Computers Math

A framework based on advanced AI techniques can solve complex, computationally intensive problems faster and in a more more scalable way than state-of-the-art methods, according to a study led by engineers at the University of California San Diego.
In the paper, which was published May 30 in Nature Machine Intelligence, researchers present HypOp, a framework that uses unsupervised learning and hypergraph neural networks. The framework is able to solve combinatorial optimization problems significantly faster than existing methods. HypOp is also able to solve certain combinatorial problems that can’t be solved as effectively by prior methods.
“In this paper, we tackle the difficult task of addressing combinatorial optimization problems that are paramount in many fields of science and engineering,” said Nasimeh Heydaribeni, the paper’s corresponding author and a postdoctoral scholar in the UC San Diego Department of Electrical and Computer Engineering. She is part of the research group of Professor Farinaz Koushanfar, who co-directs the Center for Machine-Intelligence, Computing and Security at the UC San Diego Jacobs School of Engineering. Professor Tina Eliassi-Rad from Northeastern University also collaborated with the UC San Diego team on this project.
One example of a relatively simple combinatorial problem is figuring out how many and what kind of goods to stock at specific warehouses in order to consume the least amount of gas when delivering these goods.
HypOp can be applied to a broad spectrum of challenging real-world problems, with applications in drug discovery, chip design, logic verification, logistics and more. These are all combinatorial problems with a wide range of variables and constraints that make them extremely difficult to solve. That is because in these problems, the size of the underlying search space for finding potential solutions increases exponentially rather than in a linear fashion with respect to the problem size.
HypOp can solve these complex problems in a more scalable manner by using a new distributed algorithm that allows multiple computation units on the hypergraph to solve the problem together, in parallel, more efficiently.
HypOp introduces new problem embedding leveraging hypergraph neural networks, which have higher order connections than traditional graph neural networks, to better model the problem constraints and solve them more proficiently. HypOp also can transfer learning from one problem to help solve other, seemingly different problems more effectively. HypOp includes an additional fine-tuning step, which leads to finding more accurate solutions than the prior existing methods.
This research was funded in part by the Department of Defense and Army Research Office funded MURI AutoCombat project and the NSF-funded TILOS AI Institute. More

Imagine a portable 3D printer you could hold in the palm of your hand. The tiny device could enable a user to rapidly create customized, low-cost objects on the go, like a fastener to repair a wobbly bicycle wheel or a component for a critical medical operation.
Researchers from MIT and the University of Texas at Austin took a major step toward making this idea a reality by demonstrating the first chip-based 3D printer. Their proof-of-concept device consists of a single, millimeter-scale photonic chip that emits reconfigurable beams of light into a well of resin that cures into a solid shape when light strikes it.
The prototype chip has no moving parts, instead relying on an array of tiny optical antennas to steer a beam of light. The beam projects up into a liquid resin that has been designed to rapidly cure when exposed to the beam’s wavelength of visible light.
By combining silicon photonics and photochemistry, the interdisciplinary research team was able to demonstrate a chip that can steer light beams to 3D print arbitrary two-dimensional patterns, including the letters M-I-T. Shapes can be fully formed in a matter of seconds.
In the long run, they envision a system where a photonic chip sits at the bottom of a well of resin and emits a 3D hologram of visible light, rapidly curing an entire object in a single step.
This type of portable 3D printer could have many applications, such as enabling clinicians to create tailor-made medical device components or allowing engineers to make rapid prototypes at a job site.
“This system is completely rethinking what a 3D printer is. It is no longer a big box sitting on a bench in a lab creating objects, but something that is handheld and portable. It is exciting to think about the new applications that could come out of this and how the field of 3D printing could change,” says senior author Jelena Notaros, the Robert J. Shillman Career Development Professor in Electrical Engineering and Computer Science (EECS), and a member of the Research Laboratory of Electronics.

