Thank you all for the great chats about this work and the inspiring ideas for where to go next with this project!

And a big thank you to @bendfulcher.bsky.social and @macshine.bsky.social for your guidance and support throughout this project, as with all others of my PhD 😊

Big shout-out to @jlizier.bsky.social, whose patience, dedication, and guidance in supervising me through this project have been invaluable 🙏 Thank you also to our other fab co-authors for their insightful input as always: Oliver Cliff, @macshine.bsky.social, and @bendfulcher.bsky.social!

I learned so much through working on this piece! My goal with this is to share a beginner-friendly and intuitive guide for all my fellow visual learners out there, that also doesn’t shy away from the maths and important nitty-gritty details 😊

Unbiased opinion, but I’m really interested in how we can get the most bang for our info theory buck in computational neuroscience 🧠 So, we demo the implementation and interpretation of all 11 measures with an illustrative BOLD fMRI case study:

We organized 11 measures into 6 categories according to temporal dependency and the type(s) of dynamical properties each measure captures. These measures are summarized in one large ‘pocket reference’ schematic, and we then zoom in to discuss the ‘what/how/why’ of each measure individually 🔎

Thank you very much to the COGITATE Consortium for openly sharing the MEG data we used in our analyses, and to our supervisors @bendfulcher.bsky.social and @macshine.bsky.social for supporting our first fully independent PhD project. Hope you enjoy, and we’re very open to feedback! 😊

All the data viz code is in our repo (though it's currently still a bit of spaghetti code while I finish organizing my bits 🍝): github.com/DynamicsAndN...

And lastly I combine all my figures in Inkscape!

Thank you, Tara! I swear by the R package "ggseg" (led by @drmowinckels.io) for brain map visualization and the python package @networkx.bsky.social for the circular network connectivity plots 😊

Thank you to @aditijh.bsky.social for welcoming me to join her in this project this last year, to our supervisors @bendfulcher.bsky.social, @alexfornito.bsky.social, @lawraga.bsky.social & fantastic co-authors Patrick Cahill, Brandon Lam, @stuartoldham.bsky.social, and @aurinaarn.bsky.social!

Together with new biological insights into the brain’s overlapping structural architecture, this work introduces a generalizable method for systematically comparing overlapping community detection algorithms to an ensemble of simulated networks tailored to your particular dataset at hand:

In contrast to Louvain clustering, which doesn’t allow overlapping assignments, OSLOM can disentangle communities bridged by two key regions, with the flexibility to capture the brain as a complex system characterized by fluid distributed interactions: