�Ӱ���Ƶ

ADD-H₂: AI-Driven Discovery of High-Entropy H₂ Generators

Team: Co-PI Corey Oses (WSE), Co-PI Avi Bregman (APL), Kenneth Kane (APL), Karun Kumar Rao (APL), Victor Leon (APL), Leslie Hamilton (APL), Morgan Trexler (APL), A. Shoji Hall (WSE)

Hydrogen gas (H₂), as a zero-emissions source of energy, has tremendous potential to power a post-fossil-fuel era. It could fuel the vast and ever-increasing universe of electronic technology and provide clean energy to ramp up agricultural production and feed the world’s growing population.

However, no one has yet cut the Gordian knot of clean and cheap hydrogen production. Thermochemical water splitting — the application of heat to separate hydrogen from oxygen in water molecules — is, in principle, the ideal solution. However, this method currently needs temperatures above 3,600 degrees Fahrenheit (2,000 degrees Celsius) and expensive catalysts.

ADD-H₂, led by APL’s Avi Bregman and WSE’s Corey Oses, seeks to leverage artificial intelligence (AI) to discover high-entropy materials that can serve as catalysts for thermochemical water splitting. This relatively new type of material can remain stable through many chemical reactions and can repeatedly extract hydrogen from water with heat. This means a small amount of the material could be used to produce vast quantities of hydrogen.

The goal of ADD-H₂ is to design and demonstrate a new type of reactor for this purpose, using materials that can catalyze the reaction at 1,300 degrees Fahrenheit (700 degrees Celsius) or less — low enough that waste heat from various industrial processes could be used to drive the reaction and harvest critical hydrogen outputs.

High-entropy materials are extremely complex molecules comprising five to seven elements. This complexity makes discovering the right material among millions of possible candidates impossible for humans, but not for AI.

Oses is a world-class materials scientist who has devoted his career to searching for new high-entropy material candidates, and in the process, he has developed powerful, AI-driven techniques for doing so. Bregman and APL team members bring expertise in AI for materials discovery, reactor design, creating controlled sources of high-temperature steam, and other areas critical to this project’s success.

“We envision a future where hydrogen production can be done not only on an industrial scale, but on a localized and distributed scale — where everyone has a hydrogen reactor in their house,” Bregman said.

“A reactor operating at 700 degrees Celsius and with a small, efficient and safe design would be a giant leap toward that vision, providing cheap and plentiful energy wherever it’s needed,” added Oses.

Cerebro: Enabling the Next Step of Human Evolution

Team: Co-PI Amy Foster (WSE), Co-PI Nicholas G. Pavlopoulos (APL), Jacob Khurgin (WSE), Mark Foster (WSE), Jeremiah Wathen (APL), Griffin Milsap (APL), Konstantinos Gerasopoulos (APL), Elisabeth Marsh (SOM)

Advances in brain imaging have made it possible to predict strokes and brain tumors and to identify conditions such as Alzheimer’s disease and dementia earlier than conventional diagnostic methods. Neuroimaging is also proving to be a valuable research tool in studying autism spectrum disorders and has enabled therapeutic and human enhancement concepts, such as using the power of thought to operate prostheses and remote machinery.

Unfortunately, the actual benefits of neuroimaging are currently limited because of the lack of noninvasive methods for monitoring brain activity that are also portable, wearable, sensitive, high resolution and — critically — inexpensive. Creating such a method is the mission of the Cerebro team.

With their first round of SURPASS funding, the team made critical strides toward its goal by demonstrating that it’s possible to convert weak magnetic fields like those generated by brain activity into optical signals. This in turn makes it possible to leverage existing technology to read those signals.

“In the past year, we’ve developed novel materials that change how light passes through them when you put them in a magnetic field and that can be easily integrated into chips,” said Nicholas Pavlopoulos, Cerebro co-PI at APL. “We’ve also built surrogate systems to test these materials by using photonic integrated circuits.”

In this round of SURPASS, the team hopes to demonstrate that its materials can measure biologically produced magnetic fields and, if they achieve that in their intended time frame and sensor format, compare their sensors to state-of-the-art brain monitoring equipment.

Ultimately, the Cerebro team is motivated in this ambitious effort by the lifesaving promise this technology holds.

“It’s 2024, and early diagnostics have completely changed the field for so many diseases, including some very scary cancers,” Pavlopoulos said. “If we can figure out how to get the necessary technology into the form of a microchip, we envision brain scans can become a routine part of your yearly physical, something you can get in an urgent care clinic.”

