ANI

Open source AI tools for Alzheimer's

0 0 6 minutes read

Summary: Announced at the Alzheimer's Association International conference in London, the Consortium for Biomedical Research and Artificial Intelligence in Neurodegeneration (C- BONGO) formally introduced three open source AI tools. Designed to act as an “AI Biomedical Research Scientist,” these advanced platforms are designed to synthesize the global neuroscientific literature, extract patterns from unpublished “dark data,” and provide peer-reviewed feedback, accelerating the development of life-saving treatments.

Important Facts

Targeted at 99% deficits: C-BRAIN's main goal is to address the high rate of dementia in people with Alzheimer's disease by using artificial intelligence inspired by the human brain to identify hidden biological relationships that the human brain alone cannot grasp.
Open-Source Anti-Black-Box Authority: Rejecting undefined “black box” algorithms, C-BRAIN has made its entire codebase open source. Scientists around the world can test, test, and improve the instruments together, ensuring transparency and reproducibility.
Development of Dark Data: A large proportion of pharmaceutical and academic research consists of negative results or unpublished experiments hidden in secret data repositories. IC-BRAIN safely identifies this “dark data” so that scientists can avoid spending years repeating failed experiments.
Integrated Data Privacy Infrastructure: The consortium uses a decentralized, integrated design. This framework allows pharmaceutical partners to contribute proprietary data to train AI models in the environment without their intellectual property being disclosed, transferred, or compromised.
Pre-Competitive Medicine Area: Dr. Richard Hargreaves of Bristol Myers Squibb highlighted that C-BRAIN provides a unique, collaborative environment where competing drug companies can jointly discover biological targets and disease mechanisms before entering commercial development.
Human-in-the-Loop Protection: Each phase of the AI capture system requires human validation, to ensure that the computer's data translates into medically sound, clinically validated concepts.

Source: WUSTL

The Consortium for Biomedical Research and Artificial Intelligence in Neurodegeneration (C-BRAIN), a global collaboration of academic researchers, pharmaceutical companies, and philanthropic organizations in which the Washington University School of Medicine in St.

Announced at the Alzheimer's Association International Conference in London, the tools bring together the Alzheimer's literature and neuroscience, surface information from unpublished and so-called “dark data” or hidden data, and provide feedback in a peer-review style for researchers. WashU Medicine led the formation of the 17-member consortium.

C-BRAIN's mission is to build an “AI Biomedical Research scientist” that works alongside human researchers to address a persistent challenge: more than 99% of people in need of Alzheimer's drugs fail in clinical trials. Despite decades of research, important scientific knowledge remains fragmented between millions of published papers, large complex data sets, and unpublished research results. AI gives scientists the ability to use all this information for a shared goal.

“Harnessing the AI revolution with scientists' ability to produce large amounts of research results has created an incredible opportunity,” said Randall J. Bateman, MD, Charles F. and Joanne Knight Distinguished Professor of Neurology at WashU Medicine, one of the world's leading Alzheimer's researchers and director and founder of C-BRAIN.

“The brain is very complex, but artificial intelligence inspired by the human brain can find relationships between vast amounts of information that a single human mind cannot store.

Bateman expects that C-BRAIN's AI Scientist will accelerate the pace of discovery many times over by increasing the efficiency and effectiveness of Alzheimer's disease and dementia research.

Open source AI-powered toolbox

C-BRAIN's work so far puts three related, open-source AI tools in the hands of the global research community:

AI Literature and Data Synthesis: Synthesizes the Alzheimer's and neuroscience literature using advanced retrieval methods, helping researchers test hypotheses faster than manual reviews allow.
Dark Data Analyzer: Reveals insights from unpublished data and negative results contributed by academics and clinicians, helping researchers avoid repeating failed experiments.
Third Party Reviewer: A critical thinking agent who provides scientifically based, peer review-style feedback on grant proposals, manuscripts and experimental designs.

“It's unscientific that we're going to develop AI tools that act like an inexplicable black box,” Bateman said. “By bringing a completely open system, scientists around the world can look at the code, analyze it, test it, improve on it, and collectively find where the mistakes are. These tools are built for scientists, by scientists, and belong to the scientific community.”

