Abstract
The external globus pallidus (GPe) guides action selection. However, it remains elusive how diverse GPe cells engage in routine formation while learning action sequences in repetitive reward-seeking conditioning. Here, we investigated the Ca<sup>2+</sup> signals of GPe cells during routine formation. We found that the dynamics of astrocytes and prototypic neurons’ activities are associated with the refinement of the action sequence, which enables cost-effective reward-seeking.
Introduction
In the quest for rewards, animals transition from variable trial-and-error actions into efficient, predictable "routines." While the basal ganglia are known as the brain's "habit center," most research has focused on the striatum. The role of the external globus pallidus (GPe) has remained relatively elusive. Specifically, it was unknown how its diverse cell populations—including non-neuronal astrocytes and prototypic neurons—contribute to the formation of these routines. This study sought to understand if the GPe acts as a regulatory hub that filters out unnecessary actions to streamline behavior and save cognitive energy.
Methods
The researchers employed a combination of behavioral training and high-resolution imaging. Mice were trained on a "fixed ratio 1" (FR1) operant conditioning task: perform a nose-poke (NP) to receive a sucrose reward in a magazine (ME), then exit (MX) and repeat.
To monitor activity in vivo, the team used fiber photometry to track calcium (Ca<sup>2+</sup>) signals in two specific cell types: GFAP+ astrocytes (n=12) and PV+ prototypic neurons (n=5). A mathematical "Routine Index" was developed to quantify how closely a mouse's behavior matched the ideal, efficient cycle (NP → ME → MX → NP). Advanced statistical methods, including Tensor Component Analysis (TCA), were used to extract meaningful patterns from the multi-dimensional neural data across 662 behavioral cycles.
Key Results
The study found that as mice learned, their behavior converged from "noisy" sequences to a tight 3-block cycle. GPe cells reflected this refinement by showing higher activity during "routine" blocks compared to "non-routine" blocks (errors or repetitive actions).
Cell-Type Specificity and Error Attenuation
While both neurons and astrocytes track the routine, they do so differently. As a routine becomes established, the difference in activity between "correct" and "incorrect" actions becomes more pronounced in PV+ neurons, suggesting they "sharpen" the behavioral signal over time. In astrocytes, the signal associated with "errors"—such as revisiting the magazine without a new nose-poke—actually decreased as the mouse became more skilled, suggesting the GPe helps "dampen" the representation of unsuccessful strategies.
Selective Gain Control
The "Routine Index" significantly correlated with the activity levels of GPe cells. Specifically, the discriminability of PV+ neurons for certain transitions increased as the behavior became more routine (P < 0.05). The GPe appears to provide a "selective gain control," boosting the signals for the most efficient path to a reward while suppressing others.
Discussion and Future Directions
This research demonstrates that GPe astrocytes are active participants in task representation rather than mere support cells. However, the current study provides a correlation between GPe activity and routine formation rather than proving causality. Future work using closed-loop optogenetics will be necessary to determine if the GPe actively drives the routine. Additionally, investigating the long-term transition from a newly formed "skill" to a "hard-wired habit" remains an open question for the field.
Summary updated