The goal of our nation-wide initiative is to provide clinicians intuitive and robust tools for accurate diagnosis, therapy monitoring and prognosis of cognitive decline that is based on large-scale multidomain data.
We describe a federation framework that allows for statistical analysis of aggregated brain imaging and clinical phenotyping data across memory clinics in Switzerland. The adaptation and deployment of readily available data capturing and federation modules is paralleled by developments in ontology, quality and regulatory control of brain imaging data. Our initiative incentivizes data sharing through the common resource in a way that provides individual researcher with access to large-scale data that surpasses the data acquisition capacity of a single centre. Clinicians benefit from fine-grained epidemiological characterization of own data compared with the rest additional to intuitive tools allowing for computer-based diagnosis of dementia. Finally, our concept aims at closing the loop between group-level results based on aggregate data and individual diagnosis by providing disease models, that is, classifiers for neurocognitive disorders that will enable the computer-based diagnosis of individual patients.
The obtained results will inform recommendations on best clinical practice in all relevant fields focusing on standardization and interoperability of acquired data, privacy protection framework and ethical consideration in the context of evolutive pathology.
aLREN, Department of Clinical Neurosciences, University Hospital Centre (CHUV) and University of Lausanne (UNIL), Switzerland
bMax Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany
cLeenaards Memory Centre, Department of clinical neurosciences, University Hospital Centre (CHUV) and University of Lausanne (UNIL), Switzerland
Correspondence to Bogdan Draganski, LREN, Department of Clinical Neurosciences, CHUV, University of Lausanne, Lausanne, Switzerland. Tel: +41 21 314 96 38; e-mail: firstname.lastname@example.org