Assessing Mongolian gerbil emotional behavior: effects of two shock intensities and response-independent shocks during an extended inhibitory-avoidance task

Overview of the study

We designed a step-down protocol for gerbils based on factors reported to have promoted the behavioral adjustment of this species to this task (e.g., shock frequency, platform size, habituation to the experimental chamber; Galvani, 1974; Galvani, Riddel & Foster, 1975; Lippman, Galosy & Thompson, 1970; Ögren & Stiedl, 2015; Osborne, Caul & Vanstrum, 1976; Twitty & Galvani, 1977; Walters & Abel, 1971). Though the most frequently implemented version of this procedure has entailed a single-trial training session and one testing session, additional trials/sessions of different length can also be scheduled during which the subject may be allowed to freely step down and return to the platform (Izquierdo, Furini & Myskiw, 2016). Considering the lack of information on gerbils’ long-term performance during step-down tasks, and the effects of extinction on that performance, our protocol entailed several sessions—hence the term “extended.” We adapted this protocol for implementation in the Video Fear Conditioning System (VFC; MED Associates Inc, 2009), which is a standard instrument that offers automatic scoring of rodent activity and freezing during aversive conditioning procedures (Anagnostaras et al., 2010).

In Experiment 1, the adapted protocol was used to compare the effects of implementing two foot-shock intensities considered in the low-to-mid range (0.5 and 1.0 mA) on gerbils’ (a) rate of step-down responses (platform descents) per session, (b) time spent on the platform per session, and (c) steady-state performance and extinction effects during the step-down task.

The decision to compare these two foot-shock intensities (0.5 vs. 1.0  mA) resulted from consideration of the following features: (a) Ballard, Sänger & Higgins (2001) used an analogous system to the VFC and could only reproduce defensive reactions in gerbils (flinch, vocalize, and jump—Hurtado-Parrado et al., 2015) when the intensity of the shocks was equal to or greater than 1.0 mA; (b) preliminary studies conducted in our laboratory entailing 1.0 mA shocks resulted in reliable reproduction not only of the same defensive responses reported by Ballard, Sänger & Higgins (2001) but also other responses such as running, thigmotaxis, and 360° jumps; and (c) approval of the step-down protocol by our Institutional Animal Care and Use Committee (IACUC) required the explicit effort to test a refined version of the proposed experimental procedure so the stress and pain of the animals could be reduced without compromising the experimental findings; in this case, reliably reproducing avoidance phenomena in gerbils.

Experiment 2 assessed the reliability and versatility of our step-down protocol by replicating and extending the findings of Experiment 1. Its purpose was twofold: first, testing whether an alternate procedure that entailed exposure to random delivery of foot shocks through the grid floor (i.e., response-independent shocks) was sufficient to reproduce the behavioral effects observed in Experiment 1. Second, the fact that freezing responses only accounted for 40% or less of the total session time during Experiment 1 led to the question of which other type of responses the animals were displaying during the task. Accordingly, we characterized gerbils’ exploratory behavior (rearing, digging, jumping, and probing) during the step-down protocol, which has been reported to differ importantly from that of other rodent models when exposed to aversive tasks (e.g., rats and mice; Crawford et al., 1981; Galvani, Riddel & Foster, 1975; Lippman, Galosy & Thompson, 1970; Osborne, Caul & Vanstrum, 1976; Rico et al., 2016).

Method

Apparatus

Video fear conditioning system—VFC (MED Associates Inc, 2009).

A VFC was adapted for the purposes of the present study. A clear polycarbonate (top and front), white acrylic (back), and stainless steel (sides) experimental chamber (32 cm wide, 25 cm high, 25 cm deep; Med Associates Part Number VFC-008) was encased in a white sound-attenuating box (63.5 cm wide, 35.5 cm high, 76 cm deep; Med Associates Part Number NIR-022MD). Inside the experimental chamber were located a stainless-steel grid floor (36 rods, each rod 2 mm diameter, 8 mm center to center; Med Associates Part Number ENV-005FPU-M) and drop-pan. A custom-made stainless-steel platform (25 cm × 5 cm × 0.3 cm) was hung 1 cm above the grid floor, which was secured to the back of the experimental chamber with a bar and clip 5 cm wide (see Fig. 1).

The interior of the experimental chamber was illuminated by overhead near-infrared light of 940 nm (Med Associates NIR-100). Background noise (65 dBA) was provided by a ventilation fan. Video images of the behavioral sessions were recorded at a frame rate of 30 frames per second (640 × 480 pixels) with a video camera equipped with a visible light filter (VID- CAM-MONO-2A). Appendix S1 shows a sample video frame.

