Goal-directed actions and habits in head-fixed mice

TL;DR

This study introduces a new method to study goal-directed actions and habits in mice, showing how the brain balances these behaviors.

Contribution

A novel head-fixed paradigm enables longitudinal neural interrogation of goal-directed and habitual behaviors in mice.

Findings

Mice with limited training reduced lever-pressing when outcomes were devalued, but overtrained mice did not.

Chemogenetic inhibition of the DLS in overtrained mice restored sensitivity to outcome devaluation and contingency reversal.

The paradigm allows tracking of neuronal ensembles during habit acquisition and expression.

Abstract

Behavioral control is fundamentally governed by a dynamic balance between flexible, goal-directed actions and efficient but highly automated habits. While the transition to habitual control can be adaptive, an imbalance favoring rigid habits over flexible control is thought to contribute significantly to the core pathology of many neuropsychiatric disorders. Despite their pervasiveness, the precise neural circuits that govern this critical balance remain poorly characterized. A major barrier to progress lies in the technical challenge of tracking single neurons in relevant circuits as goal-directed behavior becomes habitual.

Here, we introduce and validate a novel head-fixed instrumental learning paradigm in mice that enables the differentiation of goal-directed and habitual behavioral control. This model provides an unprecedented platform for high-resolution, longitudinal in vivo neural interrogation. To functionally validate the behavioral paradigm, we employed chemogenetic inhibition of the dorsolateral striatum (DLS) in limited-trained and overtrained mice.

Our findings demonstrate that mice rapidly acquire lever pressing, exhibiting robust active/inactive lever discrimination, and comparable initial learning between limited and overtrained cohorts. Crucially, while mice with limited training readily reduced their lever-press behavior when the resulting outcome was devalued or omitted, overtrained mice displayed characteristic insensitivity to outcome devaluation and omission. Furthermore, chemogenetic inhibition of the DLS in overtrained mice restored sensitivity to both devaluation and contingency reversal, blocking habit expression and preserving goal-directed control.

These results establish a paradigm for studying both goal-directed actions and habits in head-fixed mice and validate the DLS as a key neuronal substrate as previously demonstrated in freely moving models. This paradigm represents a vital methodological advancement that overcomes the technical barriers to longitudinal, cellular-level interrogation. By providing the optical stability necessary to track the same neuronal ensembles across acquisition and expression of habits, this platform enables investigations into the computational neural dynamics underlying goal-directed and habitual behaviors.

Figures