Joining Notaros on the paper are Sabrina Corsetti, lead author and EECS graduate student; Milica Notaros PhD ’23; Tal Sneh, an EECS graduate student; Alex Safford, a recent graduate of the University of Texas at Austin; and Zak Page, an assistant professor in the Department of Chemical Engineering at UT Austin. The research appears today in Nature Light Science and Applications.
Printing with a chip
Experts in silicon photonics, the Notaros group previously developed integrated optical-phased-array systems that steer beams of light using a series of microscale antennas fabricated on a chip using semiconductor manufacturing processes. By speeding up or delaying the optical signal on either side of the antenna array, they can move the beam of emitted light in a certain direction.
Such systems are key for lidar sensors, which map their surroundings by emitting infrared light beams that bounce off nearby objects. Recently, the group has focused on systems that emit and steer visible light for augmented-reality applications.
They wondered if such a device could be used for a chip-based 3D printer.
At about the same time they started brainstorming, the Page Group at UT Austin demonstrated specialized resins that can be rapidly cured using wavelengths of visible light for the first time. This was the missing piece that pushed the chip-based 3D printer into reality.

“With photocurable resins, it is very hard to get them to cure all the way up at infrared wavelengths, which is where integrated optical-phased-array systems were operating in the past for lidar,” Corsetti says. “Here, we are meeting in the middle between standard photochemistry and silicon photonics by using visible-light-curable resins and visible-light-emitting chips to create this chip-based 3D printer. You have this merging of two technologies into a completely new idea.”
Their prototype consists of a single photonic chip containing an array of 160-nanometer-thick optical antennas. (A sheet of paper is about 100,000 nanometers thick.) The entire chip fits onto a U.S. quarter.
When powered by an off-chip laser, the antennas emit a steerable beam of visible light into the well of photocurable resin. The chip sits below a clear slide, like those used in microscopes, which contains a shallow indentation that holds the resin. The researchers use electrical signals to nonmechanically steer the light beam, causing the resin to solidify wherever the beam strikes it.
A collaborative approach
But effectively modulating visible-wavelength light, which involves modifying its amplitude and phase, is especially tricky. One common method requires heating the chip, but this is inefficient and takes a large amount of physical space.
Instead, the researchers used liquid crystal to fashion compact modulators they integrate onto the chip. The material’s unique optical properties enable the modulators to be extremely efficient and only about 20 microns in length.
A single waveguide on the chip holds the light from the off-chip laser. Running along the waveguide are tiny taps which tap off a little bit of light to each of the antennas.
The researchers actively tune the modulators using an electric field, which reorients the liquid crystal molecules in a certain direction. In this way, they can precisely control the amplitude and phase of light being routed to the antennas.
But forming and steering the beam is only half the battle. Interfacing with a novel photocurable resin was a completely different challenge.
The Page Group at UT Austin worked closely with the Notaros Group at MIT, carefully adjusting the chemical combinations and concentrations to zero-in on a formula that provided a long shelf-life and rapid curing.
In the end, the group used their prototype to 3D print arbitrary two-dimensional shapes within seconds.
Building off this prototype, they want to move toward developing a system like the one they originally conceptualized — a chip that emits a hologram of visible light in a resin well to enable volumetric 3D printing in only one step.
“To be able to do that, we need a completely new silicon-photonics chip design. We already laid out a lot of what that final system would look like in this paper. And, now, we are excited to continue working towards this ultimate demonstration,” Jelena Notaros says.
This work was funded, in part, by the U.S. National Science Foundation, the U.S. Defense Advanced Research Projects Agency, the Robert A. Welch Foundation, the MIT Rolf G. Locher Endowed Fellowship, and the MIT Frederick and Barbara Cronin Fellowship. More

