Continuous testing of neuroimaging results across pipelines and datasets
Application to hearing loss associations with brain structure
Given the ongoing reproducibility crisis in Neuroimaging, the goal of this project is to evaluate the variability of results across a range of processing pipelines and datasets. The initial use-case to demonstrate the functionality of the framework is hearing loss associations with brain structure (in particular, left hippocampal volumes).
In order to achieve this goal, a computational platform must be constructed that is able to iteratively evaluate and update associations between the hippocampal volume and hearing loss as new data and pipelines or processing conditions are added.
Check out the eScience publication describing the initial architecture and results of the use-case. The framework is still under active development and evolving towards a more robust and user-friendly software design.
People
Jacob Sanz-Robinson
