Surprising electronic disorder in a copper oxide-based ceramic

Cuprates, a class of copper-oxide ceramics that share a common building block of copper and oxygen atoms in a flat square lattice, have been studied for their ability to be superconducting at extremely high temperatures. In their pristine state, however, they are a special kind of insulator (a material that does not readily conduct electricity) known as a Mott insulator. 

When electrical charge carriers — either electrons or the lack of electrons, known as "holes" — are added to an insulator in a process called doping, the insulator may become a metal, which readily conducts electricity, or a semiconductor, which can conduct electricity depending on the environment. Cuprates, however, behave neither like a normal insulator nor like a normal metal because of strong interactions between their electrons. To avoid the large energy cost arising from these interactions, the electrons spontaneously organize in a collective state where the motion of each particle is tied to all the other ones. 

One example is the superconducting state, where electrons move in unison and drift with zero net friction when a potential is applied, a zero-resistance state which is a defining characteristic of a superconductor. Another collective electronic state is a “charge density wave,” a term coined from the wave-like modulation in the density of electrons, in which electrons “freeze” into periodic and static patterns, at the same time hindering electron flow. This state is therefore antagonistic to the superconducting state, and, therefore, important to study and understand. In cuprates, charge-density-waves prefer to align to the atomic rows of copper and oxygen atoms that make up the underlying crystal structure, with wave “‘crests” occurring every three to five unit cells, depending on the material and doping level.

Using a technique known as resonant X-ray scattering to study these charge-density-waves in two different cuprate compounds, neodymium copper oxide (Nd₂CuO₄ or NCO) and praseodymium copper oxide (Pr₂CuO₄ or PCO) doped with extra electrons, MIT researchers made an unexpected discovery. Their work revealed a phase of the material where the electrons fall into a disordered, or “glassy,” arrangement, dubbed a “Wigner glass.” The results were recently published in a paper in Nature Physics.

Resonant X-ray scattering is a recently developed diffraction technique in which crystallography is performed on electrons rather than exclusively on the atoms as in conventional X-ray diffraction. “In the limit of low concentration of doped electrons, we observed a completely new and unexpected form of electronic phase which is neither a superfluid nor a crystal, but it rather has the characteristics of a Wigner glass. In this phase, the electrons form a collective state without any orientational preference,” says the paper’s senior author Riccardo Comin, assistant professor of physics at MIT. Such an amorphous glass of electrons is completely unprecedented in this family of materials, he adds. 

This phenomenon emerges only in a narrow window of electron doping. “Intriguingly, this exotic new state only exists in a small region of the electronic phase diagram of this material, and when more electrons are doped in the [copper oxide] planes, a more conventional electronic crystal is recovered, whose ripples align to the crystallographic axes of the underlying atomic lattice,” Min Gu Kang, the paper’s lead author, explains.

The MIT team, consisting of Comin, graduate student Kang, and postdoc Jonathan Pelliciari, designed the project and led the majority of experiments. Their research was made possible by the contributions of researchers at various institutions and facilities worldwide. Resonant X-ray scattering measurements were performed at multiple synchrotron facilities including the Berlin Electron Storage Ring in Germany, the Canadian Light Source in Saskatoon, Saskatchewan, Canada, and the Advanced Light Source, in Berkeley, California. The copper-oxide thin film samples were grown at NTT Basic Research Laboratories in Japan. Theoretical analysis was developed by researchers at the Indian Institute of Science in India.

Comin notes that the proposed theory explains the role of the electronic band structure in governing the periodic spacing and lack of orientational preference of the density waves as a function of doping level in this material. “Our theory suggests that these electronic ripples are initially formed with irregular shapes and are likely nucleated around defects or impurities in the material,” Comin says. “When the density of carriers increases, the electrons manage to find a more highly-ordered arrangement that minimizes the total energy of the system, thereby restoring the more conventional charge density waves that have been observed universally in all families of copper-oxide superconductors.”

“I was completely blown away by Riccardo's results on NCO and PCO,” says Peter Abbamonte, Fox Family Professor in Engineering at the University of Illinois at Urbana-Champaign, who developed the resonant soft X-ray scattering technique. Noting that charge density wave (CDW) order in cuprates has been at the center of the field for well over a decade, Abbamonte, who was not involved in this research, explains that the previous understanding has been that the CDW order is pinned to the crystal lattice, meaning the charge density wave must point in either of two perpendicular directions, but nowhere in between. This conventional wisdom is built on two decades of resonant scattering and scanning tunneling microscopy experiments that have always found this to be the case, he notes.

