Rats were first surgically infused with inhibitory archaerhodopsin T (ArchT) or green fluorescent protein only (GFP) and implanted bilaterally with optic fibers into the PRC (Figure 1a). Rats were then trained on a novel task in which they first acquired the incentive values of three pairs of objects (appetitive [sucrose], aversive [electric shock], and neutral [no outcome] pairs) in a customized runway apparatus comprising four successive compartments: a start box; an object box containing two objects; a neutral box in which the rat was held temporarily; and finally, a goal box, in which the associated outcomes were delivered during training (Figure 1b). Objects were selected to be visually and textually distinct from one another since PRC is implicated in the discrimination of objects with overlapping features (Murray et al., 2007).

Figure 1

Download asset Open asset

Acquisition of object valences was assessed after four (test 1) and eight (test 2) conditioning sessions, without laser treatment. PRC groups acquired the object-outcome associations successfully by test 2, with rats spending the most time in the goal box after exposure to the appetitive object pair (p<0.001), and the least time after aversive object pair exposure (p<0.001), compared to goal box time in neutral trials (valence: F(2,40)=396.99, p<0.001, η_p² = 0.95). Rats also exhibited a significant reduction in goal box time during aversive test session between the two acquisition tests (valence × test: F(2,40)=16.86, p<0.001, η_p² = 0.46) (Figure 1c). The pattern of valence acquisition was comparable between ArchT and GFP animals (construct: F(1,20)=1.51, p=0.23, η_p² = 0.07; valence × construct: F(2,40)=1.1, p=0.34, η_p² = 0.05; test × construct: F(2,40)=0.2, p=0.66, η_p² = 0.01).

Animals were then administered a series of tests in which recombinations of the learned objects were presented to elicit a high (appetitive-aversive) or low (appetitive-neutral; aversive-neutral) level of motivational conflict, or no conflict (neutral object pair was presented) (Figure 2a). When animals were exposed to a high conflict object pair, optogenetic inhibition of PRC (laser on) significantly increased time spent in the goal box compared with trials completed without inhibition (laser off), and compared to GFP control animals with laser on and off (laser × construct: F(1,20)=40.73, p<0.001, η_p² = 0.67; all post hoc: p<0.001) (Figure 2b), indicative of an increase in approach behavior under motivational conflict. An analysis of the ‘difference score’ between the two conflict test sessions (i.e., laser on–laser off) also indicated that laser-treated ArchT animals spent more time in the goal box than GFP animals (ArchT: 29.72 (M)±14.23 (SD), GFP: –1.41±8.37, t(20) = 6.38, p<0.001). Furthermore, PRC inhibition led to a significant decrease in the number of retreats from the goal box (laser × construct: F(1,20)=24.08, p<0.001, η_p² = 0.55), although there was no effect on the number of goal box entries (laser × construct: F(1,20) = 3.12, p=0.09, η_p² = 0.14) (Figure 2c) during the choice period. PRC inhibition also had no impact on the latency to enter (LTE) each of the object, neutral and goal boxes (laser × construct: F(1,20) = 2.24, p=0.15, η_p² = 0.1; laser × construct × box: F(2,40)=0.47, p=0.63, η_p² = 0.02), nor did it have an effect on the exploration of appetitive and aversive objects (laser × construct: F(1,20) = 3.10, p=0.09, η_p² = 0.13; laser × construct × object: F(1,20)=0.15, p=0.70, η_p² = 0.007) (Figure 2—figure supplement 1a–b). Collectively, these results indicate that when faced with high motivational conflict, PRC-inhibited rats entered the goal box as readily as control animals but stayed longer and retreated less, indicative of a potentiated approach bias under conflict. This effect was not observed in the neutral object test: PRC inhibition had no effect on the time spent in the goal box after exploring neutral objects (laser × construct: F(1,18)=1.02, p=0.33, η_p² = 0.05) (Figure 2d).

Figure 2 with 1 supplement see all

Download asset Open asset

When PRC-inhibited animals were exposed to ‘low conflict’ object pairs (appetitive-neutral and aversive-neutral), their time in the goal box increased (laser × construct: F(1,16)=9.37, p=0.007, η_p² = 0.37) (Figure 2e), but there was no corresponding increase in the number of entries into (laser × construct: F(1,16)=0.47, p=0.50, η_p² = 0.02) or retreats from (laser × construct: F(1,16)=0.54, p=0.47, η_p² = 0.04) the goal box during the choice period (Figure 2f–g). An analysis of the difference score (laser on–laser off) for the two low conflict tests revealed that laser treatment increased goal box time for ArchT animals compared with GFP, when faced with both App-Neu (ArchT: 22.13 (M)±21.61 (SD), GFP: 1.81±7.49, Mann-Whitney U=15, p=0.027, two-tailed) and Av-Neu (ArchT: 13.32±8.45, GFP: 1.07±10.77, t(16) = 2.71, p=0.016) pairings. PRC inhibition had no impact on the LTE into any of the runway boxes (laser × construct: F(1,16) = 0.24, p=0.63, η_p² = 0.02; laser × construct × box: F(2,32)=0.21, p=0.81, η_p² = 0.01), or valence of the object pairing (valence: F(1,16) = 2.38, p=0.15, η_p² = 0.17) (Figure 2—figure supplement 1c–d). There was no effect of PRC inhibition on the exploration of the App-Neu and Av-Neu object pairings (laser × construct: F(1,16)=1.14, p=0.30, η_p² = 0.07; laser × construct × object: F(1,16)=0.02, p=0.88, η_p² = 0.001), with comparable object exploration between valence pairings (valence: F(1,16) = 1.64, p=0.22, η_p² = 0.09; valence × construct: F(1,16)=3.73, p=0.07, η_p² = 0.19) (Figure 2—figure supplement 1e–f).

Notably, PRC inhibition did not impair the ability to discriminate appetitive and aversive valences (valence: F(1,16)=229.64, p<0.001, η_p² = 0.94; laser × valence × construct: F(1,16)=1.14, p=0.3, η_p² = 0.07), with all animals spending significantly more time in the goal box for appetitive-neutral pairings than aversive-neutral (p<0.001, Figure 2e). Furthermore, all PRC rats made fewer entries (p=0.039) and exhibited more retreat behavior (p<0.0001) when presented with an aversive-neutral pairing compared with an appetitive-neutral pairing (Figure 2f–g), indicating that valence recall was intact in PRC-inhibited rats.

