Changes in dorsomedial striatum activity during expression of goal-directed vs. habit-like cue-induced cocaine seeking

Abstract

A preclinical model of cue exposure therapy, cue extinction, reduces cue-induced cocaine seeking that is goal-directed but not habit-like. Goal-directed and habitual behaviors differentially rely on the dorsomedial striatum (DMS) and dorsolateral striatum (DLS), but the effects of cue extinction on dorsal striatal responses to cue-induced drug seeking are unknown. We used fiber photometry in rats trained to self-administer cocaine paired with an audiovisual cue to examine how dorsal striatal intracellular calcium and extracellular dopamine activity differs between goal-directed and habit-like cue-induced cocaine seeking and how it is impacted by cue extinction. After minimal fixed-ratio training, rats showed enhanced DMS and DLS calcium responses to cue-reinforced compared to unreinforced lever presses. After rats were trained on goal-promoting fixed ratio schedules or habit-promoting second-order schedules of reinforcement, different patterns of dorsal striatal calcium and dopamine responses to cue-reinforced lever presses emerged. Rats trained on habit-promoting second-order schedules showed reduced DMS calcium responses and enhanced DLS dopamine responses to cue-reinforced lever presses. Cue extinction reduced calcium responses during subsequent drug seeking in the DMS, but not in the DLS. Therefore, cue extinction may reduce goal-directed behavior through its effects on the DMS, whereas habit-like behavior and the DLS are unaffected.

Keywords: Cocaine; Cue extinction; Dopamine; Dorsal striatum; Goal-directed; habit.

Fig. 1.

Schedules of reinforcement. In this example of a fixed-ratio (FR5) schedule of reinforcement, the fifth lever press results in cocaine infusion and a 20-s timeout period during with an audiovisual cue is presented (A). In this example of an FR5(FR2S) second-order (SO) schedule of reinforcement, a brief, 1-s audiovisual stimulus (S) is presented on an FR2 schedule, and upon the fifth completion of the FR2S schedule (after 10 total lever presses), cocaine infusion and the 20-s timeout with audiovisual cue presentation occurs.

Fig. 2.

FR- and SO-trained rats do not differ in daily cocaine self-administration or cue-induced drug-seeking after cue extinction. Schematic for fiber placement and virus expression in the DLS and DMS (A). Representative images of fiber placement and virus expression in the DMS and DLS (left) and at higher magnification in the DLS (right) with fluorescent channels shown individually and merged and fiber locations outlined in dotted white lines (B). For all rats, fiber placement and virus expression were evaluated via fluorescent microscopy, and green bars represent fibers appropriately placed with confirmed virus expression at the base of the fiber, while red bars indicate fibers excluded from analysis due to either fiber misplacement and/or lack of virus expression (C). Rats (n = 16) were trained to self-administer cocaine for 20 days on different schedules of reinforcement (FR or SO) before undergoing cue extinction and a subsequent cue-induced drug-seeking test, and after acquisition, photometry recordings occurred in 15-min drug-seeking tests that preceded daily self-administration, during cue extinction, and during the subsequent cue-induced drug-seeking test (D). During self-administration, there was a main effect of training day on the number of daily cocaine infusions, where infusions increased for both groups as training progressed (E). For lever presses during daily self-administration, there was a 3-way training day × training schedule × lever interaction (F). There was no difference between groups in the ratio of active lever presses during the post-cue extinction cue-induced drug-seeking test compared to the final day of self-administration (G). During photometry recording sessions, there was a main effect of test day and a test day × training schedule interaction on the number of schedule completions (H), and there was a main effect of training schedule on the number of active lever presses (I). Graphs show group means ± SEM and individual data points where possible, and gray vertical lines separate different reinforcement schedules.

Fig. 3.

After acquisition, dorsal striatal calcium responses are greater for cue-reinforced than unreinforced active lever presses. After acquisition and prior to splitting rats into FR- and SO-trained groups (n = 16), fiber photometry recordings occurred during 15-min drug-seeking tests prior to daily self-administration, during which some active lever presses had no consequence (unreinforced) and an active lever press that completed the FR3 schedule resulted in cue presentation (cue-reinforced) and timeout (levers retracted, houselight extinguished). For DLS (A) DMS (B) calcium, there was a main effect of cue reinforcement on peak z-score amplitude during the 1 s after lever press, but there was no effect of future training schedule or interaction. For DLS (C) and DMS (D) dopamine peak z-score amplitude, there was no effect of cue reinforcement, future training group, or interaction. Graphs show group means ± SEM and individual data points. Traces show overall average trace for each event for each future group aligned to behavioral events with SEM shown with shading and dashed vertical lines indicating time of lever press. Note that the average peak for each animal does not necessarily correspond visually with the peak of the average trace. This is because the peak for each trial can occur at any point during the 1 s after each behavioral event, whereas the average trace is the average of each animal’s average response. *p < 0.05; **p < 0.01.

Fig. 4.

FR- and SO-trained rats have different patterns of dorsal striatal calcium and dopamine activity during drug seeking. Rats were separated into FR-trained (n = 9; 5 male and 4 female) and SO-trained (n = 7; 4 male and 3 female) groups for the remaining 10 days of self-administration and trained on different schedules of reinforcement accordingly for the middle and late phases of training. Dorsal striatal calcium and dopamine responses to cue-reinforced and unreinforced lever presses were compared for each training schedule and phase of training. There was a main effect of cue reinforcement on DLS calcium peak amplitude, but no main effects of training schedule or phase of training or interactions (A). For DMS calcium peak amplitude, there was a main effect of cue reinforcement and a cue reinforcement × training schedule interaction, but no other main effects or interactions (B). There was a main effect of cue reinforcement and a cue reinforcement × training schedule interaction for DLS dopamine peak amplitude (C), but there were no main effects or interactions for DMS dopamine peak amplitude (D). Graphs show group means ± SEM and individual data points. Traces show overall average trace for each event for each group aligned to behavioral events with SEM shown with shading and dashed vertical lines indicating time of lever press. *p < 0.05; **p < 0.01; ***p < 0.001.

Fig. 5.

Cue extinction results in changes in DMS, but not DLS, calcium and dopamine activity during drug seeking. To examine the effects of cue extinction on dorsal striatal calcium and dopamine activity, peak amplitudes during the post-cue extinction drug-seeking test (post-ext) were compared to the late phase of training (pre-ext) (n = 13; exclusions explained in methods). There was a main effect of cue reinforcement on DLS calcium peak amplitude, but no effects of cue extinction or training schedule or interactions (A). For DMS calcium peak amplitude, there was a main effect of cue reinforcement, a main effect of cue extinction, and significant cue reinforcement × training schedule and training schedule × cue extinction interactions, but no other interactions (B). There was a main effect of cue reinforcement on DLS dopamine peak amplitude, with no other main effects or interactions (C). Finally, there was a main effect of cue extinction on DMS dopamine peak amplitude, but no other effects or interactions (D). Data from “pre-extinction” is from the late phase data in Fig. 4. Graphs show group means ± SEM and individual data points. Traces show overall average trace for each event for each group aligned to behavioral events with SEM shown with shading and dashed verical lines indicating time of lever press. *p < 0.05.