👥 Huge thanks to my incredible co-authors M. Ángeles Serrano and Olaf Sporns, and especially to @popeme.bsky.social for guiding me into the amazing world of information theory (and for your patience throughout!)
6️⃣ To wrap up 🏁
The connectome biases how information is processed:
-Redundant subsets form tightly connected configurations ideal for reliable information maintenance.
- Synergistic subsets comprise nodes ideally positioned to integrate information across the network.
5️⃣ Why does this matter?
🧷 It links connectome topology with higher-order information theory.
Understanding this structure-function link is key to uncovering the physical basis of collective computation in the brain.
4️⃣ 🔮Predicting synergy from structure
Synergistic groups of nodes are hard to find 🍀
But we show that structure provides predictive cues: 👉 the brain’s physical wiring hints at which groups are more likely to engage in synergy
3️⃣ The synergy blueprint 🌟
Synergy follows a different rule. It relies on nodes:
- Highly central
- Not confined to a single structural community
👉These subsets can integrate information across different regions, enabling non-trivial information combinations.
2️⃣ The redundancy blueprint 🧱
Redundancy dominates in densely connected subsets of nodes, where nodes tend to have:
- High clustering
- Low centrality
- Strong local overlap
👉These groups may act as robust units, built for stable, perturbation-resistant internal communication.
1️⃣ Moving beyond pairs
The brain exhibits different modes of information sharing. Using O-information, we study:
🧱 Redundancy: Information copied across regions
🌟 Synergy: Joint information beyond the sum of parts
But can we trace these patterns back to specific features of the structural network?
🧠 How is the wiring of the brain related to higher-order function?
In our new preprint, we show that the human connectome carries distinct structural signatures of synergy and redundancy.
Link to bioRxiv: 🔗 www.biorxiv.org/content/10.6...
Curious?🧵⬇️
WiNS will be back at @netsciconf.bsky.social with our mentorship program! 🎉
Application form ⬇️, more details are on our website
excited that @wins-society.bsky.social once again will have a satellite at @netsciconf.bsky.social where our mentorship program is back by popular demand! you can submit at this link: forms.gle/nFivVAiE4gdE...
Congratulations Dr. Merritt!!!
PLOS Comput. Biol.: The multiscale self-similarity of the weighted human brain connectome
journals.plos.org/ploscompbiol/article
4️⃣ Why is this important?: first, our findings highlight the relevance of weak connections in brain architecture, and second, the observed symmetry may reflect deeper principles of network organization, possibly linked to criticality.
📄 Read the full paper here: journals.plos.org/ploscompbiol...
3️⃣ A model to rule them all 💍🏔️: the WS1 model, based on hyperbolic geometry, explains the simultaneous occurrence of both weak and strong links. We apply a geometric renormalization technique based on the model to the highest-resolution scale and successfully reproduce the experimental results.
2️⃣ Weak links play a key role: the brain isn’t just about strong connections. Weak links - often disregarded - are crucial for bridging different communities, and their organization is consistent across scales.
1️⃣ The weighted structure of the brain is self-similar: the properties of the connection weights in the brain remain consistent when examining different parcellation resolutions.
Our work on self-similarity in the brain’s weighted structural connectivity has just been published in PLOS Computational Biology!🔗 journals.plos.org/ploscompbiol...
In short: no matter the scale we examine, the brain’s connection weights follow the same patterns! Let’s dive in…🧵👇🏻
We are excited to announce the 2nd edition of “Network Geometry: Theory and Applications" Satellite at @netsciconf.bsky.social .
Join us in Maastricht on June 3!
Call for abstracts : tinyurl.com/4mt2hyz3
More info: tinyurl.com/zmsh48yk