To rule out the possibility that the observed effects of PRC inhibition were driven by a failure to respond to novel recombinations of object pairings rather than conflict processing, we repeated the runway task in which animals were trained to associate a new set of four object pairs with either appetitive or aversive outcomes (two pairs each) (Figure 3a). Animals acquired the cue-outcome associations successfully by the first test, and spent significantly more time in the goal box after exposure to the appetitive object pairs compared with the aversive pairs, with comparable acquisition behavior between sets of objects pairs within each valence (valence: F(1,14)=259.27, p<0.001, η_p² = 0.95; pairing: F(1,14)=1.67, p=0.22, η_p² = 0.11; valence × pairing: F(1,14)=0.27, p=0.61, η_p² = 0.02; Figure 3b). The pattern of object cue acquisition was comparable between ArchT and GFP animals (construct: F(1,14)=0.16, p=0.69, η_p² = 0.01; pairing × construct: F(1,14)=0.04, p=0.86, η_p² = 0.003).

Figure 3 with 1 supplement see all

Download asset Open asset

Rats were then administered a within-valence ‘no conflict’ recombination test (appetitive-appetitive; aversive-aversive). PRC inhibition did not lead to significant changes in the time spent in the goal box after exposure to novel appetitive or aversive object pairs (laser: F(1,14)=0.10, p=0.76, η_p² = 0.007; laser × construct: F(1,14)=0.01, p=0.92, η_p² = 0.001) (Figure 3c). An analysis of the difference score (laser on–laser off) for the two respective no conflict tests revealed that laser treatment had no effect on goal box time when animals encountered either App-App (ArchT: 0.65 (M)±9.27 (SD), GFP: –1.81±16.33, t(14)=0.37, p=0.72) or Av-Av (ArchT: –1.97 (M)±19.80 (SD), GFP: –0.72±12.54, t(14) = –0.15, p<0.88) pairings. Furthermore, both ArchT and GFP animals readily discriminated between valences during this test (valence: F(1,14) = 371.95, p<0.001, η_p² = 0.96; valence × construct: F(1,14)=0.05, p=0.82, η_p² = 0.004). PRC inhibition also had no effect on the number of entries into (laser × construct: F(1,14)=0.06, p=0.82, η_p² = 0.004) nor retreats from the goal box (laser × construct: F(1,14)=0.15, p=0.71, η_p² = 0.01), and furthermore, all PRC-inhibited rats made fewer entries (p<0.001) and exhibited more retreat behavior (p<0.0001) when presented with an aversive pairing compared with an appetitive pairing (Figure 3d–e). PRC inhibition also had no impact on the LTE data into any of the runway boxes (laser × construct: F(1,14)=1.7, p=0.21, η_p² = 0.11; laser × construct × box: F(2,28)=0.30, p=0.74, η_p² = 0.02) regardless of the valence of the object pairing (valence: F(1,14)=2.99, p=0.11, η_p² = 0.18); nor did it have an effect on the exploration of appetitive and aversive object pairs (laser × construct: F(1,14)=1.07, p=0.32, η_p² = 0.07; laser × construct × object: F(1,14)=0.07, p=0.80, η_p² = 0.005), with comparable object exploration between valence pairings (valence: F(1,14)=3.67, p=0.08, η_p² = 0.21; valence × construct: F(1,14)=0.39, p=0.55, η_p² = 0.03; Figure 3—figure supplement 1a–d).

Finally, to investigate whether animals exhibiting approach bias spent an increased amount of time in the vicinity of the sucrose dispenser in expectation of reward (i.e., reward location approach), we analyzed the amount of time and proportion of total goal box time that ArchT animals spent in a demarcated ‘dispenser zone’ (final 18 cm or 1/3 of goal box) during the high (App/Av), low (App/Neu), and no conflict (App/App) test sessions. When faced with high conflict, PRC inhibition led to animals spending significantly more time by the sucrose dispenser compared to high conflict test sessions without optogenetic inhibition (p<0.0001), whereas an increase in time spent by the dispenser elicited by low conflict stimuli approached significance (p=0.052; laser: F(1,18)=24.34, p<0.0001, η_p² = 0.57; laser × conflict: F(2,32)=5.43, p=0.0093, η_p² = 0.25) (Figure 4a–b). As expected, in the absence of PRC inhibition, animals spent more time by the dispenser during no conflict trials compared to high conflict (p=0.002) and low conflict (p=0.03) trials.

Figure 4

Download asset Open asset

When considering the time spent by the dispenser as a proportion of the total time in the goal box, PRC-inhibited rats spent a higher proportion of time at the dispenser during high conflict trials compared with low conflict (p=0.0008) and no conflict (p=0.0012) trials, as well as high conflict trials completed with the laser off (p=0.015; laser: F(1,18)=7.80, p=0.012, η_p² = 0.30, conflict: F(1.79,28.65)=11.19, p=0.00038, η_p² = 0.41, laser × conflict: F(2,32)=3.34, p=0.048, η_p² = 0.17) (Figure 4c). Laser off animals also spent proportionally more time by the dispenser during no conflict trials compared with low conflict trials (p=0.017), with no difference between no conflict and high conflict trials (p=0.89). To further investigate the impact of PRC inhibition on reward approach, we also analyzed the LTE into the object, neutral and goal boxes across the two levels of conflict that included an appetitive object (e.g., App+Av; App+Neu). Animals were significantly faster to enter all compartments under ‘laser on’ conditions (laser: F(1,16)=9.67, p=0.007, η_p² = 0.38), which was likely driven by the PRC inhibition group data (laser × construct: F(1,16)=3.92, p=0.065, η_p² = 0.20; laser × construct × box × conflict: F(2,32)=0.2, p=0.82; Figure 2—figure supplement 1 a,c). However, all rats exhibited longer latencies to enter the boxes during the high conflict condition, compared with the low conflict test (conflict: F(1,16)=10.47, p<0.005, η_p² = 0.40). Object exploration was also comparable between high vs. low conflict conditions, with PRC inhibition having no impact on the durations of object exploration (conflict: F(1,16)=0.003, p=0.96, η_p² = 0.0001; laser: F(1,16)=0.034, p=0.86, η_p² = 0.002; laser × construct: F(1,16)=0.056, p=0.82, η_p² = 0.003).

