Machine Learning and Big Data for Neuro-Diagnostics: Opportunities and Challenges for Clinical Translation

In this report, we examine some developments in neurodiagnostics that make use of machine learning and other algorithms, with a particular focus on the potentials and challenges for clinical translation. As the ultimate aim of development of diagnostic algorithms is for their use in the diagnosis and treatment of patients, we focus particularly on the possibilities and challenges of clinical translation. We draw attention to the challenges faced in relating probabilistic predictions derived from such algorithms to individualised clinical interventions, and we highlight the importance of trust in the relationships that enable clinical translation of technologies – trust between researchers, clinicians, patients, and regulators.