Organoid Intelligence: Synthetic Biological AI

Team: Co-PI Erik C. Johnson (APL), Co-PI Thomas Hartung (WSE and BSPH), Lomax Boyd (BSPH), Brian Caffo (BSPH), David Gracias (WSE), Erin Hahn (APL), Tim Harris (WSE), Jeffrey Kahn (BSPH), Bart Paulhamus (APL), Lena Smirnova (WSE/BSPH), Francesco Tenore (APL), Brock Wester (APL), Mike Wolmetz (APL)

Our understanding of biological intelligence is still in its infancy, and the same might be said of AI. Integrating both into a single system holds the promise to dramatically improve our understanding of each and potentially produce a quantum leap in computing as we know it.

The Organoid Intelligence (OI) team is pioneering the use of brain organoids — 3D cell culture models that are grown from adult human stem cells and mimic certain structural and functional properties of specific regions of the brain. Doing this will allow the team to gain insight into the emergence of intelligence in biological neural networks.

The OI team has made considerable progress with its first round of SURPASS funding. The team has created a prototype system for OI experiments, new methods for electrically stimulating and recording activity in organoids, innovative microfluidic devices to maintain the health and stability of the cell cultures, and novel machine learning and AI tasks to test these systems.

All of this has laid the groundwork for an even more impressive showing in the second round.

“The work we did in the first round of SURPASS has really set us up to advance discovery by growing larger organoids and developing more testing modalities and tasks to gain deeper insight into how these systems work,” said , co-PI at WSE.

“In the new cycle, I would like to increase the complexity of the model by introducing more regions of the brain involved in learning and reward processing,” said , a member of the WSE OI team.

“We’ve done some experiments to help us understand the mechanisms of long-term learning in these brain organoids — how the synapses change their connectivity strength in response to performing a task, which is one of the basic mechanisms of learning,” said Erik C. Johnson, the OI team’s co-PI at APL. “We’re also developing new interfaces and microfluidic systems that can work with the complex 3D cell cultures we’re creating and overcome the limitations of commercially available 2D systems.”

Beyond the scope of the second stage of SURPASS, he added, the team will also formalize protocols for ensuring ethical oversight of this work.

Photoacoustic Retinal Prosthesis: A New Paradigm for Prosthetic Vision

Team: Co-PI Emad Boctor (WSE), Co-PI Seth Billings (APL), Maomao Chen (SOM), Bree Christie (APL), George Coles (APL), Luke Currano (APL), Peter Gehlbach (SOM, WSE), Jeeun Kang (WSE), Jin Kang (WSE), Casey Keuthan (SOM), Dominique Meyer (WSE), Van Phuc Nguyen (SOM), Alexandra Patterson (WSE), Yannis Paulus (SOM), Erika Rashka (APL), Wayne Rodgers (WSE), Jacalynn Sharp (APL), David Shrekenhamer (APL), Hyunwoo Song (WSE), James Spicer (WSE), Christopher Stiles (APL), Francesco Tenore (APL), Chad Weiler (APL), Ji Yi (WSE), Don Zack (SOM), Mi Zheng (SOM)

Around 3 million Americans today suffer from irreversible degenerative vision disorders — a figure that is set to double by 2050. The team behind Photoacoustic Retinal Prosthesis hopes to provide a viable treatment option for individuals suffering from blindness or severe vision loss due to incurable disorders of the outer retina.

To do that, they’re harnessing the power of the photoacoustic effect, a physical phenomenon used in a variety of scientific applications whereby a material absorbs light, heats, and rapidly expands, generating pressure waves that then propagate as ultrasound. The team’s patented approach — known as photoacoustic retinal stimulation, or PARS — involves using a focused nanosecond pulsed laser to irradiate an implant material to produce ultrasound waves. These then stimulate activity within the residual neurons of the inner retina that are responsible for vision. In contrast to existing vision-restoration devices, which require implanting an invasive, active device onto the retina, this approach uses a passive implant that would be stimulated by a wearable device in an eyeglasses form factor.

In its first phase of research under last year’s SURPASS program, the Photoacoustic Retinal Prosthesis team developed a prototype test device and completed proof-of-concept experiments demonstrating its ability to successfully stimulate activity in retinal neurons.

As the researchers continue their work this year, they’re focused on maturing their technology to improve performance and expand the domain of retinal disorders that could be alleviated with the device, continuing biological studies to further characterize the safety and efficacy of the prototype and optimizing their approach through further experiments.

“No one person or small team can have all of the knowledge and experience that’s required to accomplish something like this,” said Seth Billings, APL co-PI of the Photoacoustic Retinal Prosthesis team. “It takes all of the different divisions of Johns Hopkins coming together.”