Finally, we used this to probe synaptic plasticity. Pairing climbing fibre input before granule cell activation induced potentiation of IPSPs, lasting at least 40 min. Smaller IPSPs potentiated more, and neighbours co-potentiated — revealing circuit-level learning rules.

Because we could image full dendritic arbors across multiple cells, we saw rich circuit dynamics: precise synchrony of CF-evoked events, spatially diverse inhibition across branches, and global vs local patterns of dendritic activity.

With this, we captured high-fidelity recordings of spontaneous and sensory-evoked complex spikes, optogenetically evoked IPSPs (via feedforward inhibition) + even the occasional EPSP!

We expressed JEDI-2Psub in Purkinje cell dendrites and opsins in cerebellar granule cells. This enables us to both image voltage at postsynaptic dendrites and precisely activate presynaptic inputs with light.

At the heart of this is a new voltage indicator — JEDI-2Psub — from colleagues at the St.Pierre Lab. Optimised for subthreshold signals, it gives ~3.5× larger responses near rest compared to JEDI-2P. This makes it ideal for capturing synaptic potentials.

SUPER excited that my first ever *real* neuroscience paper is published today in Nat Comms!🥳

We combined two-photon voltage imaging ⚡️and optogenetics 🔦to probe synaptic plasticity in the cerebellum of awake behaving mice. 🧵

That’s the vision we’re exploring.

If you’re working on transformative, unconventional ideas to 1000x the scale of neurotech or have answers to the following questions — I’d love to hear from you!

📖 Full post: link.aria.org.uk/jacquessubst...

I think cardiology in the 1950s is a great case study, where treating heart block required open-heart surgery.

The leap wasn’t “better surgery.”
It was the transvenous lead — a minimally invasive access route that unlocked millions of pacemaker implants/year. ❤️⚡

DBS for PD is a great proof point: effective, approved, reimbursed… yet only a fraction of people eligible receive it.

Complex surgery imposes a series of stacked, multiplicative barriers that fundamentally limit scalability.

So what do we do?

I’ve been thinking a lot about how we scale neurotech.

Current approaches (e.g. DBS for PD) aren’t meeting the need. Emerging ones may fare even worse.

That’s why I’m scoping a new @aria-research.bsky.social programme to develop neurotech built for scale.

Check out my post +share feedback!👇

wait... fungi display action potentials?!🤯

"Two types of spiking activity are uncovered: high-frequency (period 2.6 min) and low-frequency (period 14 min) .."

is this well known in biology? like, if I actually studied it at school would I know this? so cool!

Love this. Definitely the cerebellum. Specifically how different cell types map onto specific components of an adaptive filter www.nature.com/articles/nrn...