Functional heterogeneity within the rodent lateral orbitofrontal cortex dissociates outcome devaluation and reversal learning deficits

Abstract

The orbitofrontal cortex (OFC) is critical for updating reward-directed behaviours flexibly when outcomes are devalued or when task contingencies are reversed. Failure to update behaviour in outcome devaluation and reversal learning procedures are considered canonical deficits following OFC lesions in non-human primates and rodents. We examined the generality of these findings in rodents using lesions of the rodent lateral OFC (LO) in instrumental action-outcome and Pavlovian cue-outcome devaluation procedures. LO lesions disrupted outcome devaluation in Pavlovian but not instrumental procedures. Furthermore, although both anterior and posterior LO lesions disrupted Pavlovian outcome devaluation, only posterior LO lesions were found to disrupt reversal learning. Posterior but not anterior LO lesions were also found to disrupt the attribution of motivational value to Pavlovian cues in sign-tracking. These novel dissociable task- and subregion-specific effects suggest a way to reconcile contradictory findings between rodent and non-human primate OFC research.

Keywords: Pavlovian learning; autoshaping; neuroscience; orbitofrontal cortex; outcome devaluation; outcome identity; rat; reversal learning.

Figure 1.. The effects of excitotoxic OFC lesions on instrumental devaluation by taste aversion.

(A) Representative OFC lesion damage in the Non-Devalued (left) and Devalued (right) lesion groups. Semi-transparent grey patches represent lesion damage in a single subject, and darker areas represent overlapping damage across multiple subjects. Coronal sections are identified in mm relative to bregma (Paxinos and Watson, 1997). (B) Rate of lever pressing during 3 days of instrumental acquisition. (C) Mean reward consumption during taste aversion learning, consumption of rewards paired with LiCl induced nausea decreased across injection pairings (Left), whereas consumption of rewards paired with saline injections increased across injection pairings. (D) Total lever presses during the 10 mins devaluation test in extinction. Within-session responding presented in Figure 1—figure supplement 1. (E) Total lever presses during the 20 mins re-acquisition test with rewards delivered instrumentally. Error bars depict + SEM. (*) Symbol denotes statistical significance of simple or main effects following a significant interaction.

Figure 1—figure supplement 1.. The effects of excitotoxic OFC lesions on instrumental devaluation by taste aversion.

Within-session responding during test session, presented in 1 min time bins. Responding in all conditions was minimal during minute 10 and has been excluded. Error bars depict + SEM. There were significant devaluation effects in both lesion conditions, however they were expressed differently over time. Specifically, the devaluation effect was more persistent across time in the Lesion groups than the Sham groups. A mixed Lesion (sham, lesion) x Devaluation (devalued, non-devalued) x Bin (9 bins of 1 min) ANOVA revealed a main effect of Devaluation (F_{(1, 28)}=5.03, p=0.03), Bin (F_{(8, 224)}=16.74, p<0.001) and Lesion x Devaluation x Bin interaction (F_{(8, 224)}=2.53, p=0.01) (All remaining Fs <1.47, p>0.17). Follow up Devaluation x Bin ANOVAs for each lesion group revealed a significant main effect of Devaluation (F_{(1, 14)}=4.81, p=0.046) and Bin (F_{(8, 112)}=12.05, p<0.001) but no Devaluation x Bin interaction (F_{(8, 112)}=1.32, p=0.24) for the Lesion group. In contrast, Sham group showed a significant main effect of Bin (F_{(8, 112)}=6.98, p<0.001) and Devaluation x Bin interaction (F_{(8, 112)}=2.66, p=0.01, and a significant Devaluation x Bin linear trend interaction F_{(1, 14)}=6.97, p=0.02), but no significant main effect of Devaluation (F_{(1, 14)}=0.79, p=0.39). Bonferroni corrected simple effects revealed significantly higher responding in the Non-Devalued Sham group than the Devalued Sham Group during the 7^th minute (F_{(1, 14)}=12.19, p=0.004, all remaining minutes failed to reach significance, Fs <2.73, p>0.12). Linear trend contrasts revealed significant negative linear decreases in lever responding over time Bins in the Sham Devalued group (F_{(1, 7)}=126.03, p<0.001), but not the Sham Non-Devalued group (F_{(1, 7)}=3.36, p=0.109). (*) Symbol denotes statistical significance of simple or main effects following a significant interaction.

