Gabriel Stine

Anyone know what happens with NIH deadlines during a government shut down??

Interested in doing a Ph.D. to work on building models of the brain/behavior? Consider applying to graduate schools at CU Anschutz:
1. Neuroscience www.cuanschutz.edu/graduate-pro...
2. Bioengineering engineering.ucdenver.edu/bioengineeri...

You could work with several comp neuro PIs, including me.

Congrats Thomas!

Congrats! Looking forward to reading more carefully, but it sounds like this aligns closely with how I think about things!

Some journals are claiming that you need to pay big $$ for gold open access to comply with NIH's new public access policy. FYI that is total bs. You can comply by depositing the Accepted Manuscript into PubMed Central on the Date of Publication without embargo. Pass it on.

See you soon!

Please reach out if you want help with these probes. I benefited immensely by having early access to them, and so I feel an obligation/responsibility to help others adopt the tech. And, stay tuned for open-source hardware/software for a motorized microdrive to lower these (and other) probes!

Finally out! I've been using these probes for a few years and they have been truly transformative for us. The core experiments of my lab are based around them.

Please reach out if you want help with these probes. I benefited immensely by having early access to them, and so I feel an obligation/responsibility to help others adopt the tech. And, stay tuned for open-source hardware/software for a motorized microdrive to lower these (and other) probes!

Large-scale high-density brain-wide neural recording in nonhuman primates - Nature Neuroscience Neuropixels 1.0 NHP is a 45-mm, high-density silicon probe capable of recording large numbers of neurons with single-neuron resolution from most areas in a macaque’s brain.

Our paper on NHP neuropixels is finally out in Nat Neuro! These probes have already been transformative and will usher in a new era of primate neuroscience. I am extremely proud to have played a very small role in this project. I can't wait to see what our community discovers. tinyurl.com/54u3hrj8

This is amazing.

Sensorimotor Transformations for Postural Control in the Vermis of the Cerebellum The cerebellar vermis plays an essential role in maintaining posture and balance by integrating sensory inputs from multiple modalities to effectively coordinate movement. By transforming convergent s...

The cerebellar vermis keeps us balanced by integrating sensory and motor signals. In this new review, Drs. Mildren and Cullen examine how the specialized functions of the anterior and posterior vermis contribute to postural control across a variety of contexts. www.jneurosci.org/content/45/2...

A reminder that the deadline for this is coming up! Please spread the word/repost!

These charlatans seem to outdo their stupidity every single day.

The deadline for this is ⏰THIS SATURDAY!⏰ Please consider applying and share your science with the MIT BCS community!🧠

A vector calculus for neural computation in the cerebellum Null space theory predicts that neurons generate spikes not only to produce behavior but also to prevent the undesirable effect of other neurons on behavior. In this work, we show that this competitiv...

Neuronal computation in the cerebellum via a vector calculus.

Work of Mohammad Amin Fakharian, Alden Shoup, Paul Hage, and Hisham Elseweifi

www.science.org/doi/10.1126/...

A reminder that the deadline for this is coming up! Please spread the word/repost!

No, I mean your brain! We have some neural data and we're trying to understand how it relates to behavior. That's the process that is severely under constrained.

How are the subspaces defined? How do you know whether or not trajectories are moving through a null space with regard to the output? You still have to make assumptions about how neurons are combined/what signals are relevant to behavior/computation.

The manifold view (in its purest form) would miss this organizing principle entirely. One reason is because it doesn't actually define what a "population" is, beyond whatever neurons you happen to record from. Then, all bets are on for how this "pop." is linked to behavior. DOF = # of neurons.

An example where we've done this successfully is the primate visual system, where the receptive field of a neuron is informative about how it is combined downstream. I.e. there's a relationship between functional properties and connectivity that is critical to how we analyze and interpret data.

Not sure I agree re: manifolds, beyond it being trivial that structured connectivity will lead to the observation of manifolds. More important to identify the organizing principles of the system that constrain the link between brain & behav. In most cases this link is severely underconstrained.

Some examples:
www.science.org/doi/full/10....
www.jneurosci.org/content/29/2...
elifesciences.org/articles/67258
www.biorxiv.org/content/10.1...

Agreed. IMO, the key question is how much flexibility does the brain have to extract information that is present in the population activity? For a variety of reasons, I think this read-out is highly constrained, and so the brain figures out solutions that are robust and generalizable.

Many approaches assume precise control over how single neurons are combined. The "cost" of this mechanism seems high. If the brain can organize things such that random weights get you pretty far, I think that's an easier/more robust solution. I think this is one reason why topography is everywhere.

Many approaches assume precise control over how single neurons are combined. The "cost" of this mechanism seems high. If the brain can organize things such that random weights get you pretty far, I think that's an easier/more robust solution. I think this is one reason why topography is everywhere.

Agreed. IMO, the key question is how much flexibility does the brain have to extract information that is present in the population activity? For a variety of reasons, I think this read-out is highly constrained, and so the brain figures out solutions that are robust and generalizable.

I have the opposite intuition: A highly specific, highly-trained task is exactly the case where the brain could learn a task-specific, optimal decoder. In natural settings, a task-general decoder is more likely (and would thus appear suboptimal for any single task.).

Always seemed strange that this result is typically no longer taken seriously when thinking about how activity is decoded. Of course, @marlenecohen.bsky.social's group and others have done amazing work expanding on this though. All-in-all, lots of evidence that the read-out is highly constrained.