NeuroAgent: LLM-Driven Framework for Multimodal Neuroimaging Analysis
A new LLM-driven framework called NeuroAgent automates preprocessing and analysis of multimodal neuroimaging data, including sMRI, fMRI, dMRI, and PET. The system uses a hierarchical multi-agent architecture with a feedback-driven Generate-Execute-Validate engine to autonomously generate code, handle errors, and validate results. It supports interactive downstream analysis through natural-language queries, bridging the gap between raw acquisitions and reproducible scientific analysis.
Key facts
- NeuroAgent is an LLM-driven agentic framework for multimodal neuroimaging analysis.
- It automates preprocessing for sMRI, fMRI, dMRI, and PET data.
- Uses a hierarchical multi-agent architecture with a Generate-Execute-Validate engine.
- Agents autonomously generate executable preprocessing code.
- Detects and recovers from runtime errors.
- Validates outputs automatically.
- Supports interactive downstream analysis via natural-language queries.
- Aims to reduce barriers between raw data and reproducible analysis.
Entities
—