CyberNeuro: A Privacy-Preserving Agentic Workbench for Cohort-Scale Neuroimage and Clinical Data Analysis

Researchers introduced CyberNeuro, a privacy-preserving multi-agent platform that automates neuroimaging and clinical data analysis from natural-language prompts.

It combines four coordinated agents—Planner, Validator, Dispatcher, and Reporter—with a secure execution setup and the local model WandaMind to keep sensitive data on-premises.

The system achieved improved benchmark performance on NeuroBench and lower estimated token usage than cloud-based approaches.

With human-in-the-loop verification for quality control, CyberNeuro aims to make cohort-scale biomedical workflows more accessible to smaller labs while protecting clinical privacy.