Comin’s research on these particular electron-doped cuprates showed that during the glassy phase the charge order can point in any direction, independent of the crystal lattice it lives in. “The more precise statement is that the CDW order parameter is not Ising-like (that is, taking only discrete values, in this case two: x or y), as has always been assumed, but is more like an X-Y order parameter (that is, free to choose any value on a continuous range, such as all directions between x and y as is the case here) that is only weakly influenced by the crystal,” Abbamonte says.
 
“It is going to take some time for the community to fully digest this realization and its implications for understanding the relevance of CDW order,” Abbamonte adds. “What is clear is Riccardo's paper is going to lead to a serious re-reckoning of the rules of the game, and in this sense is a major advance for the field.” 

Superconductors have an immense, largely untapped potential for transformative applications such as quantum computing, lossless energy transport, magnetic sensing and medical diagnostic imaging, and plasma and nuclear fusion power technologies.

“Overall, our study has revealed yet another manifestation of the exquisite quantum character of charge carriers in high-temperature superconductors, which ultimately arises from the nature of the electronic interactions,” Comin says. “The detailed behavior of electrons uncovered in this work provides new insights on how high-temperature superconductivity is born out of a Mott insulator, and promises to bridge a gap between regions of the phase diagram with very contrasting phenomenologies.” 

de MIT News http://bit.ly/2ThDXW3

Converting Wi-Fi signals to electricity with new 2-D materials

Imagine a world where smartphones, laptops, wearables, and other electronics are powered without batteries. Researchers from MIT and elsewhere have taken a step in that direction, with the first fully flexible device that can convert energy from Wi-Fi signals into electricity that could power electronics.

Devices that convert AC electromagnetic waves into DC electricity are known as “rectennas.” The researchers demonstrate a new kind of rectenna, described in a study appearing in Nature today, that uses a flexible radio-frequency (RF) antenna that captures electromagnetic waves — including those carrying Wi-Fi — as AC waveforms.

The antenna is then connected to a novel device made out of a two-dimensional semiconductor just a few atoms thick. The AC signal travels into the semiconductor, which converts it into a DC voltage that could be used to power electronic circuits or recharge batteries.

In this way, the battery-free device passively captures and transforms ubiquitous Wi-Fi signals into useful DC power. Moreover, the device is flexible and can be fabricated in a roll-to-roll process to cover very large areas.

“What if we could develop electronic systems that we wrap around a bridge or cover an entire highway, or the walls of our office and bring electronic intelligence to everything around us? How do you provide energy for those electronics?” says paper co-author Tomás Palacios, a professor in the Department of Electrical Engineering and Computer Science and director of the MIT/MTL Center for Graphene Devices and 2D Systems in the Microsystems Technology Laboratories. “We have come up with a new way to power the electronics systems of the future — by harvesting Wi-Fi energy in a way that’s easily integrated in large areas — to bring intelligence to every object around us.”

Promising early applications for the proposed rectenna include powering flexible and wearable electronics, medical devices, and sensors for the “internet of things.” Flexible smartphones, for instance, are a hot new market for major tech firms. In experiments, the researchers’ device can produce about 40 microwatts of power when exposed to the typical power levels of Wi-Fi signals (around 150 microwatts). That’s more than enough power to light up an LED or drive silicon chips.

Another possible application is powering the data communications of implantable medical devices, says co-author Jesús Grajal, a researcher at the Technical University of Madrid. For example, researchers are beginning to develop pills that can be swallowed by patients and stream health data back to a computer for diagnostics.

“Ideally you don’t want to use batteries to power these systems, because if they leak lithium, the patient could die,” Grajal says. “It is much better to harvest energy from the environment to power up these small labs inside the body and communicate data to external computers.”

All rectennas rely on a component known as a “rectifier,” which converts the AC input signal into DC power. Traditional rectennas use either silicon or gallium arsenide for the rectifier. These materials can cover the Wi-Fi band, but they are rigid. And, although using these materials to fabricate small devices is relatively inexpensive, using them to cover vast areas, such as the surfaces of buildings and walls, would be cost-prohibitive. Researchers have been trying to fix these problems for a long time. But the few flexible rectennas reported so far operate at low frequencies and can’t capture and convert signals in gigahertz frequencies, where most of the relevant cell phone and Wi-Fi signals are.

