Comparison of AI, diagnostic generator, and expert opinion in difficult neurologic diagnosis: real-world case study

Abstract Objectives Comparative diagnosis of AI vs. diagnostic generator and expert opinion in complex neurologic cases is limited. We present the results of a comparative study of these three diagnostic methods. Methods A real-world study of two rare and challenging neurologic diseases was studied using 3 computer-assisted diagnostic programs, 2 AI (Chat-GPT4 and GLASS AI), diagnostic generator (NeurologicDx) and expert opinion. Results The correct diagnosis was suggested by diagnostic generator but not by expert opinion or either AI program. Conclusions Diagnostic generator outperformed AI and expert opinion supporting its use in difficult neurologic diagnosis.