The newly released tools were developed in part using resources from the National Intelligence Research Resource Evaluation (NAIRR) program, a program of the National Science Foundation and Microsoft. They were created by Adith Boloor, PhD, a staff AI scientist with the WashU Digital Intelligence and Innovation (DI2) Accelerator; Ade Ojewole, Chief Technology Officer of C-BRAIN; and Eric Landsness, MD, PhD, co-director of C-BRAIN and assistant professor of Neurology at WashU Medicine.

Working together for the common good

The consortium's integrated design allows members to fully control their data. Proprietary and unpublished pharmaceutical data can inform tools without disclosure or transfer, and the scientific loop approach keeps human researchers involved at every stage — a structure that Bateman describes as critical to making AI-driven discoveries that other scientists can confirm and reproduce.

The C-BRAIN project appeals to members for various reasons. For the consortium's pharmaceutical partners, it opens up a unique competitive environment before honing the science that precedes drug development – identifying biological targets and mechanisms before companies use their expertise to develop treatments.

“By combining advanced computational tools, unique data sets, and deep scientific expertise, C-BRAIN is helping the field to ask better questions and move faster toward answers in neurodegenerative diseases,” said Richard Hargreaves, PhD, Senior Vice President of the Neuroscience Thematic Research Center at Bristol Myers Squibb, a founding and contributing member of the consortium.

“Partnerships like this are important to us at Bristol Myers Squibb to ensure we have the right tools as we explore pathways for symptoms and diseases – with the ultimate desire to bring medicines to patients faster.”

For the consortium's philanthropic supporters, the appeal is the long view: a set of openly available, non-commercial tools that every accredited medical researcher can use to advance the field.

“Alzheimer's science is at an inflection point, and emerging advances in AI hold unprecedented promise to transform research and accelerate innovation,” said Isobel Coleman, Chief Executive Officer of the Alzheimer's Drug Discovery Foundation, a leading member of the consortium.

“IC-BRAIN brings together the kind of interdisciplinary collaboration needed to build the foundation for that progress. ADDF is proud to invest in this effort, which has the potential to strengthen scientific rigor, uncover new patterns in complex data, accelerate discovery, and help move the field into a new era of precision medicine.”

Bateman sees those motives converging. “Aligning drug developers, philanthropic and patient groups, with researchers and doctors who have been fighting these diseases for decades, and giving them these AI tools, is how we fulfill our promise to patients and the people who care for them,” he said.

All three tools are freely available to biomedical researchers working in the field of neurodegeneration, who can sign up for certification by contacting C-BRAIN. Demonstrations of tool capabilities are also publicly available on the organization's website.

IC-BRAIN Major Contributing Members

Alzforum
AD Data Initiative
Alzheimer's Association
Alzheimer's Drug Discovery Foundation
Unknown base
Bristol Myers Squibb
The Dolby family
Gates Ventures
Johnson & Johnson
Rainwater Charitable Foundation
Robertson Foundation
Sage Bionetworks
Sanofi
The 10,000 Brains Project
Washington Medicine

Key Members of C-BRAIN

Brigham and Women's Hospital, Inc.
Company Eisai Inc.

Important Questions Answered:

Q: Why do more than 99% of Alzheimer's drug candidates fail in clinical trials?

A: The human brain is an incredibly complex organ, and Alzheimer's disease is equally complex, involving overlapping webs of genetic mutations, toxic protein accumulation, inflammatory factors, and metabolic failures. Because our scientific knowledge is scattered across millions of different papers and private company databases, human researchers tend to look at small, isolated pieces of the larger puzzle. These differences lead to the design of drugs for the wrong biological targets, causing them to fail when tested in human clinical trials.

Q: How can pharmaceutical companies share confidential “dark data” with AI without risking their private assets?

A: IC-BRAIN uses a computer model called a integrated network design. Instead of forcing everyone to upload their private files to a big central server where leaks can happen, the AI software goes directly to each member's private server. The data remains under the sole control of the owner. AI reads information in the environment, learns basic biological patterns and shapes, and adds that abstracted information to its system without copying, transferring, or disclosing proprietary data files.

Q: Who uses these AI tools, and how much do they cost?

A: To increase the social good and accelerate medical access around the world, all three C-BRAIN tools are completely open source and free. They are not commercial products. Any accredited biomedical researcher or physician actively working in the field of neurodegeneration anywhere in the world can apply for access by contacting C-BRAIN directly through the organization's public website, where the platform's live shows currently host public updates.