To build their rectifier, the researchers used a novel 2-D material called molybdenum disulfide (MoS₂), which at three atoms thick is one of the thinnest semiconductors in the world. In doing so, the team leveraged a singular behavior of MoS₂: When exposed to certain chemicals, the material’s atoms rearrange in a way that acts like a switch, forcing a phase transition from a semiconductor to a metallic material. The resulting structure is known as a Schottky diode, which is the junction of a semiconductor with a metal.

“By engineering MoS₂ into a 2-D semiconducting-metallic phase junction, we built an atomically thin, ultrafast Schottky diode that simultaneously minimizes the series resistance and parasitic capacitance,” says first author and EECS postdoc Xu Zhang, who will soon join Carnegie Mellon University as an assistant professor.

Parasitic capacitance is an unavoidable situation in electronics where certain materials store a little electrical charge, which slows down the circuit. Lower capacitance, therefore, means increased rectifier speeds and higher operating frequencies. The parasitic capacitance of the researchers’ Schottky diode is an order of magnitude smaller than today’s state-of-the-art flexible rectifiers, so it is much faster at signal conversion and allows it to capture and convert up to 10 gigahertz of wireless signals.

“Such a design has allowed a fully flexible device that is fast enough to cover most of the radio-frequency bands used by our daily electronics, including Wi-Fi, Bluetooth, cellular LTE, and many others,” Zhang says.

The reported work provides blueprints for other flexible Wi-Fi-to-electricity devices with substantial output and efficiency. The maximum output efficiency for the current device stands at 40 percent, depending on the input power of the Wi-Fi input. At the typical Wi-Fi power level, the power efficiency of the MoS₂ rectifier is about 30 percent. For reference, today’s rectennas made from rigid, more expensive silicon or gallium arsenide achieve around 50 to 60 percent.

There are 15 other paper co-authors from MIT, Technical University of Madrid, the Army Research Laboratory, Charles III University of Madrid, Boston University, and the University of Southern California.

The team is now planning to build more complex systems and improve efficiency. The work was made possible, in part, by a collaboration with the Technical University of Madrid through the MIT International Science and Technology Initiatives (MISTI). It was also partially supported by the Institute for Soldier Nanotechnologies, the Army Research Laboratory, the National Science Foundation’s Center for Integrated Quantum Materials, and the Air Force Office of Scientific Research.

de MIT News http://bit.ly/2G7jgsi

The gift of light, through science and service

Pawan Sinha, a professor of vision and computational neuroscience in the Department of Brain and Cognitive Sciences, first met Poonam when the girl was 13 years old. She lived in a remote village far from the urban bustle of Delhi, the second most-populous city in India. Poonam had grown up among lush trees and straw-thatched roofs and white cows with pointy horns grazing languidly along the dirt-packed road. But she had never seen any of it. Like many of the patients Sinha has worked with since he began Project Prakash, Poonam had been blind since birth due to dense cataracts, a treatable form of blindness more often seen in elderly patients.

Project Prakash, from the Sanskrit word for light, is a nonprofit organization that provides surgeries to congenitally blind children in India, and observes them during recovery to track the development of sight in the brain. Volunteers from Project Prakash travel to remote villages with limited access to health care to screen blind children and, if they are eligible, enroll them in the program.

Sinha first conceived of Project Prakash in 2002 while visiting his father in Delhi. He encountered two young siblings, living in poverty on the city streets, both blinded by treatable cataracts.

“This opened my eyes to the pervasiveness of childhood blindness in India, and more broadly in the developing world,” says Sinha. “Many children have treatable forms of blindness, but they stay blind because of lack of access to medical facilities, lack of knowledge of treatment options, and lack of financial resources to pursue medical care. These children languish in their blindness and lead difficult lives with little education, almost no prospects for employment and sadly, in many cases they die very young.”

Volunteers transported Poonam from her village to Dr. Shroff’s Charity Eye Hospital in Delhi, where, after a full examination, she received cataract removal surgery. The very next day, the bandages came off. During Poonam’s post-operative check-up, the caregiver held up her fingers both close to her face and from a distance and asked how many. Poonam answered correctly. She could see.

“The name Prakash reflects our immediate goal to bring light into the lives of children who suffer from blindness,” says Sinha. “In meeting this humanitarian need, as a neuroscientist I realized we had an scientific opportunity. With surgery, we can transition a child from blindness to sight in less than an hour and from the very moment the child’s bandages are removed, you have a ringside seat into the process of visual development.”

The call to action

According to data collected by Project Prakash, current estimates suggest that between 200,000 and 700,000 children suffer from potentially treatable forms of blindness, such as cataracts or corneal opacities. Only 50 percent of these children are expected to survive into adulthood, and many suffer physical or sexual abuse at some point in their lives.

