Mechanisms of Striatal Fast-Spiking Interneuron Plasticity in Habit Learning

Abstract

The behavioral transition from goal-directed to habitual responding is known to differentially rely on dorsomedial (DMS) & dorsolateral (DLS) striatal regions; with each region’s activity exerting opposing behavioral influences - DMS activity promoting goal-directed responding and DLS activity required for habitual responding. However, relatively little is known about how local plasticity is expressed within these two regions to mediate the goal to habit behavioral transition. Prior studies establish that fast-spiking interneurons (FSIs) in the dorsolateral striatum (DLS) show experience-dependent plasticity and are required for habitual responding (O'Hare, et al., 2017). FSIs are a sparse cell type within the striatal microcircuitry that exert powerful influence over striatal output in both medial and lateral regions. As of yet, their plasticity in habit has only been studied in DLS, and only in relation to goal-directed subjects, where FSIs in habitual mice show increased excitability relative to goal (O’Hare, et al., 2017).Here, I investigate the state transitions of FSIs in both DMS and DLS across learning goal-directed and habitual responding, beginning with the naïve state. Using lever press instrumental task training, acute brain slice electrophysiological recordings and cell-specific transcriptional profiling, I make three major observations that substantially revise working models and introduce new molecular mechanisms. Although current models support opposing roles for DMS and DLS in goal versus habitual behavior, I find that habit acquisition is accompanied by increased FSI excitability in both regions. By including naïve state physiology, I further reveal that rather than progressive gains in excitability through experience, FSIs instead transition to a temporary state of decreased excitability upon learning an instrumental task in a goal-directed manner. Physiology between FSIs in naïve and habitual subjects were highly similar. Despite similar physiological transitions of FSIs between DMS and DLS regions, I found that the underlying mechanisms are distinct. Electrophysiological, immunohistochemical, and transcriptional data support a model in which DLS FSIs decrease excitability through a transient down-regulation of Kv3 potassium channels and degradation of perineuronal net (PNN) structures; while DMS FSIs reduce excitability through transient increases in parvalbumin (PV) and Ca2+-activated potassium channel expression. My findings substantially revise working models in the field and shift focus away from a model of experience-dependent gains in interneuron plasticity to a model in which early learning imposes a temporary dampening of FSI excitability. In cortical circuits, transient decreases in FSI activity are a well-known gating mechanism to facilitate induction of excitatory synaptic plasticity. My findings suggest that the striatum may use parallel mechanisms, though on distinctly prolonged time scales.