Malaria has been studied for more than 100 years, much longer than most diseases, but there is still a lot that we don’t understand about it."

Malaria is caused by several species of Plasmodium parasites that are transmitted to humans by mosquitos. Plasmodium falciparum, the most common malaria parasite on the continent of Africa, is the one responsible for the most malaria-related deaths globally.

P. falciparum has developed resistance to former first-line therapies chloroquine and sulfadoxine-pyrimethamine. Resistance has also emerged in Southeast Asia to the third and last available treatment, artemisinin combination therapy, or ACT.

One of the mysteries about malaria is why drug-resistant strains tend to emerge first in low-transmission areas, like Southeast Asia, and not appear until much later in Africa, where transmission is high.

Previous research led by de Roode and Bushman showed that when people are co-infected with drug-resistant and drug-sensitive strains of malaria, both strains are competitively suppressed.

For the current paper, the researchers wanted to get a more detailed understanding of these dynamics. Some evidence had shown that within-host competition could suppress resistance, while other studies showed that it could ramp resistance up.

“It was a little bit of a puzzle, why the findings were conflicting,” Bushman says.

The new model, driven by evidence for how malaria parasites work within the immune system and the blood cells they infect, provided a solution to the puzzle.

“Some previous models were based on the assumption that when you put two strains of malaria into a host, they split 50-50,” Bushman says. “But our model showed that the system is asymmetrical. When you put two strains in a host they virtually never split 50-50.”

The late-comer will usually go extinct, which explains why in high-transmission areas drug resistant strains are at a big disadvantage. But in low transmission areas, such as Southeast Asia, a drug resistant strain has a better chance of arriving first in a host and getting established.

The new model also showed how once a drug-resistant strain becomes established in a high transmission area, it will spread much faster than it would in a low transmission area.

“The distinction between establishment and spread just jumped out of the data,” Bushman says. “Our model validated both sides of the argument — that within-host dynamics of competing parasites could both repress and accelerate the spread of resistance. The phenomena are occurring at different stages of the process so they both can happen.”

The results offer a new explanation for why chloroquine resistance arrived relatively late in Africa, appearing in Kenya and Tanzania in 1978, but then spread rapidly across the continent.

Related:
Mixed-strain malaria infections influence drug resistance
Zeroing in on 'super spreaders' and other hidden patterns of epidemics

Monday, August 27, 2018

Sensitivity to how others evaluate you emerges by 24 months

"Image management is fascinating to me because it's so important to being human," says Sara Valencia Botto, shown posing with a toddler.  The Emory graduate student published a study on how toddlers are attuned to image, along with psychology professor Philippe Rochat. (Kay Hinton, Emory Photo/Video)

By Carol Clark

Even before toddlers can form a complete sentence, they are attuned to how others may be judging them, finds a new study by psychologists at Emory University.

The journal Developmental Psychology is publishing the results, documenting that toddlers are sensitive to the opinions of others, and that they will modify their behavior accordingly when others are watching.

“We’ve shown that by the age of 24 months, children are not only aware that other people may be evaluating them, but that they will alter their behavior to seek a positive response,” says Sara Valencia Botto, an Emory PhD candidate and first author of the study.

While previous research has documented this behavior in four- to five-year-olds, the new study suggests that it may emerge much sooner, Botto says.

“There is something specifically human in the way that we’re sensitive to the gaze of others, and how systematic and strategic we are about controlling that gaze,” says Philippe Rochat, an Emory professor of psychology who specializes in childhood development and senior author of the study. “At the very bottom, our concern for image management and reputation is about the fear of rejection, one of the main engines of the human psyche.”

This concern for reputation manifests itself in everything from spending money on makeup and designer brands to checking how many “likes” a Facebook post garners.

“Image management is fascinating to me because it’s so important to being human,” Botto says. “Many people rate their fear of public speaking above their fear of dying. If we want to understand human nature, we need to understand when and how the foundation for caring about image emerges.”

The researchers conducted experiments involving 144 children between the ages of 14 and 24 months using a remotely controlled robot toy.

In one experiment, a researcher showed a toddler how to use the remote to operate the robot. The researcher then either watched the child with a neutral expression or turned away and pretended to read a magazine. When the child was being watched, he or she showed more inhibition when hitting the buttons on the remote than when the researcher was not watching.