Compelled to act, Sinha quickly realized he wanted to create a lasting impact beyond a one-time donation to cover the cost of cataract surgeries for individual children. From the perspective of a vision researcher, Sinha felt he was in a unique position to tackle this problem on a larger scale. He recognized an opportunity for synergy between this profound humanitarian need and a longstanding neuroscientific question he was interested in as a researcher. It would also address a common problem that arises when studying visual development.

“Until now, almost all the approaches we had to observe visual development happening in real-time was by working with infants, and they are notoriously difficult to work with experimentally,” says Sinha. “Not only are they unable to follow instructions or report out, their brains have many developmental processes progressing simultaneously, so it is hard to isolate one from the others. Visual development proceeds very rapidly, so we have a very short window to work with before infants become very sophisticated visual perceivers.”

And that is one of the key scientific benefits of Project Prakash. Older blind children, like Poonam and the kids Sinha met while visiting with his father, have physiologically mature brains, but have not yet been exposed to patterns in the visual world. This makes it much easier to identify the processes of visual development as they unfold as part of the follow-up care. Using functional MRI data acquired from post-operative Prakash patients imaged at various points throughout their recovery, Sinha and the rest of the Project Prakash team observe how a child’s brain incorporates new information into existing structural and functional organization. These studies then inform Sinha’s MIT lab work on computational models of visual learning.

Prakash patients range from children as young as 6 to young adults in their mid-20s; if a young person presents to the Project Prakash team with a treatable form of blindness, they won’t be turned away. A few days after surgery, children begin visual acuity tests using the standard eye charts anyone with glasses would be familiar with. While no child winds up with perfect vision, patients gain significant functional vision, and Project Prakash provides glasses to correct their sight further.

After securing an initial round of funding from the National Institutes of Health, Sinha officially launched Project Prakash with three of his students in 2004. Since then, Sinha has added 20 team members and together they have provided 500 surgeries and 43,000 ophthalmic screenings to children in need.

A ringside seat

Since Sinha established Project Prakash, the humanitarian work has become inextricably linked to the work in his research laboratory at MIT. Many of his lab members are also involved with Project Prakash and pursue a number of research questions related to data gathered from its patients. Typically, the research team travels to India twice a year to volunteer with the humanitarian efforts of the program, interact with the patients, and of course, to gather data.

Conventional wisdom in childhood blindness suggests that older children should not see significant gains post-surgery, since their brain’s visual machinery should be set. However, in one of his first key scientific findings enabled by Project Prakash, Sinha found that even young adults can make significant gains in visual function after surgical intervention.

Members of the Sinha laboratory at MIT have also found new avenues of research probing the development of different visual skills in the Prakash children. For example, Sruti Raja, a research associate in the Sinha laboratory, is working on a project that looks at sensitivity to visual motion before and after surgery. Another ongoing study, led by Sharon Gilad-Gutnick, a staff research scientist in Sinha’s lab and Project Prakash team member, looks at how patients learn to translate what they see into drawings.

“How are they able to recognize and then copy or draw from memory basic shapes?” says Gilad-Gutnick, who has worked with Sinha since she was an undergraduate. “What does that tell us about their internal representation of these shapes, and of objects in general? We are looking at that as a function of time after sight onset.”

Another study, published in the Proceedings of the National Academy of Sciences on Oct. 30, details how some Prakash patients struggle with the task of recognizing faces. Newborn babies have notoriously bad eyesight, with an average visual acuity of 20/600. According to the American Optometric Association, good visual acuity refers to the ability to see sharply and clearly. Normal visual acuity is referred to as 20/20 vision, which means that you can see patterns as clearly at 20 feet as an average person at the same distance. In this paper, the researchers hypothesize that poor eyesight has an important function in infant visual development, acting as a visual low-pass filter. The filter induces the brain to develop visual processing strategies that emphasize the gestalt rather than local details, or as Sinha describes it, the ‘forest from the trees.”

The Project Prakash patients miss out on this benefit of poor initial acuity, which leads to difficulty organizing and recognizing the spatial relationships of distinct faces. In the paper, Sinha and the team refer to this as the high-initial acuity (HIA) hypothesis. To test this hypothesis, Sinha’s research team used a deep learning algorithm designed to mimic the many layers of the human visual system. They fed the algorithm series of images simulating different visual learning scenarios, from only blurred images to only high resolution images to a mix of the two. The series that led to the most robust recognition performance began with blurred images and progressively increased in resolution — echoing the progression in normal human development and consistent with the HIA hypothesis.

