This paper, based on my Master's thesis, details the development of a predictive biomarker from multimodal data for integration into implantable devices, aiming to enable frequency-based adaptive deep brain stimulation and improve patient outcomes.

This work investigates and benchmarks the performance of various Bayesian and frequentist inverse solvers for M/EEG source imaging, with a focus on the quality of their uncertainty estimates. The goal is to provide a framework for well-calibrated source reconstruction. Publication expected soon.

This software poster was presented at the Neural Traces conference in Berlin, Germany, showcasing new multivariate connectivity features in the MNE-Python toolbox.

This project involved constructing a lead-field matrix for the AAL2 atlas to be used within the neurolib framework for whole-brain modeling.