University of North Carolina at Chapel Hill scientists have created innovative soft robots equipped with electronic skins and artificial muscles, allowing them to sense their surroundings and adapt their movements in real-time, according to the paper, “Skin-Inspired, Sensory Robots for Electronic Implants,” in Nature Communications.
In their research, funded by the National Science Foundation and the National Institutes of Health, the robots are designed to mimic the way muscles and skin work together in animals, making them more effective and safer to use inside the body. The e-skin integrates various sensing materials, such as silver nanowires and conductive polymers within a flexible base, closely resembling the complex sensory functions of real skin.
“These soft robots can perform a variety of well-controlled movements, including bending, expanding and twisting inside biological environments,” said Lin Zhang, first author of the paper and a postdoctoral fellow in Carolina’s Department of Applied Physical Sciences. “They are designed to attach to tissues gently, reducing stress and potential damage. Inspired by natural shapes like starfish and seedpods, they can transform their structures to perform different tasks efficiently.”
These features make soft sensory robots highly adaptable and useful for enhancing medical diagnostics and treatments. They can change shape to fit organs for better sensing and treatment; are capable of continuous monitoring of internal conditions, like bladder volume and blood pressure; provide treatments, such as electrical stimulation, based on real-time data; and can be swallowed to monitor and treat conditions in the stomach.
An ingestible robot capable of residing in the stomach called a thera-gripper, can monitor pH levels and deliver drugs over an extended period, improving treatment outcomes for gastrointestinal conditions. The thera-gripper can also gently attach to a beating heart, continuously monitoring electrophysiological activity, measuring cardiac contraction and providing electrical stimulation to regulate heart rhythm.
A robotic gripper designed to wrap around a person’s bladder can measure its volume and provide electrical stimulation to treat the overactive one, enhancing patient care and treatment efficacy. A robotic cuff that twists around a blood vessel can accurately measure blood pressure in real time, offering a non-invasive and precise monitoring solution.
“Tests on mice have demonstrated the thera-gripper’s capability to perform these functions effectively, showcasing its potential as a next-generation cardiac implant,” said Zhang.
The Bai Lab collaborated on the study with UNC-Chapel Hill researchers in the Department of Biology; Department of Biomedical Engineering; Department of Chemistry; Joint Department of Biomedical Engineering and McAllister Heart Institute; North Carolina State University; and Weldon School of Biomedical Engineering at Purdue University.
The researchers’ success in live animal models suggests a promising future for these robots in real-world medical applications, potentially revolutionizing the treatment of chronic diseases and improving patient outcomes.
“This innovative approach to robot design not only broadens the scope of medical devices but also highlights the potential for future advancements in the synergistic interaction between soft implantable robots and biological tissues,” said Wubin Bai, principal investigator of the research and Carolina assistant professor. “We’re aiming for long-term biocompatibility and stability in dynamic physiological environments.” More

Boys are good at math, girls not so much? A study from the University of Zurich has analyzed the social mechanisms that contribute to the gender gap in math confidence. While peer comparisons seem to play a crucial role for boys, girls’ subjective evaluations are more likely to be based on objective performance.
Research has shown that in Western societies, the average secondary school girl has less confidence in her mathematical abilities than the average boy of the same age. At the same time, no significant difference has been found between girls’ and boys’ performance in mathematics. This phenomenon is often framed as girls not being confident enough in their abilities, or that boys might in fact be overconfident.
This math confidence gap has far-reaching consequences: self-perceived competence influences educational and occupational choices and young people choose university subjects and careers that they believe they are talented in. As a result, women are underrepresented in STEM (science, technology, engineering, math) subjects at university level and in high-paying STEM careers.
Peer processes provide nuanced insights into varying self-perceptions
A study from the University of Zurich (UZH) focuses on a previously neglected aspect of the math confidence gap: the role of peer relationships. “Especially in adolescence, peers are the primary social reference for individual development. Peer processes that operate through friendship networks determine a wide range of individual outcomes,” said the study’s lead author Isabel Raabe from the Department of Sociology at UZH. The study analyzed data from 8,812 individuals in 358 classrooms in a longitudinal social network analysis.
As expected, the main predictor of math confidence is individual math grades. While girls translated their grades — more or less directly — into self-assessment, boys with below-average grades nevertheless believed they were good at math.
Boys tend to be overconfident and sensitive to social processes
“In general, boys seem to be more sensitive to social processes in their self-perception — they compare themselves more with others for validation and then adjust their confidence accordingly,” Raabe explains. “When they were confronted with girls’ self-assessments in cross-gender friendships, their math confidence tended to be lower.” Peers’ self-assessment was less relevant to girls’ math confidence. Their subjective evaluation seemed to be driven more by objective performance.
Gender stereotypes did not appear to have negative social consequences for either boys or girls. “We found that confidence in mathematics is often associated with better social integration, both in same-sex and cross-sex friendships,” said Raabe. Thus, there was no evidence of harmful peer norms pressuring girls to underestimate their math skills.
The results of the study suggest that math skills are more important to boys, who adjust their self-assessment in peer processes, while math confidence does not seem to be socially relevant for girls. More