These results have significant clinical implications. Post-operative outcomes for cases of congenital cataracts can potentially be improved by blurring visual stimuli to mimic the acuity of a newborn. By gradually increasing the resolution of visual stimuli, the regimen may provide the Prakash children’s brains the inducement to encode larger scale structures in images and improve subsequent recognition performance.

While the Project Prakash children are a unique subset of individuals, lessons learned from them can be applied to brain development in general. The work reported in the PNAS paper illustrates how studies of newly sighted children can inform our thinking of normal visual development, and also guide the creation of more powerful computational strategies for visual recognition.

“We are essentially providing a possible answer to why normal visual development unfolds in the way that it does,” Sinha says. “It's not just a limitation imposed upon us by immaturity of the retina, but it might actually have adaptive value.”

Looking beyond vision, this idea could potentially provide insight into auditory development as well; the muffling of sound by the amniotic sac may have adaptive significance akin to the blurring of images in early development.

Looking ahead

While Sinha appreciates the opportunity to tackle these scientific questions, he doesn’t lose sight of the transformative impact Project Prakash has on real lives. In the days following Poonam’s surgery, Project Prakash staff watched her blossom as she healed. She created artwork, she made up dance moves with her caregiving team, and she even took it upon herself to lead another blind patient receiving care at the hospital around the hallways. In addition to this newfound sense of independence and self-confidence, Poonam’s follow-up exams showed a marked improvement in visual abilities.

Poonam’s outcome was not an outlier for Project Prakash participants. Even though the speed of healing and level of visual acuity varies from patient to patient, most report significant improvement both to their vision and their quality of life. Sinha and his team surveyed a group of 60 patients and their families to gauge how their experience participating in Project Prakash improved their sense of independence, their ability to perform in school, and their relationships with friends, family and their communities.

“Across all of these dimensions, they reported big gains, and both patients and their families are uniformly ecstatic about the outcomes of the treatment, says Sinha. “When you take a step back to look at the whole picture, we have made a relatively small contribution by providing this routine surgery. But the consequence of that surgery is so profound for the child and for their family that the families think of us as more. It’s incredibly rewarding.”

But Prakash patients and their families are not the only ones who have been impacted by this work.

“In general, scientific research can be very frustrating, ambiguous and at times, difficult. But when you get to work so closely with children in need and you get to have that real-world impact in addition to pursuing these interesting questions, the motivation is huge,” says Gilad-Gutnik. “As scientists, I think we need to find more ways to work at this intersection of basic science and humanitarian need, and I think that neuroscience and the study of behavior presents a lot of unique opportunities to do that.”

Looking ahead, Project Prakash aims to improve patient outcomes even further through their newest initiative: a year-long residential educational program where patients who have fallen behind in school due to their visual impairment can receive specialized instruction, bringing them to an age-appropriate grade level before integrating them in local schools. They also intend to track how learning and education affects brain structure.

For Sinha, though his experience with Project Prakash has led to many immeasurably meaningful moments, meeting the Prakash patient named Poonam and following her success through the program hit especially close to home. His older sister was a doctor before her untimely death at age 25, and she was one of Sinha’s main sources of inspiration when he established Project Prakash. Her name was also Poonam.

“We are all shaped by the people we meet, and especially by the ones we admire,” says Sinha. “Seeing my sister’s devotion as a doctor to helping those in need, even at great cost to her own health, affected me greatly. Project Prakash is a small tribute to her memory, and the future of many more Poonams.”

de MIT News http://bit.ly/2DDbyEh

Champion figure skater thrives at MIT

“Representing the Skating Club of Boston, please welcome now, Kevin Shum!” says the announcer at the 2018 U.S. Collegiate National Championship in Adrian, Michigan. Shum, wearing a sparkly dark-blue-and-black jumpsuit, begins his free skate performance to “The Sound of Silence” by Disturbed.

According to Shum, an MIT senior majoring in computer science with a concentration in theater arts, life is about discovering and pursuing passions. Shum is a two-time collegiate champion figure skater and has traveled across the nation and around the world with Team USA’s world junior team. Despite competitive figure skating’s intense environment, the freedom of skating without restrictions or boundaries has sustained Shum’s passion for the sport and has even contributed to his success at MIT.

