Almost forgot the important coda:

Lydia's celebration dessert of choice was tiramisu. This version was true to Seattle, with amped up levels of coffee over traditional ones ☕😋

Big congrats to Lydia on creating this tool that empowers better science and fuels refinements in our use of animals. Hope others find these resources as useful as we do, and we’d love ideas on ways to continue making our simulators better and more useful!

We’re also working to set this up as a fee-for-service pipeline our lab can offer to folks who might want a custom model but don’t want to do it themselves – stay tuned!

We find these SUPER useful for many things in the lab, which we describe in the paper. We’re excited to help other folks make their own by sharing the details of how we make ours. Lydia also made & shared model files for several different non-human primate species folks can use to make their own.

This method report shows how we go from an image of an animal (e.g. MRI) to a simulator that includes the skull, brain, muscle, & skin. Lydia also came up w/ clever ways to make things modular so it’s easy to adapt depending on what you want to use it for, & as animals move through an experiment.

So! Lydia combined a bunch of existing ideas to build customized, 3D, & modular simulators that include most of the key tissues for neuroscience research & implant design. We do this for non-human primates 🐵but it could work for any species, including people (Neuralink has some cool related work)!

Tissue simulators exist, but we found them a bit limited. They mostly model 1 thing (e.g. just the brain). Or they don’t capture the tissue's 3D geometry (e.g. skin suture boards are flat). Thinking about how tissue layers interact & their geometry could improve implant outcomes and improve training

Tissue simulators help people practice skills and develop new methods. For instance, most every med student first learns to suture on a suture board and skin mimics. Neuroscience researchers also use of things like gel brain replicas to test procedures (e.g., injections, inserting probes).

Epic congrats to Ryan and all of his co-authors. This is a heroic effort to collect exciting new electrophysiology data, AND a lot of thinking to make sense of what these new data tell us.

And, of course, a massive thank you to our 🐵A & B for making these discoveries possible.

Lots to unpack and curious what folks think. But we are excited to start getting a new view of how motor cortex neuronal populations represent movement and coordinate to control behavior. And this 'functional network' perspective raises lots of questions for how to best target implants for BCIs!

Interestingly, which neurons coordinate w/ shared dynamics were strongly related to whether they had info about the task (in a seemingly non-trivial way). We think our data hint that motor cortices are less a big homogenous population, and more like a mix of distributed functional networks!

He paired a giant pile of data w/ decoding & population dynamics analyses. He found a surprising amount of heterogeneity! Info about the task was sprinkled across populations without strong spatial organization. Which neurons coordinated as a population also varied a lot.

MC looks like a jumble @ neuron levels but seems coordinated when looking @ populations. But population studies mostly lump neurons all together w/out looking for spatial structure. Ryan used Neuropixels to widely sample neural populations across frontal motor cortices to quantify spatial structure.

Also, got a glimpse into how America's seen by the rest of the world. Can't say I disagree 🔥🫠😞

Also taking in the scenery and pastries in Frankfurt 😍

I was training for a half marathon this summer, so I've been hovering around 12-13 miles for a bit. I learned I kind of love these long runs (still shocking to me, honestly!) but am intimidated by the idea of a marathon. So my new goal is to try ~20mi (a 3/4 marathon) 😅.