There is abundant brain image and genetic data for researchers to understand brain structure, functions and disease. It is challenging to carry out comprehensive analyses that integrate available data. Though data generation is continuous, analyses are seldom repeated and their results updated.
The Neuro-DISK project automates the analysis of neuroscience data through artificial intelligence (AI). NeuroDISK is an AI-driven discovery engine that will continually processes neuroscience data. We use ontologies of hypotheses to represent science driving questions, semantic representations to reason about the data available, machine learning to improve accuracy, and intelligent workflows to customize the analysis to the data. Given a science question, NeuroDISK autonomously determines what data is needed, triggers the execution of relevant families of workflows, customize them to the data at hand, and alert users of interesting findings.
We are developing NeuroDISK in collaboration with researchers from the ENIGMA consortium who study different neuroscience questions.