Software
We develop and maintain open-source software that implements and benchmarks machine learning methods for neuroimaging data, supporting reproducible research and scalable analysis across studies and sites. Our software emphasizes robustness, transparency, and interoperability, enabling researchers to apply, evaluate, and extend state-of-the-art methods across different datasets and experimental settings. By prioritizing open development practices and well-documented workflows, we aim to facilitate methodological reuse, foster collaboration, and support the translation of computational methods into neuroimaging and clinical research.
In-house software
MEGaNorm is a Python package that wraps MNE-Python and PCNToolkit functionalities for extracting functional imaging-derived phenotypes (f-IDPs) from large-scale EEG and MEG datasets, and then deriving their normative ranges. It allows researchers to analyze large MEG and EEG dataset using high-performance computing facilities, and then build, visualize, and analyze normative models of brain dynamics across individuals.
External software contributions
A Python package for normative modelling, spatial statistics and pattern recognition.