But we couldn't have found that human result without the rat data to arbitrate between models, so that's reason 1 for needing the reverse translation. Reason 2 is that humans object if you slice their brains for c-fos to understand where those computational changes happen. 5/ - ThreadSky

alikwidge.bsky.social • 71 days ago

But we couldn't have found that human result without the rat data to arbitrate between models, so that's reason 1 for needing the reverse translation.
Reason 2 is that humans object if you slice their brains for c-fos to understand where those computational changes happen. 5/

Comments

alikwidge.bsky.social•71 days ago

Unsurprisingly, it's in PFC, but differentially -- bounds more in PL, drift rate more in IL. That's clinically interesting -- it suggests that to make DBS work, you need to modulate multiple PFC-thalamic-striatal sub-circuits. There's not one magic button -- it's a broad effect across PFC. 6/

alikwidge.bsky.social•71 days ago

That matches what we know clinically. For instance, think about @andreashorn.org 's sweet spots -- they are right in the middle of the capsule where the electric field can hit projections of multiple areas. We also saw widespread EEG changes in https://pubmed.ncbi.nlm.nih.gov/30948727/ . 7/

alikwidge.bsky.social•71 days ago

AND, if you happen to have a model for how you could optimize DBS for cognition (like in https://iopscience.iop.org/article/10.1088/1741-2552/acd0d5), now you have an animal system where you can test your optimizers and make sure they work before clinical trials. 8/

alikwidge.bsky.social•71 days ago

AND, you could start to opto dissect what individual components do, why the heck 130 Hz matters, and all the "why does DBS even work???" stuff that's bothered me for years.
So that's why rats, but because it's reverse translation, we can really hope what we learn will matter to people. 9/

alikwidge.bsky.social•71 days ago

It also highlights the power of the RDoCian approach we argued for in https://pubmed.ncbi.nlm.nih.gov/37415081/ : there aren't many clinical effects that look the same in humans and rats. This does. It's the closest I know to a "model of" in psychiatry. 10/

alikwidge.bsky.social•71 days ago

It also took 8 years. Adriano Reimer, the first author, started the first pilots in 2016. We had the model by 2018 when we moved, but it took till 2022 to have all of the cohorts and controls run (damn pandemic) and till 2022 to figure out how to do DDM in the presence of learning (Evan again). 11/

Comments

Posting Rules

Reply