Thanks to everyone in the lab for getting this out the door. And we've got way cooler M1-RN stuff in the works - so stay tuned!

Unfortunately, the methods we used weren't sensitive enough to pick that up since that kind of detailed kinematics wasn't the point of that study. But we're doing our best to be sure that doesn't happen again! Here's recent data from the lab, showing we can now track limbs in 3D throughout movement.

Turns out - we see the same pattern, bursting for some twitches but inhibition for others. What's super cool is that it's sped up. The burst and the inhibition is only ~25% as long as during wake, so it's like the whole motor system is accelerated. So what kind of twitches happen during the bursts?

Now for the unpublished part:

At P24, we found RN neurons that showed activity during specific wake behaviors (Fig. 6). The one we show in this fig fires preferentially during grooming - and shows strong inhibition during other wake movements. So what's that same RN neuron doing during REM sleep?

So can you use EMGs to detect twitches. In short - yes. But there's a couple of catches.

1) you're recording from a single muscle (maybe even just a few muscle fibers) so you're getting a limited look at movement.

2) There's WAY more false negatives (only ~20% of twitches are detected vs. video)

Same basic story for the P24 M1-RN data (Fig. S4) - there's a huge discrepancy between EMG and movement during wake, but for twitches, the correspondence is pretty good.

The reviewer-suggested addition: We had EMGs, but triggered twitches and wake movements on video. How do they compare?

The figure is in the supplement (Figure S3). Twitches = purple; wake movements = green.

Basically, for twitches, EMG tracks 100 fps video quite well - less so for wake movements

This has been a long time coming - but our M1 Development/M1-RN paper is finally out in its final form!

Rather than re-hash the original story, I'm going to talk about one of the biggest reviewer-suggested additions, as well as some unpublished data!

www.nature.com/articles/s41...

I keep going back to how tough it would be to explain REM sleep and twitching to an alien (assuming they don't do it too, which maybe isn't a given)

....I notified them as well - and was relieved when they responded within the 48 hours... Im not much of a rebel

a bald child with the words there is no spoon written below him ALT: a bald child with the words there is no spoon written below him

I got this like a month ago, and learned (sadly for the first time) its a total bluff. The only rules are the ones you believe...

Actually don't really know if it's cortically driven or not. I used to think that was a sure thing - but now I'm less sure on that. We argue in the paper that it reflects cortical magnification, but honestly - collicular motor maps show similar magnification patterns. So who knows

...part of me wonders if they start to reflect Graziano's "ethological action maps". In P20 rats (but not earlier), I see alternating bimanual twitches (reaching/holding), twitches of all 4 limbs (locomotion), and whisker/forelimb twitch bouts (grooming, feeding) that occur together quite frequently

Lots of thoughts!

Since I've mostly looked at twitches in a developmental context, I can say that I've definitely noted that twitches go from *looking* random (not actually random, but internal structure isn't obvious from the outside) to looking more ethological... (cont'd)

Big thanks of course to @marksblumberg.bsky.social and everyone over at Iowa. And of course to @currentbiology.bsky.social for their continued support and love of weird animals!

REM-sleep twitching in adults and the maintenance of specialized sensorimotor systems Blumberg et al. present video evidence of twitching during adult sleep in a diverse sample of animals. Adult twitching appears to selectively involve appendages used for active sensing, mirroring thei...

tl;dr
Babies twitch to learn about their bodies; adults twitch to keep their most specialized skills sharp. So next time you see a twitching paw, know that you could be watching a brain-calibration in progress. As always, check out the full paper for the deep dive!

www.cell.com/current-biol...

This framework suggests twitching shifts from building the system in infants to maintaining and repairing it in adults. It might even explain how humans adapt after a stroke or limb loss. Twitching isn't a "glitch"—it's a neuroplasticity feature across the animal kingdom.

(pig video just for fun!)

But this might not just be limited to cortex and it might not be limited to mammals. Octopuses and cuttlefish— kings of complex appendages—show intense twitching of their arms and suckers during "active sleep." This suggests the intriguing possibility that if you want precision, you need twitches.

So why is this the case? We think it’s because active sensing is hard. To use a trunk or a whisker, the brain needs a perfect "internal model" to predict where that sensor is in space. We propose twitches act as biological test pulses to calibrate these models while the animal is asleep.

These body parts have something in common: Cortical Magnification. The brain dedicates a massive amount of space to these body parts and their sensors. For example, a platypus's bill takes up about 90% of its somatosensory cortex! And guess what? Platypus bills twitch like crazy during REM!

In adults, twitching isn’t random. It’s highly selective, and it seems like it happens most in the most specialized parts of animal bodies. The parts they use to probe the world – including:

Elephant trunks
Rat whiskers
Pig snouts
Raccoon hands

the list goes on and on!

For decades, people thought twitches were just accidental motor commands escaping the REM-sleep "paralysis" system. But if you look at who twitches and where on their body the twitches are, a much cooler pattern emerges.

REM-sleep twitching in adults and the maintenance of specialized sensorimotor systems Blumberg et al. present video evidence of twitching during adult sleep in a diverse sample of animals. Adult twitching appears to selectively involve appendages used for active sensing, mirroring thei...

New paper (with lots of cute animal videos!)

Ever watch your dog "run" while asleep and wonder what’s going on in their brain? In Current Bio we suggest that those twitches aren't just leaky dreams—they’re a vital maintenance system for the most precise movements

www.cell.com/current-biol...

I mean, if youre hallucination-worthy I think that means you're "recognized in the field" - so there's that.

Id put it in my tenure packet for sure

Of course, the full paper has more details (oscillations, state transitions, developmental timing, behavioral correlates, etc.) but these are the main highlights.

Happy to answer questions - and stay tuned for more about how sleep states shape early motor development!
12/end

The broader point:

Infant REM isn’t a tiny version of adult REM.

It’s a developing state that gains new structure over time - and this new structure likely changes the plasticity that’s possible during REM sleep.

“Sleep like a baby” turns out to be…a lot more complicated than we thought.
11/x

So why does this matter?

Because REM sleep isn’t just “sleep.”

It’s a major driver of early sensorimotor development.

By showing how tonic REM emerges (and becomes distinct from phasic REM) we can start to understand what neural representations are even possible in the developing brain.
10/x

So here’s the key result:

At P12, REM doesn’t have two substates - it’s one undifferentiated state. The two-stage structure of adult REM just doesn’t exist yet

That two-stage structure emerges during the 3rd postnatal week via a progression that looks a bit like the development of NREM sleep
9/x

Meanwhile, phasic REM keeps looking like the REM seen in early infancy, complete with:
• High neural activity
• Frequent twitches
• Elevated gamma

And by P20, phasic REM also has theta oscillations
8/x

Importantly, tonic REM doesn't just appear, fully formed. Instead, its features appear one-by-one

At P16, early REM bouts are mostly twitch free

At P20, tonic REM has slower theta oscillations than phasic REM

At P24, the alpha rhythm characteristic of tonic REM first appears
7/x