Shum grew up in the San Francisco Bay Area and first got involved with figure skating at age 6. What started as a hobby quickly evolved into a passion. At age 10, Shum performed in his first skating competition. “I love the big arenas. I love the bright lights. I love the big audience,” Shum says. “It’s a pretty surreal feeling.”

Shum is also passionate about his work at MIT. Here, Shum has participated in an Undergraduate Research Opportunities Program (UROP) through the MIT Media Lab. In addition, Shum is an active student blogger for the Office of Admissions, where he writes freely about his MIT experience, internships, projects, travels, and skating. He especially enjoyed writing posts about the time he spent in Germany while participating in the MIT International Science and Technology Initiatives Global Teaching Labs (MISTI GTL) during Independent Activities Period (IAP) in 2018. While there, Shum taught computer science to high school students, and shortly thereafter traveled to Zurich, Switzerland, for the inaugural exchange program between ETH Zurich and the MIT Department of Electrical Engineering and Computer Science.

Recently, Shum worked on a project for the class 6.810 (Engineering Interactive Technologies). He and a partner applied concepts related to adaptive sports to the process of learning to ride a skateboard. Shum’s skateboard detects a rider’s skill level by monitoring how often the board wobbles, and it shrinks or expands in length to fit the rider’s abilities.

Each semester, Shum trains at the Cronin Skating Rink in Revere, Massachusetts, up to six days a week for two hours a day. Training includes both on- and off-ice preparation: practicing jumps and footwork on the ice, body conditioning and strengthening at the gym, physical therapy, and a lot of stretching. This Independent Activities Period, Shum is competing in the 2019 U.S. Figure Skating Championships in Detroit, Michigan, skating to Frank Sinatra’s rendition of “Luck Be a Lady” from the musical “Guys and Dolls.”

His extraordinary commitment to skating has taught Shum to manage and prioritize his time working in the classroom, doing research, and completing projects; applying learning from one passion to another. “If there’s something I want to accomplish, I know I just need to put in the work, put in those hours. Work smart, work hard, and get it done,” he says. “I know I only have 24 hours in a day. It has really forced me to really be intentional about what I spend my time on.”

de MIT News http://bit.ly/2S8zZSb

Tsai earns Hans Wigzell's Prize in Medicine

The Hans Wigzell Research Foundation announced Jan. 23 that neuroscientist Li-Huei Tsai, Picower Professor and director of The Picower Institute for Learning and Memory at MIT, is the winner of Hans Wigzell's Prize in Medicine for 2018. Tsai will travel to Stockholm to receive the prize and deliver a lecture on her research Feb. 14.

“The prize is given to professor Li-Huei Tsai for her innovative research in trying to understand the etiology and possible treatment of Alzheimer's disease,” the Foundation stated in the announcement of the $100,000 prize. “Professor Tsai has in her research made a series of impressive findings with regard to this disease.”

In decades of research with collaborators, postdocs, and students, Tsai has led several fundamental and translatable discoveries about biological mechanisms underlying neurodegeneration including specific aberrations in epigenetic gene regulation, enzyme pathways, and repair of DNA damage. Tsai and collaborators have also uncovered substantial evidence that impaired neuronal synchrony may underlie Alzheimer’s progression, an insight that has allowed her to demonstrate a non-invasive treatment approach using light and sound stimuli to drive neural oscillations, engage the brain’s immune system, reduce pathology, and improve functionality in multiple mouse models. Testing of the technique has recently begun in humans.

“This treatment has resulted in dramatic improvements of the diseased animals both with regard to pathology and performance,” the foundation noted. “Her research has rapidly resulted in the start of advanced, clinical trials in Alzheimer’s patients.”

Tsai, who also directs the Aging Brain Initiative at MIT, said she was honored to earn the Wigzell Foundation’s recognition. Hans Wigzell is a former president of the Karolinska Institute and chair of the Nobel Prize Committee of the Institute.

“I am deeply grateful to Professor Wizgell and the foundation for this award,” Tsai says. “The prize provides my team with great inspiration and resources to continue our work to understand the biology of neurodegeneration and to translate our findings to effective treatments for Alzheimer’s and other diseases.”

de MIT News http://bit.ly/2DyeeD7

Kevin Woller: From student to mentor

A year after receiving his ScD from MIT, research scientist Kevin Woller is looking back. Years of study at Tarleton State University in Stephenville, Texas, followed by graduate research at MIT’s Plasma Science and Fusion Center (PSFC), have indebted him to a long list of teachers and mentors. His appreciation may stem from his own experience advising his young colleagues, as early in his career he is already carving out a reputation as someone who can skillfully guide undergraduate researchers and visiting students to do their best work.

