Neuro-VL Brings Conversational Analysis to Clinical Neuroimaging
Chenhui Zhao and 21 other authors presented Neuro-VL, a vision-language model trained on nearly one million neuroimaging studies from routine clinical care. The paper reports that it interprets complete imaging examinations, discusses findings in natural language, compares serial examinations, and outperformed general-purpose and specialized medical models on internal and international datasets. These are author-reported results that have not been independently verified in the supplied report.
Neuro-VL combines whole-exam neuroimaging interpretation, conversational interaction, radiology-report generation, and longitudinal comparison within one clinical-focused model.
Computer Science > Artificial Intelligence
arXiv:2611.15094 (cs)
[Submitted on 11 Nov 2026]
Title:A frontier machine intelligence for human neuroimaging
Authors:Chenhui Zhao, Akshay Rao, Akhil Kondepudi, Yiwei Lyu, Xinhai Hou, Rikhil Seshadri, Cheng Jiang, Soumyanil Banerjee, Rushikesh Joshi, Samir Harake, Ashok Srinivasan, Shruti Mishra, Francisco Rivas-Rodriguez, Seunghong Rhee, Nathaniel Roddenberry, Neeraj Chaudhary, Zach Wilseck, Paul Park, Honglak Lee, Vikas Gulani, Aditya Pandey, Todd Hollon
View a PDF of the paper titled A frontier machine intelligence for human neuroimaging, by Chenhui Zhao and 21 other authors