Collectively, these findings indicate that the PRC inhibition-induced increase in approach behavior during the high and low conflict tests arises from alterations in conflict processing, rather than impairments in valence retrieval or novel stimulus recombination processing. Furthermore, in the absence of the normal PRC functioning, animals spent more time by the sucrose dispenser and traversed through the runway more quickly, reflecting a greater anticipation for a reward outcome.

To rule out encoding impairments leading to the observed PRC inhibition effects, a separate control cohort of animals (N = 16; ArchT = 8; GFP = 8) was trained on the runway paradigm, wherein the optogenetic inhibition was applied when the animal entered the neutral box (i.e., after object exploration) and prior to the choice period (Figure 5a).

Figure 5 with 1 supplement see all

Download asset Open asset

All rats acquired the object-outcome associations successfully by test 2, with rats spending the most time in the goal box after exposure to the appetitive object pair (p<0.001), and the least time after aversive object pair exposure (p<0.001), compared to goal box time in neutral trials (valence: F(2,28)=231.3, p<0.001, η_p² = 0.94; Figure 5b); rats exhibited a significant increase in goal box time during appetitive trials and decrease in aversive trials between the two acquisition tests (valence × test: F(2,28)=54.91, p<0.001, η_p² = 0.80). The pattern of valence acquisition was comparable between ArchT and GFP animals (construct: F(1,14)=1.09, p=0.3, η_p² = 0.07; valence × construct: F(2, 28)=1.53, p=0.23, η_p² = 0.10; test × construct: F(1, 14)=1.64, p=0.22, η_p² = 0.10).

When animals were exposed to a high conflict object pair, optogenetic inhibition of PRC prior to goal box access significantly increased time spent in the goal box compared to trials without inhibition, and compared to GFP control animals (laser × construct: F(1,14)=81.3, p<0.001, η_p²=0.85; all post hoc: p<0.001) (Figure 5c). Turning the laser on following neutral box entry did not impact time spent in the goal box when compared with animals that received whole-session laser application (laser timing: F(1,34)=1.26, p=0.27, η_p²=0.04; laser timing × laser: F(1,34)=0.06, p=0.81, ηp2=0.002; laser timing × construct: F(1, 34)=0.12, p=0.7, η_p²=0.005), suggesting that PRC mediates object-based motivational conflict rather than encoding or retrieving representations of highly conflicting objects. The difference score of the two conflict test sessions (laser on–laser off) revealed that laser treatment increased goal box time for ArchT animals only (ArchT: 32.13±5.95, GFP: –2.22±8.98, t(14) = 9.02, p<0.001).

PRC inhibition during the choice period led to a significant decrease in the number of retreats from the goal box (laser × construct: F(1,14) = 38.11, p<0.001, η_p² = 0.73), while having no effect on the number of goal box entries (laser × construct: F(1,14)=0.88, p=0.37, η_p² = 0.06) (Figure 5d), the LTE each of the object, neutral, and goal boxes (laser × construct: F(1,14)=0.17, p=0.69, η_p² = 0.01; laser × construct × box: F(2,28)=0.26, p=0.77, η_p² = 0.02), or exploration of appetitive and aversive objects (laser × construct: F(1,14) = 0.007, p=0.93, η_p² = 0.001; laser × construct × object: F(1,14)=0.05, p=82, η_p² = 0.004) (Figure 5—figure supplement 1a–b). PRC inhibition had no effect on the time spent in the goal box during the neutral object test (laser × construct: F(1,16)=0.87, p=0.37, η_p² = 0.07; Figure 5e).

After a refresher conditioning session, rats underwent a low conflict recombination test. Similar to rats that received whole-session PRC inhibition, inhibiting PRC during the choice period increased the time spent in the goal box for both App-Neu and Av-Neu object pairs (laser × construct: F(1,14)=15.77, p<0.005, η_p²=0.53) (Figure 5f). The laser treatment difference score analysis showed that ArchT animals spent more time in the goal box when faced with App-Neu (ArchT: 22.58±11.72, GFP: –0.34±9.22, t(14) = 4.35, p=0.001), but not Av-Neu (ArchT: 13.17±15.35, GFP: 0.40±16.71, t(14) = 1.59, p=0.13) pairings. PRC inhibition had no effect on the number of entries or retreats from the goal box during the choice period (entries: laser × construct: F(1,14)=0.14, p=0.71, η_p²=0.01; retreats: laser × construct: F(1,14)=0.24, p=0.64, η_p²=0.05) (Figure 5g–h). PRC inhibition also had no impact on the LTE for any of the runway boxes (laser × construct: F(1,14)=1.91, p=0.19, η_p²=0.10; laser × construct × box: F(2,28)=0.11, p=0.11, η_p²=0.07), or the exploration of the App-Neu and Av-Neu object pairings (laser × construct: F(1,14)=1.42, p=0.25, η_p²=0.09; laser × construct × object: F(1,14)=0.1, p=0.92, η_p²=0.001), with comparable object exploration between the valanced pairs (valence: F(1,14)=2.36, p=0.15, η_p²=0.14; valence × construct: F(1,14) = 0.001, p=0.97, η_p²<0.0001) (Figure 5—figure supplement 1c–f).

Notably, PRC inhibition during the choice period did not impair the ability to discriminate appetitive and aversive valences (valence: F(1,14)=253.86, p<0.001, η_p² = 0.95; laser × valence × construct: F(1,14)=1.0, p=0.34, η_p² = 0.07), with all animals spending significantly more time in the goal box for appetitive-neutral pairings than aversive-neutral (p<0.001, Figure 5f). Furthermore, all PRC rats made fewer entries (p=0.001) and exhibited more retreat behavior (p<0.001) when presented with an aversive-neutral pairing compared with an appetitive-neutral pairing (Figure 5g–h). Thus, valence recall was intact in PRC-inhibited rats.