Woller credits physics professor Daniel Marble at Tarleton with setting him on the path to fusion research. His professor was studying gun barrel alloy composition in an ion accelerator lab that had been established for materials analysis.

“My project was to set up a target chamber in his lab that could be accessed remotely by high school students,” Woller says. “[Marble] is a big advocate for education; he wanted schools to have access to university level research facilities.”

Also tasked with keeping the accelerator conditioned to take high voltages, Woller began working with plasmas, generating and using them as a source for the ion beam. He became even more familiar with plasmas when the ion source broke and he had to repair it.

“Once you start learning about plasma it points you towards fusion,” Woller says. “Before then I was mostly interested in renewables — solar, wind, geothermal — all that kind of stuff. When I discovered there was fusion, that’s what led me to MIT.”

When he arrived for graduate studies at MIT, Woller was primed to do research in the Cambridge Laboratory for Accelerator Study of Surfaces (CLASS) under the supervision of Department of Nuclear Science and Engineering Professor Dennis Whyte, now PSFC director. As a user facility that housed an ion accelerator, the laboratory was available to interested researchers, inside or outside MIT. With his working knowledge of ion accelerators, Woller found himself beginning to mentor others with the help of his colleagues.

“It was just me and a few other grad students,” he says. “There were [undergraduate researchers] working in the lab then, as there are today, and we would help them understand lab practices, how the accelerator worked, where the pitfalls could be.”

As a research scientist, Woller has become a reliable starting point for students and staff new to the PSFC. In the summer of 2017 he offered a project to visiting student Joran Savenije from Eindhoven University of Technology in the Netherlands, focusing on lithium coating preparation and analysis, as well as deuterium plasma generation. In 2018, working with the MIT Summer Research Program (MSRP), he oversaw visiting student Christopher Reis’ work using ion beams to evaluate materials exposed in the Experimental Advanced Superconducting Tokamak in China.

“At first, they are like shadows,” Woller explains. “It's useful to peek over someone’s shoulder. The accelerator is a large machine. It’s not easy to put someone in front of it and say ‘OK, operate it!’ I always try to explain the physical concept involved because I know that helped me learn the process." 

Once he is confident his mentees understand how the accelerator works, he puts them in control of it and observes.

“I believe physically controlling the machines or tools is very helpful in the learning process,” he says. “I take the same approach with my children. I encourage them to use tools to fix toys that break.”

Both visiting students were moved to write thank you notes to Dennis Whyte for providing the opportunity, and for pairing them with Woller. Describing his time at the PSFC as “one of the greatest opportunities I have ever received,” Reis wrote, “I am especially thankful for Kevin’s delegation as my supervisor. I learned something new from him every single day we worked together these past two months.”

Joran Savenije described Woller’s way of working and careful explanations as “probably the most valuable thing I took home with me.”

Woller, who continues to oversee students in the MIT Undergraduate Research Opportunities Program, visiting students, and postdocs, is full of praise for the colleagues he mentors, describing their work as “high-impact and technically challenging.” He credits Whyte for not only teaching him how to run the laboratory, but for modeling how to pass that knowledge on to others.

But his primary mentors will always be his parents.

“Dad was into tinkering. If something broke he wanted to fix it,” he says. “And I like to work in a lab with nuts and bolts and wrenches. Mom provided more of a way of thinking about things and going about doing things.”

Woller has been thinking about his undergraduate education, and the teachers who were key to his progress. With the rigors of writing a thesis behind him, he feels he may have time to write to his undergraduate professors and let them know where he is, and how their guidance helped. Marble has stayed in touch.

“He was the first to show me the ion accelerator,” says Woller. “He showed me the way to turn the knobs. Now I can tell him that I know there is a lot more to it than knobs and switches.”

de MIT News http://bit.ly/2RORY0Q

Filling the gaps in a patient’s medical data

MIT researchers have developed a model that can assimilate multiple types of a patient’s health data to help doctors make decisions with incomplete information.

The field of “predictive analytics” holds promise for many health care applications. Machine learning models can be trained to look for patterns in patient data to predict a patient’s risk for disease or dying in the ICU, to aid in sepsis care, or to design safer chemotherapy regimens.  

The process involves predicting variables of interest, such as disease risk, from known variables, such as symptoms, biometric data, lab tests, and body scans. However, that patient data can come from several different sources and is often incomplete. For example, it might include partial information from health surveys about physical and mental well-being, mixed with highly complex data comprising measurements of heart or brain function.