Lasers have revolutionized the world since the 60’s and are now indispensable in modern applications, from cutting-edge surgery and precise manufacturing to data transmission across optical fibers.
But as the need for laser-based applications grows, so do challenges. For example, there is a growing market for fiber lasers, which are currently used in industrial cutting, welding, and marking applications.
Fiber lasers use an optical fiber doped with rare-earth elements (erbium, ytterbium, neodymium etc) as their optical gain source (the part that produces the laser’s light). They emit high-quality beams, they have high power output, and they are efficient, low-maintenance, durable, and they are typically smaller than gas lasers. Fiber lasers are also the ‘gold standard’ for low phase noise, meaning that their beams remain stable over time.
But despite all that, there is a growing demand for miniaturizing fiber lasers on a chip-scale level. Erbium-based fiber lasers are especially interesting, as they meet all the requirements for maintaining a laser’s high coherence and stability. But miniaturizing them has been met by challenges in maintaining their performance at small scales.
Now, scientists led by Dr Yang Liu and Professor Tobias Kippenberg at EPFL have built the first ever chip-integrated erbium-doped waveguide laser that approaches the performance with fiber-based lasers, combining wide wavelength tunability with the practicality of chip-scale photonic integration. The breakthrough is published in Nature Photonics.
Building a chip-scale laser
The researchers developed their chip-scale erbium laser using a state-of-the-art fabrication process. They began by constructing a meter-long, on-chip optical cavity (a set of mirrors that provide optical feedback) based on ultralow-loss silicon nitride photonic integrated circuit.

“We were able to design the laser cavity to be meter-scale in length despite the compact chip size, thanks to the integration of these microring resonators that effectively extend the optical path without physically enlarging the device,” says Dr. Liu.
The team then implanted the circuit with high-concentration erbium ions to selectively create the active gain medium necessary for lasing. Finally, they integrated the citcuit with a III-V semiconductor pump laser to excite the erbium ions to enable them to emit light and produce the laser beam.
To refine the laser’s performance and achieve precise wavelength control, the researchers engineered an innovative intra-cavity design featuring microring-based Vernier filters, a type of optical filter that can select specific frequencies of light.
The filters allow for dynamic tuning of the laser’s wavelength over a broad range, making it versatile and usable in various applications. This design supports stable, single-mode lasing with an impressively narrow intrinsic linewidth of just 50 Hz.
It also allows for significant side mode suppression — the laser’s ability to emit light at a single, consistent frequency while minimizing the intensity of other frequencies (‘side modes’). This ensures “clean” and stable output across the light spectrum for high-precision applications.
Power, precision, stability, and low noise
The chip-scale erbium-based fiber laser features output power exceeding 10 mW and a side mode suppression ratio greater than 70 dB, outperforming many conventional systems.

