From rodents to algorithms: The rise of virtual animals in drug development and neuroscience

Animal models have long been the foundation of biomedical research and experimentation. However, the utility of animal models in both research and preclinical testing of drugs has been challenged by the insufficiency and cost of such methods, including the fact that over 90 % of preclinically successful compounds ultimately fail in human trials, as well as concerns about irreproducibility, scalability, and ethics in animal studies; their limitations are increasingly recognized. Advances in artificial intelligence (AI) now offer a promising complementary approach through the creation of "virtual animals." Recent examples include the U.S. Food and Drug Administration's AnimalGAN to predict toxicology outcomes in thousands of virtual rats, and Google DeepMind's virtual fruit fly, an anatomically accurate model that can reproduce locomotion and sensorimotor behavior. In these examples, AI systems may not replace, but rather support traditional experimentation, by increasing accuracy, reducing the number of animals needed for studies, and accelerating translational insight. Potential early applications in neurology and neurosurgery are just now emerging, such as neural-circuit simulation, prediction of neurotoxicity, and modeling of movement disorders. Legislative acts such as FDA Modernization Act 2.0, as well as European initiatives to implement the 3Rs, now include AI-based virtual testing as a possible evidentiary basis. Virtual organisms represent a potentially significant ethical and methodological shift that may influence the future of experimental neurology and drug development, while still upholding scientific rigor.