Towards Heterogeneous Data Integration: The NeuroLang Approach

Gastón E Zanitti (1) , Majd Abdallah (1) , Demian Wassermann (1, 2)

1 PARIETAL - Modelling brain structure, function and variability based on high-field MRI data
2 MIND - Modèles et inférence pour les données de Neuroimagerie

A major goal of neuroscience research is to understand the circuits and patterns of neural activity that give rise to mental processes and behavior. But a task of this complexity often requires the integration of datasets, which are usually presented under a variety of topologies –ontologies, images, text–. This makes those analyses prone to human error and time-consuming. We believe that a tool with the ability to integrate heterogeneous datasets in a comprehensive and unified manner would alleviate the workload required for such a task. To this end, we will demonstrate how we can combine different datasets to perform an analysis that allows us to use histological knowledge of the brain and relate it to a variety of cognitive processes. Our main proposal is that having a probabilistic language based on first-order logic able to combine data under different topologies, solve complex queries and deal with uncertainty in a single framework is a great fit for this type of analysis.

Domaines: Neurosciences; Informatique [cs]

Gastón E Zanitti, Majd Abdallah, Demian Wassermann. Towards Heterogeneous Data Integration: The NeuroLang Approach. OHBM 2022 - Organization for Human Brain Mapping, Jun 2022, Glasgow, United Kingdom. ⟨hal-03786717⟩