Here’s a wrap-up of my PhD work with new published and unpublished results on theta phase coding. 2 Years ago we showed that during theta cycles, sequences of place fields represent trajectories that start at the position the animal reached a fixed time interval in the past and…

The EU demands X to prevent the amplification of hate, disorder or misinformation. Explicit threats of violence should be banned, but otherwise, we need free discourse. Sometimes hate and disorder are justified, and often "misinformation" ends up being accurate. We are not kids.

Or maybe it's all noise 😅

Before migraines I usually feel depressed, and after them I feel much more energized. I've used >10 years of data from https://t.co/B4vOo099nE and from my migraine records to check if this gets reflected in the music I listen to. Maybe it does? Before migraines I seek the

"If you were just a rational agent acting in the most self-interested way possible as a researcher in academia, you would cheat."

Our new 📝 in @PLOSCompBiol With @Julio_EI & @EnrRodSeb we describe the #StructureIndex, a graph-based metric to quantify how feature values are distributed over a point cloud in any dim-space (genes, neural manifolds, signals, pixels..) https://t.co/CFjt8L8ljw #DeepCode

🔔Our new 📝 with @EnrRodSeb is out in @NatureNeuro Using topological analysis we show that ripple waveforms can be represented in a low-dimensional space, which conveys information about layer-specific synaptic inputs https://t.co/w7ItgPtqWd Follow 👉🏼🧵 to discover the Minotaur!

Free tiger energy minimization. The feeling of being scratched triggers scratching movements.

Come join us! Liset is a very smart and supportive PI with a great team, plus you can have lunch under the sun almost every day ☀️

New 📝 in @Cancer_Cell! Together with @valientelab we look at the electrophysiological signatures of 🧠 metastasis, using multi-electrode recordings and #MachineLearning . By Alberto Sanchez-Aguilera, @mariam_masmudi @acnavasolive & many others👉🏼 📝 https://t.co/k2zKwUFjjw 👇🏼🧵

👀

New @biorxiv_neursci 📝 !! https://t.co/0YNKsK9dKw By @acnavasolive & Adrian Rubio. An opensource ML toolbox for sharp-wave ripple detection https://t.co/J2U85zn8mm Trained with 🐭 data and applied to 🙊 in collaboration with @perpl_lab & @SAbbaspoor #DeepCode by @CaixaResearch

For many more ideas and a crazy amount of references, check out our paper. And if you disagree with something, bring it on! We are always happy to discuss.

A topological graph (yellow) connecting known places (blue pins) is overlaid on top of metric representations (gray grid) that are approximately consistent locally but have gradually shifting reference frames and sudden breaks which result in global inconsistencies.

These results fit with reports that the regularity of grid cell firing patterns breaks down in multiple conditions. Overall, we think the evidence points to cognitive maps integrating topological and metric aspects as sketched here:

But that's not what happens. Participants could reproduce each angle by walking with the VR switched off, indicating they had learned the environment even though it violted Euclidean geometry. What's more, most didn't even notice there was anything odd going on!

Another cool one: participants learned to walk in "impossible" virtual environments with wormholes (A). If participants tried to learn a coherent metric map, they would need to deform it to preserve local continuity (B).

For example, we tend to remember road junctions as being at 90º. As a result, the remembered inner angles of 3 roads intersecting in a triangle end up adding up to more than 180º.

I'm just going to highlight one of the topics we discuss: People doing rodent electrophysiology tend to think of grid cells as providing a metric map (it works out in small boxes). However, behavioral studies show that animals and humans are generally bad at building metric maps.

We use the taxonomy as a guide to review a ton of behavioral and neural studies. We think the framework proves useful in many instances; helping to interpret neural data, point to holes in our understanding (the whole thing's a hole), and suggesting new experiments.

This taxonomy builds and expands upon others. You can also see how messy the terminology is. Different names are applied to the same processes, and different processes are referred to by the same name.