Key Features
Powerful tools for neuroscience knowledge extraction and management
Ingest Knowledge Graphs
Upload Knowledge Graph files in JSON-LD or Turtle format. Seamlessly integrate your neuroscience data into the BrainKB ecosystem.
NER Extraction
Extract Neuroscience Named Entities from text using multi-agent systems. Leverage advanced AI to identify and classify neuroscience concepts automatically.
Resource Extraction
Extract structured resources from unstructured sources and documents. Transform raw text and PDFs into structured, queryable knowledge.
SynthScholar
Run end-to-end PRISMA literature reviews — from search strategy to synthesis, screening, and critical appraisal — orchestrated across multiple LLMs.
What is BrainKB?
BrainKB is a platform designed to support neuroscience research by structuring and organizing scientific knowledge using knowledge graphs (KGs) for delivering evidence-based insights.
Structured Knowledge
BrainKB structures neuroscience knowledge into an accessible and scalable knowledge graph.
Data Exploration
It provides tools for searching, exploring, visualizing, and analyzing neuroscience data.
Community Contribution
Researchers can contribute new findings, represented as assertions with supporting evidence from publications.
Collaboration Hub
BrainKB aims to be the go-to resource for neuroscientists worldwide, fostering collaboration and accelerating discoveries.
Use Cases
BrainKB is actively developing use cases to support neuroscience research and knowledge management
HMBA Taxonomy
Hierarchical taxonomy management for neuroscience data classification and organization.
Resources Extraction
Extract and structure resources from unstructured documents and publications.
Resource Metadata Expansion
Expand and enrich resource metadata including datasets with comprehensive information.
Neuroscientific Entity Extraction
Extract and identify neuroscience entities from text using agentic AI.
Assertion Evidence
Link scientific assertions with supporting evidence from publications and research.
SynthScholar
PRISMA-guided literature review — search strategy, screening, critical appraisal, and synthesis orchestrated across multiple LLMs.
Brain Visualization
Connect all information through interactive brain visualization for comprehensive knowledge exploration.
About Use Cases
Each use case in BrainKB includes three key components that work together to create a comprehensive knowledge management system:
- •Ingestion Process: Data submission mechanisms (public or maintainer-only, depending on the use case)
- •Public View & Interactions: Accessible interfaces for exploring and interacting with the knowledge
- •Public Feedback: Community input and validation mechanisms
All use cases are integrated with (or are on a process of integration) the BrainKB Knowledge Graph, enabling seamless data flow and comprehensive knowledge representation.
View Full Planning DiscussionStructured Models
Structured models used in BrainKB.
Genome Annotation Schema
A data model designed to represent types and relationships of an organism's annotated genome.
Anatomical Structure Schema
A data model designed to represent types and relationships of anatomical brain structures.
Library Generation Schema
A schema designed to represent types and relationships of samples and digital data assets generated during multimodal genomic data processes.
Evidence Assertion Ontology
A data model designed to represent types and relationships of evidence and assertions.
Powered by Advanced AI Agents
BrainKB leverages cutting-edge agentic frameworks for structured information extraction. Our platform utilizes STRUCTSENSE, a task-agnostic agentic framework that enables sophisticated structured information extraction. SynthScholar extends this foundation to support PRISMA-guided literature synthesis, including structured data extraction and critical appraisal of scientific studies.
Chhetri, T.R., Chen, Y., Trivedi, P., Jarecka, D., Haobsh, S., Ray, P., Ng, L. and Ghosh, S.S., 2025. STRUCTSENSE: A Task-Agnostic Agentic Framework for Structured Information Extraction with Human-In-The-Loop Evaluation and Benchmarking. arXiv preprint arXiv:2507.03674.
See Research Paper