Back

Striatal ensembles specify and control granular forelimb actions

Rodrigues-Vaz, I.; Athalye, V. R.; Peterka, D. S.; Costa, R. M.

2025-12-08 neuroscience
10.64898/2025.12.03.692128 bioRxiv
Show abstract

The ability of the brain to control specific fine actions is crucial for survival. The striatum is a critical brain center for both movement and learning, and its dysfunction underlies numerous movement disorders1-9. Whereas activity in the striatum has been classically viewed as invigorating5,10-13 and reinforcing movements12-20, recent studies suggest that striatal activity encodes specific movements21-24. However, it is not known how granular this activity is, and if it indeed controls specific ongoing movements. We designed a task where mice performed two minimally-different forelimb actions, consisting of a push or pull isometric force on an immobile joystick, and imaged the activity of medium spiny neurons (MSNs) in the dorsolateral striatum using 2-photon microscopy. We observed that striatal activity encoded both the preparation and execution of specific actions, even when those actions were not reinforced. Furthermore, both populations of D1 and D2-MSNs - classically viewed as promoting versus inhibiting movement2,3,25 - equally encoded action identity. We developed a closed-loop system to model and stimulate action-specific neural ensembles deep in the brain, using holographic optogenetics through a GRIN lens. Stimulation of action-specific ensembles of both D1- and D2-MSNs increased the force of ongoing actions, but only when the ensemble stimulated was congruent with the ongoing action. These results reveal that specific ensembles of both D1- and D2-MSNs causally control specific ongoing actions, as granular as different muscle co-contractions of the same forelimb. Such granularity provides a mechanistic framework for understanding how striatal dysfunction can produce highly specific movement impairments in Huntingtons disease2,9 and dystonia2,6.

Matching journals

The top 2 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.