And now: The revealing neuro-evo-trivia game - what's what? 😁

Not yet - but will

Only one way to figure out

...only good old mechanics. My youtube algorithm washed this onto my timeline. But: this would really be cool for preventing craniotomy brain damage. Probably unaffordable for small animal science. www.youtube.com/watch?si=qHW...

Excellent! Will come back!

Was about to suggest BonnBrain! Let me come back to you - I will organize something. We als have fantastic human action segmentation people around.

💪 (if only that would be how one spells Ratatouille :) )

GPT says: "Rodent Action Tracking Across Tasks: Objects, Utilities, Levers, Locomotion, Interaction, and Exploration — a mouse “kitchen” benchmark for dense, sequential, ethological behavior."

...more like a dense foragy/approach-avoidy mix of Josh's FOMO, SWC's AEON, some closed-loop stim control, some risky decision with shelter-seeking /-building, some sequential goal / spatial decision, some digging, some exploring. And yes: difficult. The acronym would have to be RATATOULLIE.

Really not sure what that would be. But fun to think about...

Would be totally up for it! Not easy, though. Social aside (@antihebbiann.bsky.social has that covered), a mouse-kitchen should be dense (many actions per time), rich (max affordances), with interpretable goals, sequential dependencies, ethological grounding, and transferable across labs...

nice - did not know this one. Sadly - mouse ego-cams need cables and commutators. So the IntelliMaze RFID gates are an issue (which we currently face ourselves)

Our (and others’) unrestrained ego/exo tracking is certainly not at the scale of your human datasets, but it at least begins to approach Meta’s ego-exo4D technically...

Wow - very cool. At some point, we small-animal folks should come together to define what annotated ego/exo benchmark data _we_ want to generate—and how best to share it. What would be the mouse equivalent of the human “kitchen”?