A general activity index (Motion Index) was derived in real time from the video stream using computer software (Video Freeze^®; Med Associates Part Number SOF-843). The Motion Index was subjected to a threshold to generate automatic freezing scores (frequency, duration, and percentage of session time). The freezing threshold was set to the default Minimum Freeze Duration setting of the VFC system (1 s = 30 frames). Though this value was originally validated by Anagnostaras et al. (2010) with mice, preliminary analyses in our laboratory, analogous to those conducted by Anagnostaras et al. (2010), support the feasibility of implementing the same criterion with Mongolian gerbils.

An aversive stimulator/scrambler (Med Associates Part Number ENV-414S) connected to the grid floor harness delivered 0.5 s foot shocks of two intensities (0.5 or 1.0 mA; see Design and Procedure sections for details).

JWatcher+ Video 1.0 (Blumstein & Daniel, 2007).

Using the video files produced by the VFC for each experimental session, step-up and step-down responses and the time spent by each subject on and off the platform were manually scored and analyzed using the JWatcher+Video 1.0, which is a freeware available at http://www.jwatcher.ucla.edu (Blumstein & Daniel, 2007; University of California & Macquarie University). Necessary files for scoring the videos obtained from the VFC using the JWatcher (global and focal behavioral definition files) are available as Appendix S2.

Results

Table 1 shows effect-size scores (Tau-U) and corresponding statistical significance (p values) for each shock intensity (0.5 mA and 1.0 mA) per subject and for each measure of interest—i.e., % of session time on platform, % of session time freezing, and step-down rate.

Figures 2A–2D shows the percentage of session time (630 s = 100%) spent by each subject (S1–S4) on the platform (solid line) and freezing (dashed line) during each session of baseline (panels BL1, BL2, BL3, and BL4) and treatment conditions (0.5 mA and 1.0 mA). Figures 3A–3D show the rate of step-down responses per minute for each subject across the baseline and treatment conditions. Note that S1 and S2 differed from S3 and S4 regarding the order in which they were exposed to the treatment conditions—i.e., S1 and S2 initiated the experiment with foot shocks of 0.5 mA, whereas S3 and S4 received 1.0 mA shocks first.

Discussion

The extended step-down protocol for gerbils—via implementation in the VFC system—produced clear and reliable evidence of avoidance learning in terms of rapid and major increments in freezing and on-platform time, and decrements in rates of step-down. However, a comparison of the effect sizes of the two shock intensities implemented (0.5 and 1.0 mA) indicated that 1.0 mA shocks resulted in higher and steadier on-platform and freezing times, as well as more pronounced suppression of step-down responses. These findings confirm previous reports that intensities below 1.0 mA do not produce reliable and clear effects in aversive conditioning in gerbils (Ballard, Sänger & Higgins, 2001). They also suggest that implementation of dramatically higher intensities (e.g., 2–4 mA; Amano et al., 1993; Karasawa et al., 1990) is not necessary for reproducing learning and emotional phenomena in gerbils, including immobility and avoidance responses during step-down tasks.

The arrangement of extensive numbers of treatment sessions and foot-shock discontinuance phases permitted the observation of gerbils’ steady step-down performance and its extinction and reacquisition. The on-platform time and freezing measures typically reached stability in fewer sessions during the 1.0 mA conditions than during the 0.5 mA conditions. In the case of platform descents, they typically fluctuated across sessions, even during the 1.0 mA conditions, and very often failed to reach stability by the end of a given condition.

Extinction phases generally produced noticeable effects when such manipulation followed 1.0 mA treatment conditions. The typical course of immobility and on-platform time following cessation of the 1.0 mA shocks consisted of gradual decrements across sessions that in some cases reached baseline levels. In the case of step-down responses, foot-shock cessation produced increments in this response that tended to be more sudden than gradual. Reinstatement of 1.0 mA shocks after extinction periods typically produced rapid reestablishment of freezing, time on the platform, and step-down behavior to levels observed during previous treatment conditions. The reliability of these extinction and reacquisition effects showed promise for further refinement and implementation of the step-down protocol in pharmacological and behavioral neuroscience research of emotional-memory phenomena (e.g., Izquierdo, Furini & Myskiw, 2016; Knox, 2016).