Using machine learning to analyze all available data could help doctors better diagnose and treat patients. But most models can’t handle the highly complex data. Others fail to capture the full scope of the relationships between different health variables, such as how breathing patterns help predict sleeping hours or pain levels.

In a paper being presented at the AAAI Conference on Artificial Intelligence next week, MIT researchers describe a single neural network that takes as input both simple and highly complex data. Using the known variables, the network can then fill in all the missing variables. Given data from, say, a patient’s electrocardiography (ECG) signal, which measures heart function, and self-reported fatigue level, the model can predict a patient’s pain level, which the patient might not remember or report correctly.

Tested on a real sleep study dataset — which contained health surveys, and ECG and other complex signals — the network achieved 70 to 80 percent accuracy in predicting any one of eight missing variables, based on the seven other known variables.

The network works by stitching together various submodels, each tailored to describe a specific relation among variables. The submodels share data as they make predictions, and ultimately output a predicted target variable. “We have a network of models that communicate with each other to predict what we don’t know, using the information we do know from these different types of data,” says lead author Hao Wang, a postdoc at the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL). “If you have, say, eight different types of data, and I have full information on a patient from seven, the communication between the models will help us fill in the missing gaps in the eighth type of data from the other seven types.”

Joining Wang on the paper are Chengzhi Mao, an undergraduate student at Tsinghua University; CSAIL PhD students Hao He and Mingmin Zhao; Dina Katabi, the Andrew and Erna Viterbi Professor of Electrical Engineering and Computer Science at MIT and director of the MIT Center for Wireless Networks and Mobile Computing; and Tommi S. Jaakkola, the Thomas Siebel Professor in the Department of Electrical Engineering and Computer Science and the Institute for Data, Systems, and Society.

Bi-directional predictions

Using traditional machine-learning models to analyze the number of variables the researchers’ network can handle is practically infeasible, because the number of models scales exponentially with the number of variables.

“We asked, ‘Is it possible to design a single model that can use all these groups of data, despite the fact that in each group we have different information?’” Wang says.

The key innovation was breaking the network into individual submodels each tailored to fit a different type of input data. A neural network is an interconnected network of nodes that work together to process complex data. One node does relatively simple computations before sending the output to the next node. In networks with submodels, however, each node can function as a separate network that can handle more complex computations. Submodels can be much more efficient, depending on the application.

In their work, the researchers created one probabilistic submodel for each variable output. They also developed a technique to let the submodels communicate with one another while making predictions, called Bi-directional Inference Networks (BIN). This technique leverages a neural network training technique known as backpropagation. In training, backpropagation sends computing errors back through the nodes to update the network’s parameter values. But this technique is never used in testing, especially when there are complex conditional dependencies involved. Instead, in traditional testing, inputted data get processed from node to node in one direction, until a final node at the end of the sequence outputs a prediction.

The researchers programmed their network to use both the traditional method and backpropagation during testing. In this context, backpropagation is basically taking a variable output, then predicting an input from that output, and sending the input value backward to a previous node. This creates a network where all submodels are working together and co-dependent on one another, to output a target probability.

Filling in the blanks

The researchers trained their network on the real-world Sleep Heart Health Study 2 (SHHS2) dataset. The data include electroencephalography (EEG) readings, which measure brain function; ECG; and breathing pattern signals. It also includes information from a health survey to measure eight health variables — including emotional well-being, social functioning, and energy/fatigue — ranked on a scale of 0 to 100.

In training, the network learns patterns for how each variable may affect another. For instance, if someone holds their breath for long durations, they may be tense, which can indicate physical pain. In testing, the network is able to analyze the relationships to predict any of the eight variables, based on any of the other information, with 70 to 80 percent accuracy.

The network could help quantify sometimes-ambiguous health variables for patients and doctors, such as pain and fatigue levels. When patients sleep after surgery, for instance, they may wake up in the middle of the night in pain, but may not remember an appropriate pain level the next day.

Next, the researchers hope to implement the network as a software component for a device they built, called the EQ Radio, which can track someone’s breathing and heart rate using only wireless signals. Currently, the device analyzes that information to infer if someone is happy, angry, or sad. With the network, the device could potentially make continuously updated predictions about a patient’s health, passively, given only partial information. “This could be so helpful in assisted-living facilities, where doctors can monitor both emotional and physical dimensions of a patient’s health all day, every day,” Wang says.

de MIT News http://bit.ly/2HA0nAq