It also has a very narrow linewidth, which means the light it emits is very pure and steady, which is important for coherent applications such as sensing, gyroscopes, LiDAR, and optical frequency metrology.
The microring-based Vernier filter gives the laser broad wavelength tunability across 40 nm within the C- and L-bands (ranges of wavelengths used in telecommunications), surpassing legacy fiber lasers in both tuning and low spectral spurs metrics (“spurs” are unwanted frequencies), while remaining compatible with current semiconductor manufacturing processes.
Next-generation lasers
Miniaturizing and integrating erbium fiber lasers into chip-scale devices can reduce their overall costs, making them accessible for portable and highly integrated systems across telecommunications, medical diagnostics, and consumer electronics.
It can also scale down optical technologies in various other applications, such as LiDAR, microwave photonics, optical frequency synthesis, and free-space communications.
“The application areas of such a new class of erbium-doped integrated lasers are virtually unlimited,” says Liu.
The lab spin-off,EDWATEC SA, is an Integrated Device Manufacturer that can now offer Rare-Earth Ion-Doped Photonic Integrated Circuit-based Devices including high-performance amplifiers and lasers. More

A team of Vanderbilt researchers has developed a wirelessly activated device that mimics the wavelike muscular function in the esophagus and small intestine responsible for transporting food and viscous fluids for digestion.
The soft-robotic prototype, which is driven by strong magnets controlled by a wearable external actuator, can aid patients suffering from blockages caused by tumors or those requiring stents. For example, traditional esophageal stents are metal tubes used in patients with esophageal cancer, mostly in an aging population. These patients risk food being blocked from entering the stomach, potentially causing a dangerous situation where food instead enters the lung.
Restoring the natural motion of peristalsis, the wavelike muscular transport function that takes place inside tubular human organs, “paves the way for next-generation robotic medical devices to improve the quality of life especially for the aging population,” researchers wrote in a new paper in the journal Advanced Functional Materials describing the device.
The study was led by Xiaoguang Dong, Assistant Professor of Mechanical Engineering. This work was done in collaboration with Vanderbilt University Medical Center colleague, Dr. Rishi Naik, Assistant Professor of Medicine in the Division of Gastroenterology, Hepatology and Nutrition.
The device itself consists of a soft sheet of small magnets arrayed in parallel rows that are activated in a precise undulating motion that produces the torque required to pump various solid and liquid cargoes. “Magnetically actuated soft robotic pumps that can restore peristalsis and seamlessly integrate with medical stents have not been reported before,” Dong and the researchers report in the paper.
Dong, who also holds appointments in Biomedical Engineering and Electrical and Computer Engineering, said further refinements of the device could aid in other biological processes that may have been compromised by disease. For example, he said the design could be used to help transport human eggs from the ovaries when muscular function in the fallopian tubes has been impaired. In addition, the researchers said with advanced manufacturing processes, the device could be scaled down to adapt to even narrower passageways.
Vanderbilt University School of Engineering provided funding support. Oak Ridge National Laboratory provided facility support for this research. The research team is affiliated with the Vanderbilt Institute for Surgery and Engineering (VISE). More

Researchers at University of Galway have created digital babies to better understand infants’ health in their critical first 180 days of life.
The team created 360 advanced computer models that simulate the unique metabolic processes of each baby.
The digital babies are the first sex-specific computational whole-body models representing newborn and infant metabolism with 26 organs, six cell types, and more than 80,000 metabolic reactions.
Real-life data from 10,000 newborns, including sex, birth weight and metabolite concentrations, enabled the creation and validation of the models, which can be personalised — enabling scientists to investigate an individual infant’s metabolism for precision medicine applications.
The work was conducted by a team of scientists at University of Galway’s Digital Metabolic Twin Centre and Heidelberg University, led by APC Microbiome Ireland principal investigator Professor Ines Thiele.
The team’s research aims to advance precision medicine using computational modelling. They describe the computational modelling of babies as seminal, as it enhances understanding of infant metabolism and creates opportunities to improve the diagnosis and treatment of medical conditions during the early days of a baby’s life, such as inherited metabolic diseases.
Lead author Elaine Zaunseder, Heidelberg University,said: “Babies are not just small adults — they have unique metabolic features that allow them to develop and grow up healthy. For instance, babies need more energy for regulating body temperature due to, for example, their high surface-area-to-mass ratio, but they cannot shiver in the first six months of life, so metabolic processes must ensure the infant keeps warm.