Finally, PRC-inhibited ArchT animals spent significantly more time by the sucrose dispenser after encountering both high conflict (p<0.001) and low conflict (p<0.01) stimuli, compared to test sessions without optogenetic inhibition (laser: F(1,7)=130.11, p<0.0001, η_p² = 0.95; laser × conflict: F(1,7)=10.70, p=0.014, η_p² = 0.61) (Figure 5i). PRC-inhibited rats also spent a higher proportion of time by the dispenser during high conflict trials compared with low conflict (p<0.0001) trials, as well as high conflict trials completed with the laser off (p=0.001; laser: F(1, 7)=34.10, p=0.001, η_p² = 0.83; conflict: F(1,7)=24.41, p=0.002, η_p² = 0.78; laser × conflict: F(1,7)=5.95, p=0.045, η_p² = 0.46) (Figure 5j). An analysis of the LTE data for each of the boxes between high and low conflict sessions revealed that PRC inhibition did not change LTE behavior (laser: F(1,14)=0.803, p=0.39, η_p² = 0.05; laser × construct: F(1,14)=1.48, p=0.24, η_p² = 0.1), with all rats exhibiting overall longer latencies to enter all boxes during the high conflict test compared with the low conflict test (p=0.043; conflict: F(1,14)=9.99, p=0.007, η_p² = 0.42; construct: F(1,14)=0.17, p=0.69, η_p² = 0.01; conflict × construct: F(1,14)=0.28, p=0.60, η_p² = 0.02; laser × construct × box × conflict: F(2,28)=0.16, p=0.85).

To investigate whether the observed increase in goal box preference time exhibited by PRC-inhibited rats was specific to the conditions elicited by the runway task, the same PRC rats completed another novel behavioral paradigm conducted in a modified two-way active avoidance ‘shuttle box’ apparatus (Figure 6a–c). In the first phase of the paradigm, appetitive or aversive object cue pairs used in the ‘no conflict’ recombination test in the runway task were placed in opposite ends of a two-compartment apparatus, behind transparent barriers. In each trial, animals were allowed to visually sample the objects for 1 min, after which the transparent barrier was raised to allow animals access to the associated outcome (sucrose or shock). At the same time, a central guillotine door dividing the two compartments was raised to give rats the opportunity to shuttle into the opposite compartment to escape the shock outcome associated with the aversive object cue pair. Acquisition of escape behavior was assessed after four (test 1) and eight (test 2) conditioning sessions, without laser treatment. PRC rats demonstrated the expected behavior by test 2, with rats exhibiting significantly shorter escape latencies after exposure to the aversive object pair compared with exposure to the appetitive object pair (valence: F(1,13)=45.97, p<0.0001, η_p² = 0.78; valence × test: F(1,13)=16.17, p=0.001, η_p² = 0.55) (Figure 6d). The pattern of escape acquisition was comparable between ArchT and GFP animals (construct: F(1,13)=4.55, p=0.052, η_p² = 0.26; valence × construct: F(1,13)=4.22, p=0.06, η_p² = 0.25; test × construct: F(1,13)=2.69, p=0.13, η_p² = 0.17). PRC rats also spent more time in the outcome-associated area (stay behavior) when exposed to the appetitive object pair compared with the aversive object pair (valence: F(1,13)=456.04, p<0.0001, η_p² = 0.97; valence × test: F(1,13)=21.18, p<0.001, η_p² = 0.62) (Figure 6e). This pattern of stay behavior was also comparable between ArchT and GFP animals (construct: F(1,13)=3.57, p=0.082, η_p² = 0.22; valence × construct: F(1,13)=8.17, p=0.013, η_p² = 0.39; test × construct: F(1,13)=1.65, p=0.22, η_p² = 0.11).

Figure 6

Download asset Open asset

When animals were next exposed to a high conflict pairing, optogenetic inhibition of PRC (laser on) significantly prolonged escape latencies compared with trials completed without inhibition (laser off; p<0.001), and compared to GFP control animals (p<0.005) with laser on and off (laser × construct: F(1,13)=18.97, p=0.001, η_p² = 0.59) (Figure 6f), indicative of a decrease in avoidance behavior under motivational conflict. Furthermore, despite making a comparable number of re-entries into the outcome-associated area during the test session (laser × construct: F(1,13)=0.99, p=0.34, η_p² = 0.07), PRC-inhibited animals spent significantly longer in the outcome-associated area containing the conflict object pair compared with trials completed without inhibition (laser off; p<0.001), and compared to GFP control animals (p<0.005) with laser on and off (laser × construct: F(1,13)=12.32, p=0.004, η_p² = 0.49) (Figure 6g). An analysis of the difference score of the two conflict test sessions (laser on–laser off) revealed that laser treatment both decreased escape latency (ArchT: 32.91±19.85, GFP: –1.54±9.8, Mann-Whitney U=2, ArchT = 7, GFP = 8, p=0.001, two-tailed), and increased cumulative duration in the outcome-associated area (ArchT: 50.04±36.64, GFP: –1.57±18.70, Mann-Whitney U=7, p=0.014, two-tailed) for ArchT animals compared with GFP.

Following ‘refresher conditioning,’ PRC rats completed a set of control tests, in which conditioned cue acquisition tests with and without PRC inhibition were completed. PRC inhibition had no effect on the duration of escape latencies, defined as the length of the first ‘stay’ only in the cued side, for either appetitive or aversive valences (laser: F(1,13)=0.88, p=0.37, η_p² = 0.06; laser × construct: F(1,13)=0.15, p=0.71, η_p² = 0.01; valence × construct: F(1,13)=0.04, p=0.84, η_p² = 0.003), with all rats exhibiting significantly shorter escape latencies for aversive than appetitive trials (valence: F(1,13)=84.84, p<0.001, η_p² = 0.87; construct: F(1,13)=0.06, p=0.82, η_p² = 0.004) (Figure 6h). Similarly, PRC inhibition had no effect on the amount of time spent in the cued side of the apparatus, defined as the total time spent across multiple stays in the chamber (laser: F(1,13)=0.59, p=0.46, η_p² = 0.04; laser × construct: F(1,13)=0.003, p=0.96, η_p² < 0.001; valence × construct: F(1,13)=0.22, p=0.65, η_p² = 0.02), with all rats spending significantly less time in the cued side during aversive than appetitive trials (valence: F(1,13)=470.47, p<0.0001, η_p² = 0.97; construct: F(1,13)=0.55, p=0.47, η_p² = 0.04) (Figure 6i). Collectively, these results suggest that PRC inhibition led to both a decrease in avoidance behavior and/or an increase in approach/stay behavior under motivational conflict, and that PRC-inhibited animals could readily discriminate between appetitive and aversive object cues.

