Reading papers from 100 years ago is so much better.

Here is a cartoon for the modern age. Bullet to Text to Bullet

bsky.app/profile/newr...

You know Trump-Vance are wrong. Another case of eating your pets nonsense.

Framing it as a question is doing harm. Fix your headline.

Sorry. Haven’t been here in a while. Minor health issues.

I am interested to see your notebook.

Everybody would like the war to stop and the hostages be returned. And Hamas to leave.

These two have failed to get it done after saying they would. Repeatedly.

Witkoff struggled to name the five regions of Ukraine that are either annexed or partially occupied by Russian forces. He mentioned "the so-called four regions. Furthermore, his statements echo Russian propaganda. So, not great.

Also, 3 long range rockets from Gaza to Tel Aviv yesterday.

Understanding learning at 2 timescales (e.g., fast neuron-neuron and slower neuron-glia-neuron) seems important. We learn much over long time periods.

One learning mechanism, synaptic strength, might be possible. But ould mechanisms at 2 timescales have some value?

Right. But the hippocampus is unusual in many ways.

Modeling these types of circuits, even with a simplified basic model, seems like a real need. Something beyond the single scalar neural network.

If differentiable, it could immediately be a target to explore in applications.

Thanks. Very helpful. I see your point and am grateful for the link.

These are the doubts, in this case about gliotransmitters.

en.wikipedia.org/wiki/Gliotra...

The authors make a point I find quite interesting. The energy consumption is very different in these regions. Hence the molecular signaling and neural-glial circuitry are also likely to be very different.

If so, the principle of a canonical cortical circuit is unlikely. #neuroskyence

Thanks. Should I take your comment as suggesting that current measurements are not yet convincing to the cellular signals community? I am reading reviews about the tripartite synapse (e.g., A conceptual framework for astrocyte function, Nat. Neuro) and many related. Are there doubts?

In sensorimotor neuroscience, it is a given that one needs to know the input and/or output to understand the neural signals.

A good principle, but harder to operationalize in cognitive and affective neuroscience. Surely worth the effort, IMHO.

Another tool for noise correction has been developed by Kendrick Kay. Largely for task-based fMRI. Originally, GLMDenoise and now extended to GLMSingle. Kendrick used to maintain a table comparing different methods, but I haven't seen that lately.
github.com/cvnlab/GLMsi...