Figure 2.. The effects of subregion specific OFC lesions on Pavlovian devaluation by taste aversion.

(A) Representative OFC lesion damage in the anterior (left) and posterior (right) LO lesion groups; additional histology presented in Figure 2—figure supplement 1. Semi-transparent grey patches represent lesion damage in a single subject, and darker areas represent overlapping damage across multiple subjects. Coronal sections are identified in mm relative to bregma (Paxinos and Watson, 1997). (B) Quantification of percent bilateral OFC damage in anterior and posterior lesion groups at each coronal plane, in mm relative to bregma. (C) Rate of acquisition to the Pavlovian CSs in blocks of 3 days. Response rates presented as duration of magazine activity during the CS minus activity in the PreCS period. (D) The acquisition of a specific taste aversion following pairings of one outcome with LiCl injections (Devalued) or saline injections (Non-Devalued). The mean weight of each outcome consumed prior to each injection pairing is plotted. (E) An additional pairing of each outcome with LiCl or saline injections was conducted in the experimental chambers following non-contingent delivery of reward into magazine. Data presented as total duration of magazine activity in the session. This allowed for a measure of the transfer of the taste aversion to the testing context. (F) Magazine responding (CS – PreCS) to the CSs associated with the devalued and non-devalued outcomes, presented in extinction. (G) An outcome specific reinstatement test in which responding to the CSs was assessed after brief exposure to its associated outcome. Error bars depict + SEM. (*) Symbol denotes statistical significance of simple or main effects following a significant interaction. Effect of OFC lesions on locomotor activity presented in Figure 2—figure supplement 2.

Figure 2—figure supplement 1.. Representative OFC lesion damage across the anterior-posterior plane within the same rat from the sham (left), anterior (middle), and posterior lesion groups.

Coronal sections are identified in mm relative to bregma (Paxinos and Watson, 1997).

Figure 2—figure supplement 2.. OFC lesions do not affect general locomotor activity.

Mean number of beam breaks before (left) and after (right) surgery, in blocks of 10 mins. Error bars depict + SEM. A mixed Group x Surgery (Pre, Post) x Block (6 blocks of 10 mins) ANOVA revealed a significant main effect of Surgery (F_{(1, 41)}=10.18, p=0.003), Block (F_{(5, 205)}=659.29, p<0.001) and a Surgery x Block interaction (F_{(5, 205)}=4.88, p<0.001, all remaining effects F < 1.71, p>0.08). The significant effect of surgery was a reduction in locomotor activity post-surgery which is likely to be a result of habituation to the locomotor chamber.

Figure 3.. The effects of subregion specific OFC lesions sign-tracking behaviour and reversal learning.

Lever (A) and magazine (B) CS responding during 12 days of acquisition (left), and reversal (right) of a rewarded (CS+) and non-rewarded (CS-) lever cue. Response competition during acquisition (C) and reversal (D) of the CS+. Lever response bias calculated as the difference between standardised lever and magazine responding so that positive scores represent greater lever bias and negative scores represent greater magazine response bias. Error bars depict + SEM. (*) Symbol denotes statistical significance of simple or main effects following a significant interaction.

Appendix 1—figure 1.. The effects of excitotoxic OFC lesions on specific Pavlovian-to-instrumental transfer (PIT).

(A) Representative OFC lesion damage in the lesion group. Semi-transparent grey patches represent lesion damage in a single subject, and darker areas represent overlapping damage across multiple subjects. Coronal sections are identified in mm relative to bregma (Paxinos and Watson, 1997). Rate of lever pressing (B) and magazine entry behaviour (C) during extinction of the instrumental response prior to PIT testing. Instrumental lever pressing (D) and magazine entry behaviour during the specific PIT test. Responding plotted as the mean response rate per minute during each cue minus the preceding baseline no-cue period. Same and different conditions indicate whether the Pavlovian CS predicted the same or different liquid reinforcer to the instrumental response, the general condition indicates responding during the CS that predicted pellets which were never an instrumental reinforcer. Error bars depict + SEM. (*) Symbol denotes statistical significance of simple or main effects following a significant interaction.