As with the runway task, a separate cohort of animals received optogenetic inhibition immediately after object exploration (i.e., prior to the raising of the central guillotine door; Figure 7). All rats exhibited significantly shorter escape latencies after exposure to the aversive object pair compared with exposure to the appetitive object pair by the second acquisition test (valence: F(1,14)=123.22, p<0.001, η_p² = 0.9; valence × test: F(1,14)=116.05, p<0.001, η_p² = 0.89) (Figure 7b). The pattern of escape acquisition was comparable between ArchT and GFP animals (construct: F(1,14)=0.23, p=0.63, η_p² = 0.02; valence × construct: F(1,14)=0.04, p=0.85, η_p² = 0.003; test × construct: F(1,14)=1.06, p=0.32, η_p² = 0.07). Rats also spent more time in the outcome-associated area (stay behavior) when exposed to the appetitive object pair compared with the aversive object pair (valence: F(1,14)=419.32, p<0.0001, η_p² = 0.97; valence × test: F(1,14)=140.61, p<0.0001, η_p² = 0.91) (Figure 7b). This pattern of stay behavior was also comparable between ArchT and GFP animals (construct: F(1,9)=2.66, p=0.13, η_p² = 0.23; valence × construct: F(1,14)=0.35, p=0.56, η_p² = 0.02; test × construct: F(1,14)=0.95, p=0.35, η_p² = 0.06).

Figure 7

Download asset Open asset

When animals were exposed to a high conflict object pair, optogenetic inhibition of PRC (laser on) during the choice period significantly prolonged escape latencies compared with trials completed without inhibition (laser off; p<0.001), and compared to GFP control animals (p<0.001) with laser on and off (laser × construct: F(1,14)=47.74, p<0.001, η_p² = 0.77) (Figure 7c), indicative of a decrease in avoidance behavior under motivational conflict. Furthermore, despite making a comparable number of re-entries into the outcome-associated area during the test session (laser × construct: F(1,14)=2.07, p=0.17, η_p² = 0.13), PRC-inhibited animals spent significantly longer in the outcome-associated area containing the conflict object pair compared with trials completed without inhibition (laser off; p<0.001), and compared to GFP control animals (p<0.005) with laser on and off (laser × construct: F(1,14)=14.98, p=0.002, η_p² = 0.52) (Figure 7d). An analysis of the difference score of the two conflict test sessions (laser on–laser off) revealed that laser treatment both decreased escape latency (ArchT: 33.93±10.09, GFP: –1.02±10.14, t(14) = 6.91, p<0.001), and increased cumulative duration in the outcome-associated area (ArchT: 35.55±19.35, GFP: 3.21±13.58, t(14) = 6.91, p<0.001 t(14)=3.87, p=0.002) for ArchT animals compared with GFP. These findings align with the data obtained from animals receiving continuous PRC inhibition throughout the shuttle box test, indicating that impairments in encoding or valence recall when faced with high motivational conflict are unlikely to contribute to the decreased avoidance behavior following PRC inhibition.

To investigate whether the observed alteration in object-based AA conflict processing extended to conflict represented by ‘context-like’ cues, we administered an established Y-maze AA task that is vHPC-dependent, with large vHPC lesions and ventral CA3 (vCA3) or dentate gyrus inactivation increasing approach behavior in the face of high AA conflict (Schumacher et al., 2018; Schumacher et al., 2016Yeates et al., 2020).

Rats first learned the valences of three visuotactile cues (appetitive, aversive, and neutral) that spanned the length of three different maze arms (Figure 8a). Analysis of time spent in a given arm after four (test 1) and eight (test 2) conditioning sessions, without laser treatment, revealed that all rats acquired the cue-outcome associations successfully by test 2 (valence: F(2,32)=146.02, p<0.0001, η_p² = 0.9; valence × test: F(2,32)=24.26, p<0.0001, η_p² = 0.6), regardless of construct (construct: F(1,16)=0.17, p=0.69, η_p² = 0.01; test × construct: F(1,16) = 0.01, p=0.92, η_p² = 0.001; valence × construct: F(2,32)=1.51, p=0.24, η_p² = 0.09) by spending significantly more time in the appetitive arm (both tests p<0.001) and less time in the aversive arm (both p≤0.003) compared with the neutral arm (Figure 8b). Rats then completed two AA conflict tests during which they could freely explore between a conflict arm containing a combination of appetitive and aversive cues, and a neutral arm containing the neutral cue. PRC inhibition did not significantly change the time spent in either arm (laser: F(1,16) = 4.06, p=0.060, η_p² = 0.2; laser × construct: F(1,16)=1.06, p=0.32, η_p² = 0.06) (Figure 8c), and both ArchT and GFP animals spent a comparable duration of time in both the conflict and neutral arms (arm: F(1,16)=0.09, p=0.77, η_p² = 0.005; construct: F(1,16)=0.31, p=0.58, η_p² = 0.02; arm × construct: F(1,16)=2.47, p=0.14, η_p² = 0.13). Moreover, PRC inhibition did not significantly alter the number of entries made into or retreats from either the conflict or neutral arms (entries: laser: F(1,16)=0.001, p=0.98, η_p² <0.001; construct: F(1,16)=3.36, p=0.09, η_p² = 0.17; laser × construct: F(1,16)=0.07, p=0.8, η_p² = 0.004; retreats: laser: F(1,16)=0.03, p=0.86, η_p² = 0.002; construct: F(1,16)=0.5, p=0.49, η_p² = 0.3; laser × construct: F(1,16)=0.2, p=0.66, η_p² = 0.12) (Figure 8d–e). The number of entries and retreats also did not differ between conflict and neutral arms (entries: arm: F(1,16)=0.001, p=0.98, η_p² <0.001; retreats: arm: F(1,16)=1.19, p=0.29, η_p² = 0.07) for both ArchT and GFP animals (entries: arm × construct F(1,16)=0.07, p=0.8, η_p² = 0.004; retreats: arm × construct: F(1,16)=0.43, p=0.52, η_p² = 0.03). Thus, while PRC is critical for AA conflict processing associated with discrete objects, it may play a minimal role in contextual AA conflict.

