Don’t forget to check out our poster at SfN 2024 meeting! Say hi to @ezeyulu00 at poster board number F31 (PSTR075) in MCP Hall A, on October 6th, 2024 between 8AM-12PM.

Also shoutouts to @eazim, @timwdunn, @jessedmarshall, @BOlveczky, @b_ignatowska, @spencelocolab (and others) for inspiration. And the keypoint tracking community has created some amazing tools that are actually easy to use.

This work was driven by the singular vision and dedication of Zeynep Ulutas, @ezeyulu00 , who is applying to grad school this cycle (don’t miss out!). We could not have finished this without @AmartyaPradhan. @DKoveal provided key guidance on chemistry.

We’re pumped about the future of this technique and are excited to share it with the community. Datasets with millions of frames are easily within reach, and we're thinking about using this in other species as well.

Using our method we built a dataset with 120,000 automatically labeled frames. Since the labels are localized through fluorescence, no labeling ambiguity! Turns out with this many frames you can train techniques like SLEAP and DLC to label body parts with sub-mm precision.

We also worked out a machine learning pipeline to take data where we fluorescently tagged 10 body parts and automatically labeled all frames without human intervention.

We successfully targeted the fatty tissue beneath the skin and even the knee joint! Both of these were readily resolvable with off-the-shelf cameras. If you’re handy with a screwdriver you can pluck out the IR filter in your cameras and try it yourself.

Standard particle mixtures diffused away within 48 hours. We formulated an alternative and now we can watch the insides of mice glow for weeks as they freely move! We call the end-result Quantum Dot-based Pose estimation in vivo or…QD-Pi.

With the first person in the lab, @ezeyulu00 , we came up with the idea of injecting near-infrared-fluorescent nanoparticles to make little orbs of light inside the body that we could see from the outside.

Markerless trackers like DLC, SLEAP, DANNCE, Lightning Pose, and others have totally revolutionized tracking movement since they don’t require the mice to wear anything. BUT, their precision still lags behind marker-based methods.

But mice don’t like swimsuits and they don’t like anything being attached to their skin. And even if you did manage to get something onto the skin they have lots of fatty tissue, which makes tracking of key body parts like joints and bone hard to track from the “outside-in”.

If we want to understand how the brain controls movement, we have to track it. In people, we can do this by putting on…a swimsuit bedazzled with ping pong balls!

First paper from the lab Long story short: @ezeyulu00, @AmartyaPradhan, @DKoveal and I developed a method using injectable nanoparticles to turn mice into…constellations in motion.

RT @KyleAJohnsen: Want to use closed-loop or all-optical methods to improve your systems neuro experiments but are lost in the sea of new m….

RT @talmop: Keypoint-MoSeq is out!. Related 😁:.

RT @joachimgoedhart: Very happy to see this published: Make the most of the data that is gathered by students e.g. in lab courses by "Stude….

RT @biorxivpreprint: Dynamics of striatal action selection and reinforcement learning #bioRxiv.

RT @Shaughness95562: 1/ I am excited to introduce my PhD work with @twit_tats2, @greg_stephens, and @shaevitz, on the 3D tracking and analy….

RT @IshmailSaboor: Excited to share our new pre-print where we discover behavior signatures of eusociality and social touch in the East Afr….

RT @IDsignals: New preprint!. Led by PhD student @VogtCaleb with @MatthewZipple @DanielSprockett @andrewhmoeller and others not on Twitter.….