In a second experiment, the researcher used two different remotes when demonstrating the toy to the child. While using the first remote, the researcher smiled and said, “Wow! Isn’t that great?” And when using the second remote, the researcher frowned and said “Uh-oh! Oops, oh no!” After inviting the child to play with the toy, the researcher once again either watched the child or turned to the magazine.

The children pressed the buttons on the remote associated with the positive response from the researcher significantly more while being watched. And they used the remote associated with the negative response more when not being watched.

During a third experiment, that served as a control, the researcher gave a neutral response of “Oh, wow!” when demonstrating how to use the two remotes. The children no longer chose one remote over the other depending on whether the researcher was watching them.

The control experiment showed that in the second experiment the children really did take into account the values expressed by the experimenter when interacting with the toy, and based on those values changed their behavior depending on whether they were being watched, Botto says.

A final experiment involved two researchers sitting next to one another and using one remote. One researcher smiled and gave a positive response, “Yay! The toy moved!” when pressing the remote. The second researcher frowned and said, “Yuck! The toy moved!” when pressing the same remote. The child was then invited to play with the toy while the two researchers alternated between either watching or turning their back to the child. Results showed that the children were much more likely to press the remote when the researcher who gave the positive response was watching.

“We were surprised by the flexibility of the children’s sensitivity to others and their reactions,” Botto says. “They could track one researcher’s values of two objects and two researchers’ values of one object. It reinforces the idea that children are usually smarter than we think.”

Botto is continuing to lead the research in the Rochat lab for her PhD thesis. She is now developing experiments for children as young as 12 months to see if the sensitivity to being evaluated by others emerges even earlier than the current study documents.

And she is following the 14- to 24-month-old children involved in the published study, to see if the individual differences they showed in the experiments are maintained as they turn four and five. The researchers are measuring social and cognitive factors that may have predictive power for individual differences — such as language ability, temperament and a child’s ability to pick up on social norms and to understand that people can have beliefs different from their own.

“Ultimately, we hope to determine exactly when children begin to be sensitive to others’ evaluations and the social and cognitive factors that are necessary for that sensitivity to emerge,” Botto says.

Such basic research may translate into helping people in a clinical environment who are at the extremes of the spectrum of such sensitivity, she adds.

“It’s normal and necessary to a certain extent to care about our image with others,” Botto says. “But some people care so much that they suffer from social anxiety, while others care so little that it is not optimal in a society where cooperation is essential.”

The American Psychological Association contributed to this report. 

Related:
Babies have logical reasoning before age one, study finds
Babies' spatial reasoning predicts later math skills

Wednesday, August 22, 2018

Students develop personal cooling device to help cope with climate change

The Vimband was developed by Emory undergraduates Ryan James, Jesse Rosen-Gooding and Hieren Helmn, in the hopes of winning the Hult Prize.

A trio of Emory students is on a globe-trotting million-dollar quest this summer to address one of the world’s most urgent challenges — helping people find physical comfort in the face of climate change.

One answer, they believe, might be the “Vimband,” their idea for a personal temperature-regulation device that could be worn to cool the body in extremely hot weather or warm individuals enduring severely cold temperatures.

Amid scientific reports that global temperatures are climbing, direct body cooling could go far in providing personal relief, especially for populations living in increasingly hot climates, says Ryan James, a sophomore from Highland, Maryland, majoring in business and computer science, who convened a team of Emory students eager to pose a solution to the problem.

“World-wide, the use of air-conditioning is expected to nearly triple by 2050, and with detrimental environmental effects, that isn’t a sustainable solution,” James says. “There needs to be an alternative.”

So instead of controlling the temperatures of large buildings or residences, the Emory team set their sights on a smaller, more efficient target — the individual. Together, they’ve created a prototype for a rechargeable device that essentially functions as a small, personalized heating and cooling unit. The compact box may be worn around the wrist, neck or head — pulse points on the human body near major arteries that play a critical role in regulating body temperature.

Click here to read more about the Vimband, and the students' quest to win the the Hult Prize, an annual business innovation challenge open to students around the world.