“Therefore, an essential part of this research work was to identify these metabolic processes and translate them into mathematical concepts that could be applied in the computational model. We captured metabolism in an organ-specific manner, which offers the unique opportunity to model organ-specific energy demands that are very different in infants compared to adults.
“As nutrition is the fuel for metabolism, we can use breast milk data from real newborns in our models to simulate the associated metabolism throughout the baby’s entire body, including various organs. Based on their nutrition, we simulated the development of digital babies over six months and showed that they will grow at the same rate as real-world infants.”
Professor Ines Thiele, study lead on the project, said: “New-born screening programmes are crucial for detecting metabolic diseases early on, enhancing infant survival rates and health outcomes. However, the variability observed in how these diseases manifest in babies underscores the urgent need for personalised approaches to disease management.
“Our models allow researchers to investigate the metabolism of healthy infants as well as infants suffering from inherited metabolic diseases, including those investigated in newborn screening. When simulating the metabolism of infants with a disease, the models showed we can predict known biomarkers for these diseases. Furthermore, the models accurately predicted metabolic responses to various treatment strategies, showcasing their potential in clinical settings.”
Elaine Zaunseder added: “This work is a first step towards establishing digital metabolic twins for infants, providing a detailed view of their metabolic processes. Such digital twins have the potential to revolutionise paediatric healthcare by enabling tailored disease management for each infant’s unique metabolic needs.” More

Neurons communicate electrically so to understand how they produce brain functions such as memory, neuroscientists must track how their voltage changes — sometimes subtly — on the timescale of milliseconds. In a new paper in Nature Communications, MIT researchers describe a novel image sensor with the capability to substantially increase that ability.
The invention led by Jie Zhang, a postdoctoral scholar in The Picower Institute for Learning and Memory lab of Sherman Fairchild Professor Matt Wilson, is a new take on the standard “CMOS” technology used in scientific imaging. In that standard approach, all pixels turn on and off at the same time — a configuration with an inherent trade-off in which fast sampling means capturing less light. The new chip enables each pixel’s timing to be controlled individually. That arrangement provides a “best of both worlds” in which neighboring pixels can essentially complement each other to capture all the available light without sacrificing speed.
In experiments described in the study, Zhang and Wilson’s team demonstrates how “pixelwise” programmability enabled them to improve visualization of neural voltage “spikes,” which are the signals neurons use to communicate with each other, and even the more subtle, momentary fluctuations in their voltage that constantly occur between those spiking events.
“Measuring with single-spike resolution is really important as part of our research approach,” said senior author Wilson, a Professor in MIT’s Departments of Biology and Brain and Cognitive Sciences (BCS), whose lab studies how the brain encodes and refines spatial memories both during wakeful exploration and during sleep. “Thinking about the encoding processes within the brain, single spikes and the timing of those spikes is important in understanding how the brain processes information.”
For decades Wilson has helped to drive innovations in the use of electrodes to tap into neural electrical signals in real-time, but like many researchers he has also sought visual readouts of electrical activity because they can highlight large areas of tissue and still show which exact neurons are electrically active at any given moment. Being able to identify which neurons are active can enable researchers to learn which types of neurons are participating in memory processes, providing important clues about how brain circuits work.
In recent years, neuroscientists including co-senior author Ed Boyden, Y. Eva Tan Professor of Neurotechnology in BCS and The McGovern Institute for Brain Research and a Picower Institute affiliate, have worked to meet that need by inventing “genetically encoded voltage indicators” (GEVIs), that make cells glow as their voltage changes in real-time. But as Zhang and Wilson have tried to employ GEVIs in their research, they’ve found that conventional CMOS image sensors were missing a lot of the action. If they operated too fast, they wouldn’t gather enough light. If they operated too slow, they’d miss rapid changes.
But image sensors have such fine resolution that many pixels are really looking at essentially the same place on the scale of a whole neuron, Wilson said. Recognizing that there was resolution to spare, Zhang applied his expertise in sensor design to invent an image sensor chip that would enable neighboring pixels to each have their own timing. Faster ones could capture rapid changes. Slower-working ones could gather more light. No action or photons would be missed. Zhang also cleverly engineered the required control electronics so they barely cut into the space available for light-sensitive elements on a pixels. This ensured the sensor’s high sensitivity under low light conditions, Zhang said.

