Multi-fiber photometry reveals 3-dimensional gradients in dopamine trajectory error encoding across the striatum
OA Version
Citation
Abstract
Striatal dopamine is a powerful modulator of goal-directed behavior. Recent work has sought to unify diverse dopamine release dynamics under a common framework through modifications and extensions of the prevailing prediction error algorithm. However, due to fragmented data and experimental limitations, the organizational principles of dopamine release across the striatum and its underlying computations remain elusive, limiting our understanding of its function during the rich complexities of everyday animal behavior. Here, we develop a multi-optical fiber array for simultaneous large-scale, high-resolution measurements and manipulations of dopamine release across the three-dimensional volume of the striatum to address the technological limitations of previous studies. To address behavioral limitations, spatial navigation has proven to be a fruitful paradigm to study the underlying principles of cognitive maps and episodic memory in the hippocampus. The striatum serves a complementary role in navigating towards a goal, supporting landmark-based egocentric strategies, but the underlying spatiotemporal dynamics and computational principles supporting this function are less well understood. Here, we combine our multi-optical fiber array with a task which mimics aspects of landmark-based navigation to uncover anatomical gradients in cue-evoked dopamine release dynamics containing information about cue identity and ongoing movement necessary to learn or execute adjustments in ongoing continuous behavior. Across the anterior striatum cue-evoked dopamine release encodes bi-directional ‘trajectory errors’ reflecting relationships between ongoing speed and direction of locomotion and visual flow relative to optimal goal trajectories. Trajectory error encoding is multiplexed with a trajectory independent positive cue response which varies along a near-orthogonal spatial axis. These two components combine to produce spatially varying expressions of positive and negative trajectory errors which evolve over different time courses in the medial and lateral striatum, enabling region specific contributions to learning. Dopamine trajectory error signaling and task performance were reproduced in a reinforcement learning model incorporating a conjunctive state space representation, extending the traditional reward prediction error framework to incorporate the speed and direction of ongoing movement relative to goals. Our results establish overlapping anatomical gradients in dopamine signals positioned to guide distinct aspects of landmark-based navigation and reveal organizational principles of dopamine dynamics across the striatum.
Description
2025
License
Attribution 4.0 International