This paper is now published in Biometrics. View the preprint here.
This is a fully automated framework for translating radiological diagnostic criteria into image-based biomarkers of brain edema (swelling), inspired by a project in which children with cerebral malaria were imaged using low-field 0.35 Tesla MRI.
Code to replicate the pre-processing pipeline for T1 and T2 brain MRI scans, calculate imaging biomarkers, and assess the performance of the predictive model can be found at https://danni-tu.github.io/cerebral-malaria/.