Figure 8

Download asset Open asset

To investigate whether the vHPC plays a role in object-associated AA conflict processing, rats with either ArchT or GFP and optical fiber implants in the vCA3 (Figure 9a) were administered the object runway paradigm. vCA3 rats demonstrated successful valence learning across the two acquisition tests (valence: F(2,26)=160.44, p<0.0001, η_p² = 0.93; valence × test: F(2,26)=23.73, p<0.0001, η_p² = 0.65) (Figure 9b), and at test 2 spent the most time in the goal box after appetitive object exposure and the least time after aversive object exposure in comparison to neutral trials (both p<0.001). Object valence acquisition was similar for ArchT and GFP animals (construct: F(1,13)=3.85, p=0.07, η_p² = 0.23; valence × construct: F(2,26)=0.20, p=0.82, η_p² = 0.02) with the exception that ArchT rats spent less time in the goal box compared to GFP rats during acquisition test 1 (test × construct: F(1,13)=7.13, p=0.019, η_p² = 0.35; test 1 construct: F(1,13)=10.94, p=0.006, η_p² = 0.46). ArchT and GFP animals did not, however, differ at the end of training (test 2 construct: F(1,13)=2.21, p=0.16, η_p² = 0.15).

Figure 9

Download asset Open asset

In contrast to PRC inhibition, there was no effect of vCA3 inhibition on time spent in the goal box in the high conflict test session (laser: F(1,13)=2.92, p=0.11, η_p² = 0.18; construct: F(1,13)=0.001, p=0.99, η_p² <0.001; laser × construct: F(1,13)=0.93, p=0.60, η_p² = 0.02) (Figure 9c). Indeed, an overall comparison across regions revealed that the impact of laser manipulation was significantly different between PRC and vCA3 animals (laser × region × construct: F(1,29)=22.83, p<0.0001, η_p² = 0.44). vCA3 inhibition also did not impact other behavioral measures (Figure 9d) including the number of retreats (laser × construct: F(1,13)=0.03, p=0.86, η_p² = 0.003), entries made into the goal box (laser × construct: F(1,13)=0.58, p=0.46, η_p² = 0.04), and LTE (laser × construct: F(1,13)=0.37, p=0.55, η_p² = 0.03). vCA3 inhibition also did not affect time spent in the goal box when rodents were presented with the neutral object pair (laser: F(1,11)=0.03, p=0.90, η_p² = 0.03; laser × construct: F(1,11)=2.36, p=0.15, η_p² = 0.18) (Figure 9e).

Similar to the high conflict test, vCA3 inhibition did not impact the time spent in the goal box during the low conflict test (laser: F(1,13)=0.11, p=0.75, η_p² = 0.008; laser × construct: F(1,13)=0.29, p=0.60, η_p² = 0.02) (Figure 9f) and crucially, there was a significant difference between vCA3 and PRC rats when compared directly (laser × area × construct: F(1,25)=5.13, p=0.032, η_p² = 0.17). All vCA3 animals could readily discriminate between valences as reflected by time spent in the goal box (valence: F(1,13)=550.18, p<0.0001, η_p² = 0.98; valence × construct: F(1,13)=0.25, p=0.63, η_p² = 0.02) and retreats (valence: F(1,13)=13.80, p=0.003, η_p² = 0.52). There was, however, no effect of vCA3 inhibition on entries into and retreats from the goal box (both valence × construct: F(1,13)≤0.55, p≥0.47, η_p²≤ 0.09) (Figure 9g–h).

Following successful acquisition of a new set of appetitive and aversive object pairs (valence: F(1,13)=361.37, p<0.0001, η_p² = 0.97; pairing: F(1,13)=0.006, p=0.94, η_p² <0.001; valence × pairing: F(1,13)=0.19, p=0.67, η_p² = 0.02; pairing × construct: F(1, 13)=0.14, p=0.72, η_p² = 0.01) (Figure 9i), vCA3 inhibition also did not impact time spent in the goal box in the within-valence no conflict recombination test (laser: F(1,13)=0.14, p=0.72, ηp²=0.01; laser × construct: F(1,13)=0.040, p=0.85, η_p² = 0.003) (Figure 9j), with entries into and retreats from the goal box also unaffected (both valence × construct: F(1,13)≤1.10, p≥0.31, η_p²≤ 0.2) (Figure 9k–l). Both ArchT and GFP animals could discriminate between the two valences by spending more time in the goal box and retreating less for appetitive pairs compared to aversive pairs (valence: F(1,13)=481.57, p<0.0001, η_p² = 0.97; valence × construct: F(1,13)=0.03, p=0.88, η_p² = 0.002). In sum, vCA3 inhibition had no impact on any aspect of performance on the object AA runway task.

vCA3 rats also completed the object-based shuttle box, and all rats demonstrated successful establishment of an active avoidance response toward aversive, but not appetitive object pairings, across two acquisition tests, exhibiting significantly shorter escape latencies after exposure to the aversive object pair compared with exposure to the appetitive object pair (valence: F(1,13)=119.71, p<0.0001, η_p² = 0.90; valence × test: F(1,13)=92.12, p<0.0001, η_p² = 0.88; construct: F(1,13)=1.43, p=0.25, η_p² = 0.10; valence × construct: F(1,13)=1.18, p=0.30, η_p² = 0.08; test × construct: F(1,13)=0.03, p=0.87, η_p² = 0.002) (Figure 10a). All vCA3 rats also spent more in the outcome-associated area when exposed to the appetitive object pair compared with the aversive object pair (valence: F(1,13)=384.84, p<0.0001, η_p² = 0.97; valence × test: F(1,13)=9.41, p=0.0090, η_p² = 0.42; construct: F(1,13)=0.01, p=0.98, η_p² < 0.001; valence × construct: F(1,13)=0.68, p=0.42, η_p² = 0.05; test × construct: F(1,13)=1.28, p=0.28, η_p² = 0.09) (Figure 10b).

