InsightEngine
Integrated AI platform for Alzheimer’s & neurodegeneration research
Insight Engine is C-BRAIN’s software platform for Alzheimer’s and neurodegeneration research. It combines three agents that share a common knowledge graph and provenance layer, supporting three stages of research work: synthesizing existing literature and data, testing hypotheses against experimental data, and reviewing manuscripts and proposals. Outputs are linked to their sources, and researchers can review, modify, or correct the system’s intermediate steps. The platform is intended for biomedical researchers, principal investigators, and reviewers, and is accessed through a web-based interface.
InsightEngine Agents
Literature Review & Data Synthesis
Ingests Alzheimer’s literature, including PubMed articles and consortium datasets, and extracts entities and relationships into a curated, persistent knowledge graph. It answers natural-language queries by retrieving from the graph and the indexed corpus, linking each statement to its source publication. An interactive graph view and a correction workflow allow domain experts to flag or amend extracted facts.
Data Analysis
An agentic system for hypothesis testing and failure-mode analysis that uses both published results and unpublished (“dark”) data, including negative and replication datasets. Given a hypothesis, it generates a test plan specifying the required datasets, replication checks, key variables, and analysis steps; harmonizes heterogeneous inputs into a common schema that links each result to the procedure that produced it; and runs statistical or code-based analyses in a sandboxed environment. It reports likely failure points, confidence levels, and probable reasons for discrepancies between studies. Outputs include data identifiers, references, code, and execution logs.
Critical Reasoning
Parses manuscripts, proposals, and reports, including figures, graphs, and tables, and evaluates them across significance, innovation, approach, rigor, and ethics. It applies several reviewer perspectives — editorial, methodological, and domain-expert — to produce structured feedback that identifies weaknesses and proposes improvements. Each comment is linked to its supporting evidence.
Architecture
The three agents run on shared infrastructure. The knowledge graph produced by the synthesis agent provides evidence grounding for the analysis and review agents. All outputs carry source provenance — identifiers, citations, code, and execution logs — and dataset, protocol, and workflow versions are tracked for auditability. The platform supports switchable underlying models and is containerized for deployment across major cloud providers (AWS, GCP, or Azure). Components are released open-source within the consortium, with documentation for adding new datasets, analysis capabilities, review templates, and reviewer personas.
Source Code
InsightEngine User Interface
The InsightEngine Web application provides the primary user interface for interacting with the InsightEngine platform. It enables conversational AI workflows, project management, knowledge graph exploration, and access to research-oriented analysis services exposed by the platform backend.
InsightEngine Backend
The C-BRAIN Backend API Service acts as the central application layer within the InsightEngine platform. It coordinates interactions between users, AI workflows, knowledge graph services, and supporting infrastructure.
InsightEngine AI
A research-oriented AI platform for evidence-backed hypothesis testing, literature retrieval, and scientific reasoning over biomedical content.
InsightEngine User Service
A user management and identity service built on .NET and PostgreSQL. The User Service provides comprehensive user authentication, authorization, role management, and organizational hierarchy capabilities for the InsightEngine platform.