Computational neuroscience, Physics of Complex Systems, Bio-Inspired Intelligence, Foundations of Physical Computing
https://neurovium.science/
https://compneuro.mit.edu/home
Loading...
As local neighborhoods were progressively excluded, performance decreased.
So local electrodes carry the strongest predictive information.
But reconstruction remained above zero even under strict local masking.
The signal is locally anchored, but not purely local.
5/n 🧵👇
The conclusion is not that field potentials are local or global.
They are both.
Electrode signals are local observations of distributed dynamical systems, filtered through tissue, sensor geometry, and network organization.
SMR gives us a way to quantify that balance.
7/n 🧵👇
If reconstruction collapses when nearby electrodes are removed, the signal is mostly locally redundant.
If reconstruction survives local masking, then some information about that target is distributed across more distant parts of the array.
4/n 🧵👇
arxiv.org/html/2606.11...
The key move is spatial masking.
For each target electrode, we reconstruct its time series from other electrodes while progressively excluding its local neighborhood.
This turns “locality” into an experimental knob.
3/n 🧵👇
neurovium.science/posts/pblog-...
Try it!
You can use this for any form of 2D/3D electrode array.
n/n 🍸
github.com/neurovium/Sp...
Surrogate controls showed that SMR was not simply exploiting marginal statistics.
Phase-shuffled, IAAFT, and block-shuffled surrogates all substantially reduced performance.
So reconstruction depends on structured temporal and cross-channel organization, not just spectra or amplitudes.
6/n 🧵👇
How local is a “local” field potential?
In our new paper 📜, we introduce Spatially Masked Regression (SMR), a reconstruction-based framework for asking how much of an electrode signal is locally redundant, and how much is embedded in broader distributed dynamics.
1/n 🧵👇
arxiv.org/abs/2606.11415
The usual approach is functional connectivity: correlation, coherence, PLV, mutual information, graph edges, etc.
These methods ask whether two signals are statistically related.
SMR asks a different question:
Can one electrode’s signal be reconstructed from the rest of the array?
2/n 🧵👇
Recent work from the lab: big contrast between the synaptic dynamics of excitatory and inhibitory synapses!
See the paper card (for link and the blog) on the recent manuscript on inhibitory neurons across different layers of the cortex!
ps...the universality paper (quoted tweet) will be a follow up on this one.
neurovium.science/papers/Inhib...