Figure 10

Download asset Open asset

In contrast to PRC rats, there was no effect of vCA3 inhibition on escape latency behavior when exposed to the high conflict object pairing (laser: F(1,13)=0.48, p=0.5, η_p² = 0.04; construct: F(1,13)=0.03, p=0.87, η_p² = 0.002; laser × construct: F(1,13)=0.29, p=0.87, η_p² = 0.02) (Figure 10c). An overall comparison across regions revealed that the impact of laser manipulation was significantly different between PRC and vCA3 animals (laser × region × construct: F(1,22)=13.78, p=0.0012, η_p² = 0.39). vCA3 inhibition also had no effect on the time spent in the outcome-associated area containing the conflict object pair (laser: F(1,13)=0.26, p=0.62, η_p² = 0.02; construct: F(1,13)=0.12, p=0.74, η_p² = 0.009; laser × construct: F(1,13)=0.020, p=0.89, η_p² = 0.002) (Figure 10d). Similarly, there was a significant difference between vCA3 and PRC animals when compared directly (laser × region × construct: F(1,22)=7.98, p=0.010, η_p² = 0.27).

vCA3 inhibition also had no effect on the duration of escape latencies for either appetitive or aversive valences during the control acquisition tests (laser: F(1,13)=0.51, p=0.49, η_p² = 0.04; laser × construct: F(1,13)=3.13, p=0.1, η_p² = 0.19; valence × construct: F(1,13)=0.25, p=0.63, η_p² = 0.02), with all rats exhibiting significantly shorter escape latencies for aversive than appetitive trials (valence: F(1,13)=826.68, p<0.0001, η_p² = 0.99; construct: F(1,13)=0.03, p=0.88, η_p² = 0.002) (Figure 10e). Similarly, vCA3 inhibition had no effect on the amount of time spent in the cued side of the apparatus (laser: F(1,13)=0.86, p=0.37, η_p² = 0.06; laser × construct: F(1,13)=0.04, p=0.84, η_p² = 0.003; valence × construct: F(1,13)=0.69, p=0.42, η_p² = 0.05), with all rats spending significantly less time in the cued side during aversive than appetitive trials (valence: F(1,13)=565.77, p<0.0001, η_p² = 0.98; construct: F(1,13)=0.30, p=0.59, η_p² = 0.02) (Figure 10f).

In sum, these results indicate that vCA3 inhibition had no impact on object-associated motivational behavior whether under AA conflict or the presence of appetitive or aversive cues alone.

To confirm optogenetic inhibition of cellular activity in the PRC and vCA3, cFos immunohistochemistry was performed following a PRC-dependent task: novel object recognition (NOR), in which rats explored an open maze containing a novel and a familiar object; or a vHPC-dependent elevated plus maze (EPM), in which rats explored two anxiogenic open arms and two ‘safe’ closed arms.

For the NOR task, an analysis of the discrimination ratio (difference between novel and familiar object exploration divided by total exploration) revealed that laser-treated PRC ArchT animals exhibited a significant NOR impairment (laser: F(1,34)=45.6, p<0.0001, η_p² = 0.57, construct: F(1,34)=28.46, p<0.0001, η_p² = 0.46; laser × construct: F(1,34)=39.63, p<0.0001, η_p² = 0.54) (Figure 11a) compared with laser-off ArchT and PRC-GFP control animals (all p≤0.001). In the EPM, laser-treated vCA3-ArchT animals spent a significantly increased proportion of time in the open arms (laser: F(1,11)=8.02, p=0.003, η_p² = 0.42; construct: F(1,11)=14.42, p=0.003, η_p² = 0.57; laser × construct: F(1,11) = 7.1, p=0.022, η_p² = 0.39) compared with laser-off Arch T and vCA3-GFP control animals (all p≤0.001) (Figure 11b).

Figure 11

Download asset Open asset

ArchT animals that completed either the NOR (PRC) or EPM (vCA3) tasks with the laser on prior to sacrifice demonstrated a significant reduction in cFos labeling in the PRC (laser: F(1,34)=98.11, p<0.0001, η_p² = 0.74; construct: F(1,34)=56.49, p<0.0001, η_p² = 0.62; laser × construct: F(1,34)=64.82, p<0.0001, η_p² = 0.66) or vCA3 (laser: F(1,11)=5.81, p=0.035, η_p² = 0.35; construct: F(1,11)=3.77, p=0.07, η_p² = 0.26; laser × construct: F(1,11)=12.78, p=0.004; η_p² = 0.54), compared with ArchT animals that completed the tasks with laser off and GFP control animals (all p≤0.001) (Figure 11c–e). Thus, laser treatment in ArchT animals significantly reduced neural activation.

Finally, histological analysis confirmed robust bilateral expression of GFP/ArchT and optic fiber tip placement immediately dorsal to the viral injection site in all PRC animals (Figure 11). Thus, no exclusions were made based on optic fiber placement/viral expression. In the vCA3 group, data from three ArchT-expressing animals were excluded based on the viral expression and optic fiber placement presenting too medially to the CA3 subfield.

Subjects were 53 male Long Evans rats (Charles River Laboratories, QC, Canada) weighing between 320 and 380 g at the start of the experiment. Following surgery, rats were individually housed to prevent damage to the implanted optic fiber, and maintained at a constant room temperature of 22°C under a 12 hr light/dark cycle (lights on at 7:00am). Water was available ad libitum, and animals were fed ad libitum up until food-motivated experimental training commenced. Food-restricted animals were maintained at 85–90% of their baseline free-feeding weight. All experiments were conducted during the light cycle and were done in accordance with the regulations of the Canadian Council of Animal Care and approved by the University and Local Animal Care Committee of the University of Toronto (Protocol no. 20012479). We anticipated a moderate to large effect size (0.6–1.1) based on unpublished behavioral data from our laboratory obtained from three previous cohorts of animals. These values were used for an a priori power analysis to compute the required sample size for the current experiment, in which a within-subjects design for optogenetic inhibition was used (Faul et al., 2007).