Two demos
In the study the researchers demonstrated two ways in which the chip improved imaging of voltage activity of mouse hippocampus neurons cultured in a dish. They ran their sensor head-to-head against an industry standard scientific CMOS image sensor chip.
In the first set of experiments the team sought to image the fast dynamics of neural voltage. On the conventional CMOS chip, each pixel had a zippy 1.25 millisecond exposure time. On the pixel-wise sensor each pixel in neighboring groups of four stayed on for 5 milliseconds, but their start times were staggered so that each one turned on and off 1.25 seconds later than the next. In the study, the team shows that each pixel, because it was on longer, gathered more light but because each one was capturing a new view every 1.25 milliseconds, it was equivalent to simply having a fast temporal resolution. The result was a doubling of the signal-to-noise ratio for the pixelwise chip. This achieves high temporal resolution at a fraction of the sampling rate compared to conventional CMOS chips, Zhang said.
Moreover, the pixelwise chip detected neural spiking activities that the conventional sensor missed. And when the researchers compared the performance of each kind of sensor against the electrical readings made with a traditional patch clamp electrode, they found that the staggered pixelwise measurements better matched that of the patch clamp.
In the second set of experiments, the team sought to demonstrate that the pixelwise chip could capture both the fast dynamics and also the slower, more subtle “subthreshold” voltage variances neurons exhibit. To do so they varied the exposure durations of neighboring pixels in the pixelwise chip, ranging from 15.4 milliseconds down to just 1.9 milliseconds. In this way, fast pixels sampled every quick change (albeit faintly), while slower pixels integrated enough light over time to track even subtle slower fluctuations. By integrating the data from each pixel, the chip was indeed able to capture both fast spiking and slower subthreshold changes, the researchers reported.
The experiments with small clusters of neurons in a dish was only a proof-of concept, Wilson said. His lab’s ultimate goal is to conduct brain-wide, real-time measurements of activity in distinct types of neurons in animals even as they are freely moving about and learning how to navigate mazes. The development of GEVIs and of image sensors like the pixelwise chip that can successfully take advantage of what they show is crucial to making that goal feasible.

“That’s the idea of everything we want to put together: large-scale voltage imaging of genetically tagged neurons in freely behaving animals,” Wilson said.
To achieve this, Zhang added, “We are already working on the next iteration of chips with lower noise, higher pixel counts, time-resolution of multiple kHz, and small form factors for imaging in freely behaving animals.”
The research is advancing pixel by pixel.
In addition to Zhang, Wilson and Boyden the paper’s other authors are Jonathan Newman, Zeguan Wang, Yong Qian, Pedro Feliciano-Ramos, Wei Guo, Takato Honda, Zhe Sage Chen, Changyang Linghu, Ralph-Etienne Cummings, and Eric Fossum.
The Picower Institute for Learning and Memory, The JPB Foundation, the Alana Foundation, The Louis B. Thalheimer Fund for Translational Research, the National Institutes of Health, HHMI, Lisa Yang and John Doerr provided support for the research. More