Method

Results

The results of Experiment 2 are summarized in Figs. 4 and 5. Repeated-measures ANOVA showed differences in the percentage of the session time allocated to the platform (F(3, 19) = 24.04, p < .001, ${\eta }_p^2=.86$) and the average duration of visits to the platform (F(3, 19) = 7.238, p = .005, ${\eta }_p^2=.65$) across sessions of the study (D1-R, D2-R, D3-3s, and D4-3s). As shown in Fig. 4A, post-hoc comparisons indicated that gerbils significantly spent more time on the platform during both sessions with periodic shocks (D3-3s and D4-3s) than during sessions in which the shocks were delivered randomly (D1-R and D2-R). By the last session of the experiment (D4-3s), the subjects were spending close to 90% of the session time on the platform. Cohen’s d calculations indicated large effects for the significant differences between D1-R–D3-3s (p = .004, d = 2.481), D1-R–D4-3s (p < .001, d = 3.604), D2-R–D3-3s (p = .002, d = 2.554), and D2-R–D4-3s (p < .001, d = 3.637).

As shown in Fig. 4C, the average visit to the platform was significantly longer during the second session of periodic shocks (D4-3s) compared to both sessions with shocks delivered with the random-time schedule (D1-R and D2-R). The average visit duration during that last session was 73 s. Cohen’s d calculations indicate large effects for the significant post-hoc comparisons D1-R –D4-3s (p = .008, d = 1.988) and D2-R–D4-3s (p = .01, d = 1.907).

No significant effect was observed for percentage of freezing time per session (F(3, 19) = 0.57, p = .644, ${\eta }_p^2=.13$) and step-down rates (F(3, 19) = 1.29, p = .321, ${\eta }_p^2=.27$) when comparing randomly delivered shocks (D1-R and D2-R) and periodic shocks (D3-3s and D4-3s)—See Figs. 4B and 4D.

The exploratory behavior of gerbils varied across sessions of the experiment (see Fig. 5). Repeated-measures ANOVA showed differences in three behaviors that were typically displayed on the grid floor: digging rate (F(3, 19) = 4.04, p = .034, ${\eta }_p^2=.53$), percentage of session time rearing (F(3, 19) = 13.33, p < .001, ${\eta }_p^2=.76$) and jumping rate (F(3, 19) = 8.39, p = .003, ${\eta }_p^2=.662$). In addition, the Friedman test of ranks showed differences in the rate of probing from the platform (X²(3) = 12.62, p = .006, W = .71).

As shown in Figs. 5A and 5B, post-hoc comparisons between random-time based and periodic delivery of foot shocks indicated a significant reduction on frequency of digging and percentage of time rearing. Cohen’s d calculations for these significant differences indicated large effects in digging (D1-R vs D3-3s, p = .046, d = 1.405; D1-R vs D4-3s, p = .046, d = 1.405) and rearing (D1-R vs. D4-3s, p = .002, d = 2.447; D1-R vs. D3-3s, p = .004, d = 2.165; D2-R vs. D3-3s, p = .009, d = 2.845; D2-R vs. D4-3s, p = .004, d = 3.382) across sessions.

Figure 5C shows that the rate of probing markedly increased when gerbils were switched from random to periodic foot shock delivery; Cohen’s d for significant differences in probing across these sessions indicated large effects (D1-R vs D3-3s, p < .05, d = 0.922; D1-R vs D4-3s, p < .05, d = 2.000; D2-R vs D3-3s, p < .05, d = 0.925; D2-R vs D4-3s, p < .05, d = 2.101). Finally, the rate of jumping on the grid floor (Fig. 5D) was significantly higher during the first day of periodic shocks (D3-3s) than all other sessions, with large effects across all comparisons (D3-3s vs. D1-R, p = .009, d = 1.952; D3-3s vs. D2-R, p = .007, d = 2.058; D3-3s vs. D4-3s, p = .006, d = 2.117).

Discussion

Experiment 2 had a dual purpose: first, assessing a variation in the step-down protocol implemented during Experiment 1—effects of delivering the shocks on a RT 30 s schedule; and second, characterizing the activity of gerbils during the step-down protocol.

By the second session of the periodic shock delivery condition (D4-3s), the gerbils were spending nearly 90% of the session time on the platform, which replicated the findings of the 1.0 mA treatment conditions in Experiment 1 (see Fig. 2). The observation that such time allocations were significantly higher than those observed during sessions in which the shocks were delivered with a random schedule indicates that sporadic exposure to the foot shocks is not sufficient to produce substantial changes in preference for the platform area. This finding supports the effectiveness of the 1.0 mA protocol, and together with the observation that only two sessions with this intensity produced such significant changes, provides evidence for its suitability for further research.