Rats were randomly assigned to receive viral infusions to either the PRC (n=38; n=18 ArchT, n=20 GFP controls) or the vCA3 region of the HPC (n=15; n=9 ArchT, n=6 GFP controls). Rats were anesthetized with isoflurane (Benson Medical, ON, Canada) and secured in a stereotaxic frame (Steolting Co, IL, USA) with the incisor bar set at –3.3 mm below the interaural line. An incision along the midline of the skull was made and the fascia retracted by small skin clips to reveal bregma. For PRC surgeries, the temporalis muscle was carefully peeled back from the temporal ridge to access lateral injection sites. Small burr holes were created directly over the injection sites using a dental drill, and 0.5 µl of virus (either AAV1-CaMKIIa-ArchT-GFP or AAV8-CaMKIIa-GFP, Addgene, MA, USA) was infused bilaterally with a 1 µl Hamilton syringe into either the PRC (AP: –5.2 mm, ML: ±6.7 mm, DV: –7.3 mm) or the vCA3 (AP: –5.2 mm, ML: ±4.8 mm, DV: –7.0 mm) over 5 min, with the needle left in situ for a further 5 min. Optic fibers were then implanted 0.5 mm dorsal to the infusion site, and were secured in place with dental cement anchored to jeweler screws implanted in the skull. Rats were given at least 7 days to recover in their home-cages before beginning behavioral experiments. The viral infusion and optic fiber implantation coordinates were chosen based on two previous cohorts of animals (unpublished observations) that demonstrated refined viral expression patterns and optimal fiber implantation localized to the PRC and vCA3 regions.

PRC rats and VCA3 rats first completed testing in the object-based AA runway task (Figure 1). They then underwent testing in the object-based AA shuttle box task. The PRC rats were then trained in the RAM AA task, prior to completing a final NOR test before euthanasia. vCA3 rats were administered a final EPM task before euthanasia.

All behavioral testing was recorded and tracked using Noldus EthoVision XT (Noldus, Netherlands). Data were analyzed using SPSS statistical package version 26.0 (IBM, ON, Canada). For the object-based AA task, conditioned cue acquisition test data for the time spent in the goal box, the LTE, number of entries and retreats, and object exploration latencies were analyzed by separate 3×2 repeated-measures ANOVA with a within-subjects factor of valence (appetitive, aversive, neutral), and a between-subjects factor of group. Analyzed data for the high conflict test were identical to the acquisition test (including object exploration data) and were subject to a 2×2 repeated-measures ANOVA with within-subjects factors of laser treatment, and a between-subjects factor of group. Each rat only completed one session of the neutral test, with between-subjects factors of laser treatment and group, and data were analyzed by univariate ANOVA. Data for the low conflict and no conflict control tests were each first analyzed by separate 2×2×2 repeated-measures ANOVA with within-subjects factors of valence and laser treatment, and a between-subjects factor of group. Activity analysis by the sucrose dispenser and generation of heatmap figures were completed using Noldus EthoVision XT, in which an 18 cm (L) × 11.5 cm (W) region of interest was demarcated within the goal box (around the dispenser; 1/3 of the total goal box), and could reliably contain an entire rat.

For the Y-maze AA task, conditioned cue acquisition data for the time spent in each of the cued arms, as well as the number of entries and retreats were analyzed by a 3×2 repeated-measures ANOVA with a within-subjects factor of valence (appetitive, aversive, neutral) and a between-subjects factor of group. Analyzed data for the conflict test were identical to the acquisition test and were subject to a 2×2×2 repeated-measures ANOVA with within-subjects factors of laser treatment and trial type (conflict vs neutral), and a between-subjects factor of group. For the EPM test data, 2 x 2×2 repeated-measure ANOVAs were conducted to compare the time spent in the open and closed arms, as well as the number of entries made into both, with a within-subjects factor of arm type (open vs closed) and between-subjects factors of laser treatment and group.

For the NOR task, a 2×2 repeated-measures ANOVA was conducted on the ‘discrimination ratio’ (d₂ = ${\displaystyle \frac{\left(\mathrm{t}\right)\mathrm{N}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{l}-\left(\mathrm{t}\right)\mathrm{F}\mathrm{a}\mathrm{m}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{a}\mathrm{r}}{\left(\mathrm{t}\right)\mathrm{T}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}}$), with a within-subjects factor of laser treatment and between-subjects factor of group. Behavioral data from the NOR task revealed that consistent with previous findings (Dix and Aggleton, 1999; Winters et al., 2004; Winters and Bussey, 2005), both novel object exploration and novel object discrimination performance were significantly higher during the first minute of the test session (time point: F(2,28)=20.79, p<0.001; object × time point: F(2,28)=16.41, p<0.001), and thus, this time point was selected for subsequent analysis. For cFos quantification, a 2×2 ANOVA was conducted on the density of cFos labeled cells in both the PRC and vCA3, with between-subjects factors of laser treatment (on vs off) and group.

Significant main effects and interactions were further investigated with simple contrasts and post hoc tests with a Bonferroni correction where appropriate. Violations to sphericity were addressed with a Huynh-Feldt correction. All manually scored data (e.g., entries, time spent in a given zone) were blindly scored by a second experimenter. None of the data collected were excluded in the final analysis.

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

This work was funded by the Canadian Institutes of Health Research (#156070 to ACHL and RI). The authors gratefully acknowledge Dr Maithe Arruda-Carvalho and the members of her DevNeuro lab in allowing us to use their Ni-U epifluorescence microscope.

All experiments were conducted in accordance with the regulations of the Canadian Council of Animal Care and approved by the University and Local Animal Care Committee of the University of Toronto (Protocol no. 20012479).

1. Received: June 29, 2022
2. Preprint posted: July 2, 2022 (view preprint)
3. Accepted: June 6, 2023
4. Accepted Manuscript published: June 14, 2023 (version 1)
5. Version of Record published: June 26, 2023 (version 2)

© 2023, Dhawan et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.