A new home for LaBGAS on the web

This site brings together things that were previously scattered: a description of what the lab actually works on, who is in it, and documentation for the analysis code we develop and release.

That last part is the reason for building it. The labgas GitHub organisation holds two dozen repositories — two carrying reusable methods, the rest study-specific code released alongside papers — with no index and no entry point. Useful code that nobody outside the lab could find, and that new lab members learned by asking someone.

The pipeline documentation now walks through a study from raw scanner output to archived result reports: BIDS conversion and prep, first-level modelling, second-level analysis with the CANlab batch system, and our PLS-DA, PLSR and Elastic Net pipelines, plus PET, MR spectroscopy, connectivity and atlas tooling.

Throughout, we have tried to be explicit about what is ours and what is not. Our neuroimaging code sits on top of the toolset built by the Cognitive and Affective Neuroscience Lab — CanlabCore, Neuroimaging_Pattern_Masks, and the second-level batch system. LaBGAScore is a layer, not a replacement, and the docs link out to the CANlab material rather than restating it.

The Tools page indexes every repository in the organisation, and includes a table setting out which layer of the stack comes from where.

Corrections and additions are welcome at lukas.vanoudenhove@kuleuven.be.