King's College London

Computational brain models use machine learning, algorithms and statistical models to harness big data for delivering disease-specific diagnosis or prognosis for individuals. While intended to support clinical decision-making, their translation into clinical practice remains challenging despite efforts to improve implementation through training clinicians and clinical staff in their use and benefits. Drawing on the specific case of neurology, we argue that existing implementation efforts are insufficient for the responsible translation of computational models. Our research based on a collective seven-year engagement with the Human Brain Project, participant observation at workshops and conferences, and expert interviews, suggests that relationships of trust between clinicians and researchers (modellers, data scientists) are essential to the meaningful translation of computational models. In particular, efforts to increase model transparency, strengthen upstream collaboration, and integrate clinicians' perspectives and tacit knowledge have the potential to reinforce trust building and increase translation of technologies that are beneficial to patients.