�Ӱ���Ƶ

The National Institutes of Health (NIH) has  for two projects supported by the Johns Hopkins Applied Physics Laboratory (APL) that will accelerate neuroscience research, illuminate foundational principles governing the neural circuit basis of behavior, and inform new approaches to the treatment of human brain disorders.

The funding comes through NIH’s  (BRAIN CONNECTS) program, a transformative project within Phase 2 of the Brain Research Through Advancing Innovative Neurotechnologies Initiative, or the BRAIN Initiative.

“There is a long history of neuroscience informing and inspiring new approaches to the treatment of human brain disorders,” said Sarah Herman, who leads the Biological and Chemical Sciences program at APL. “We’re already starting to see these intricate cortical maps helping to fill critical gaps in our understanding of biological and artificial neural networks.”

The two projects, BRAIN CONNECTS: A Center for High-Throughput Integrative Mouse Connectomics and BRAIN CONNECTS: Rapid and Cost-Effective Connectomics with Intelligent Image Acquisition, Reconstruction and Querying, are being led by Harvard University professors  and  and supported by APL. Led by researchers Brock Wester, assistant manager of the Biological and Chemical Sciences program, and Will Gray Roncal, a principal staff member and project manager, the APL team will bring scalable cloud-computing pipelines, systems engineering and machine-learning expertise to the projects.

“Connectomics, the study of structural and functional connectivity maps in the brain, is still a relatively new field, and these NIH initiatives have been instrumental in pushing technology and our understanding of the brain forward,” said Wester. “We have worked with several of these research teams on past projects, such as on NIH  and IARPA , and we are excited to see what we’ll be able to accomplish in this next phase of work.”

The Intelligence Advanced Research Projects Activity (IARPA) Machine Intelligence from Cortical Networks (MICrONS) program, overseen by Jacob Vogelstein and David Markowitz, was launched in 2015 to develop machine-learning capabilities by modeling how the brain processes information. APL supported the program by developing novel proofreading tools and a scalable cloud-based capability for storing, accessing and processing petabytes (millions of gigabytes) of neuroanatomical data and the Brain Observatory Storage Service and Database (BossDB).

The BossDB ecosystem has since been supported by NIH (R24MH114785) to store high-resolution volumetric neuroimaging datasets generated on the BRAIN Initiative, and enable researchers to collaborate and conduct large-scale, multi-team research.

“One of the main goals of MICrONS was to scale up data collection and processing, and the performers on the program were able to extend beyond a , so about the size of a large grain of salt in the visual cortex of mice,” said Wester. “BRAIN CONNECTS will build off that collection, the associated capabilities developed, and the incredible advancements demonstrated by many research teams such as those led by Jeff Lichtman and Aravi Samuel, and hopefully we’ll be able to further scale up to map an entire mouse brain.”

Gray Roncal will support Jeff Lichtman’s project by establishing a platform to help accelerate scientific discovery, especially for researchers from diverse perspectives. This overall project will look to dramatically accelerate each step of the pipeline, from brain image data collection with the animal to image processing, connectome development and hypothesis testing.

The project led by Aravi Samuel is more focused on speeding up and optimizing image generation to improve downstream steps, with Wester’s team developing scalable, cloud-integrated image processing and data analysis pipelines. While Lichtman will support both projects, Samuel’s initiative will also be supported by Hanspeter Pfister, the academic dean of computational sciences and engineering at Harvard, and Nir Shavit, a principal investigator and computer scientist at the Massachusetts Institute of Technology’s Computer Science and Artificial Intelligence Laboratory. Shavit’s team will be responsible for co-developing machine-learning algorithms and improving their efficiency so that larger and more accurate deep-learning technology can be applied in the imaging pipeline.

“Our goal is to lower the cost of electron microscopy acquisition in order to allow every neuroscience laboratory to combine connectomics into their routine lab methodologies,” said , a researcher at Harvard and lead architect of the new intelligent image acquisition pipeline, which uses deep learning in real time during the microscopy process.

Increasing Diversity While Accelerating Discovery

In addition to providing data science and analysis expertise on these projects, APL is leading the effort to include diverse perspectives and educate early-career engineers about connectomics.

“While we were working on MICrONS, we partnered with trailblazing students to develop a new model for activating talent for students with high potential but with limited access to opportunities. One unique component of this program is that the mentoring and training activities directly supported core project contributions,” said Gray Roncal. “Those students helped us verify that the machine-learning algorithms were correctly mapping neuron connections. While the students were producing critical data for a sponsored effort, they were also gaining familiarity with revolutionary connectomics research and learning about the brain.”

Since 2017, participation in this program, called Cohort-based Integrated Research Community for Undergraduate Innovation and Trailblazing (CIRCUIT), expanded to serve a variety of academic institutions, sponsors and technical domains. Over 250 students from multiple institutions participated over seven program cycles. During these grants, CIRCUIT’s peer-reviewed model will be leveraged in new settings to support engagement across the community.

“Our continued scientific pursuit of connectomics through CONNECTS will allow us to understand the structure and function underlying brain computation. We have a critical role in both scientific discovery and inclusion across diverse communities,” continued Gray Roncal.

In 2023, the BRAIN CONNECTS initiative will provide 11 awards to institutions across the country, projected to total $150 million over 5 years.

Research reported in this press release was supported by the NIH BRAIN Initiative under award numbers UM1NS132250 and U01NS132158.