Neuroscience research introduces virtual brain twins for clinical use

Professor Viktor Jirsa of Aix-Marseille Université presented research on the evolution of neuroimaging from static brain maps to the development of ‘virtual brain twins.’ Unlike standard simulations or statistical models, a digital brain twin integrates three components: individual brain data from multimodal imaging, a generative mechanistic model of network activity, and a bidirectional inference loop to estimate latent parameters.

This approach aims to provide clinical actionability by addressing mechanistic and individual questions, such as why specific brains generate seizures. The technology is currently being applied to clinical trials for epilepsy and schizophrenia, using Bayesian inversion to estimate parameters within the subject’s connectome.