Distinct spatially organized striatum-wide acetylcholine dynamics for the learning and extinction of Pavlovian associations

ABSTRACT

Striatal acetylcholine (ACh) signaling has been proposed to counteract reinforcement signals to promote extinction and behavioral flexibility. ACh dips to cues and rewards may open a temporal window for associative plasticity to occur, while elevations may promote extinction. Changes in multi-phasic striatal ACh signals have been widely reported during learning, but how and where signals are distributed to enable region-specific plasticity for the learning and degradation of cue-reward associations is poorly understood. We used array fiber photometry in mice to investigate how ACh release across the striatum evolves during learning and extinction of Pavlovian associations. We report a topographic organization of opposing changes in ACh release to cues, rewards, and consummatory actions across distinct striatum regions. We localized reward prediction error encoding in particular phases of the ACh dynamics to a specific region of the anterior dorsal striatum (aDS). Positive prediction errors in the aDS were expressed in ACh dips, and negative prediction errors in long latency ACh elevations. Silencing aDS ACh release impaired behavioral extinction, suggesting a role for ACh elevations in down-regulating cue-reward associations. Dopamine release in aDS dipped for cues during extinction, but glutamate input onto cholinergic interneurons did not change, suggesting an intrastriatal mechanism for the emergence of ACh elevations. Our large scale measurements indicate how and where ACh dynamics can shape region-specific plasticity to gate learning and promote extinction of Pavlovian associations.

FILE DESCRIPTIONS

This repository contains the following files:

• _README_Source_Data (.txt) - File with detailed information on content within each individual .zip folder that relates to a figure.
• Key Resources Table (.xlsx) - Table containing details on key resources (mouse lines, virus strains, software), and the persistent identifiers for protocols and MATLAB custom-written code used and generated in this study.
• Fig1 to 8 (.zip) - Individual folders contaning raw preprocessed and processed datasets for each Main Figures 1 to 8.
• Supp Fig1 to 7 (.zip) - Individual folders contaning raw preprocessed and processed datasets for each Supplementary Figures 1 to 7.