Digital twins in neuroscience: A narrative review of technical challenges for child neurology

Abstract

Digital twins are dynamically updated, patient-specific computational representations that integrate multidimensional data to support prediction and clinical decision-making. This narrative review examines the conceptual foundations and enabling technologies of digital twins in neuroscience, representative proto-twin platforms such as The Virtual Brain and the Virtual Child, and the contribution of digital biomarkers. Our main finding is that most neurological applications remain patient-specific simulations or proto-twins rather than clinically mature digital twins, and prospective evidence of decision impact are usually absent. Pediatric evidence is especially limited and currently consists of two proof-of-concept applications, the infant microbiome digital twin for neurodevelopmental outcome prediction and the neuromotor digital twin for longitudinal monitoring of preterm infants. Translation into child neurology will require developmentally aware longitudinal datasets, interoperable data standards, external and multicenter validation, transparent uncertainty reporting, and prospective implementation studies. Digital twins should therefore be positioned as clinician-supervised decision-support systems, not autonomous substitutes for clinical expertise.