Data on the average visit to the platform, together with the lack of differences in step-down rates across conditions, indicate that the effects of frequent shocks were not on the rate but on the length of the visits—i.e., the duration of each visit to the platform increased when gerbils were changed from random-time to periodic shock delivery. This finding is consistent with reports by Galvani, Riddel & Foster (1975) regarding differences between rats and gerbils in behavioral adjustments to step-down tasks. Galvani, Riddel & Foster (1975) found that in contrast to rats, which typically perform all their step-down responses after first exposure to the shocks, gerbils spaced a similar number of step-down responses over a longer period of observation.

The lack of differences in freezing time across the experimental conditions (percentages of session time remained approximately 35–40%) suggest that temporal patterning of shocks does not differentially affect this measure. Conversely, exploratory behavior did seem to be affected by a change from random-time to periodic shocks, as evidenced by the decrease in digging and rearing and the increments in probing and jumping. The observed patterning of these responses may relate to the notably-active defensive system of the gerbil (Ellard, 1993; Ellard, 1996; Ellard & Chapman, 1991), including its adjustment to the frequency and temporal distribution of aversive events. Variations in the probing response are of special relevance for further research (e.g., pharmacological) because they could not only be interpreted as a form of risk assessment but are also not displayed in step-down tasks by other popular rodent species (Galvani, Riddel & Foster, 1975).

Adaptation of the step-down protocol

The simple integration of a custom-made platform allowed for successful implementation of the designed step-down task for gerbils in the VFC system. This protocol produced reliable and clear evidence of avoidance learning in terms of rapid and major increments in freezing and on-platform time, as well as decrements in the rates of platform descents. The ease of adaptation of the protocol in a commercially available system, and the versatility of the system in terms of automatic scoring of activity and freezing measures, are expected to contribute to the need of a standard step-down protocol for this species. Further refinement and testing of this protocol could ultimately improve the commensurability of findings across laboratories, which in the past have implemented highly dissimilar protocols.

One major limitation of our protocol is that latency of the first step-down response—a widely used measure in related research (e.g., Borba Filho et al., 2015; Mina et al., 2014)—was not registered because no method for confining the subject to the platform prior to the start of the session was available. Further versions of the protocol should provide this datum, not only because it often has been the primary outcome of step-down tasks, but also because the relation between this measure and the others reported here could be interesting on its own. We are presently exchanging information with the developers of the VFC (MED Associates, Inc.) for this and related refinement purposes, including automated registering of the step-down and step-up responses.

Effects of two foot-shock intensities (0.5 vs. 1.0 mA)

A comparison of the effects of the foot-shock intensities (0.5 versus 1.0 mA; Experiment 1) indicated that 1.0 mA shocks produced clearer and more reliable performance in terms of higher on-platform and freezing times, and more pronounced suppression of step-down responses. Reproduction of these effects during the analogue condition of Experiment 2 (D3–D4) confirmed the effectivity of that same intensity, and overall the suitability of our step-down protocol when implemented in the VFC. The extensive discrepancy in shock-intensity levels across the relevant literature (0.4–4.0 mA) warrants further research examining the effects of different intensities on the adjustment of gerbils to not only step-down tasks but also to other related protocols that have been implemented with diverse shock intensities (e.g., step-through type avoidance tasks; Imai et al., 2007; Ji et al., 2007; Kaundal & Sharma, 2011).

The observation that 1.0 mA foot shocks were efficient for producing clear and reliable avoidance behavior suggests that implementation of higher intensities is unnecessary for reproducing aversive conditioning effects in gerbils. Evidently, this notion requires further confirmation in studies manipulating more systematically different shock intensities and extend the analysis to female gerbils, which have been reported to perform less efficiently in step-down tasks (Galvani, Riddel & Foster, 1975). These efforts are needed to continue minimizing the pain and distress of laboratory animals through the refinement of experimental procedures, as strongly endorsed by agencies worldwide (National Research Council, 2011).

Acquisition, extinction, and reacquisition of step-down performance

The gerbils’ steady step-down performance and its extinction and reacquisition were characterized in Experiment 1, which were efforts that have not been previously reported in the relevant literature. The on-platform time and freezing measures reached clear steady states in fewer sessions when the 1.0 mA foot shocks were scheduled. Psychological, pharmacological, and behavioral neuroscience research on emotion and learning phenomena involving operant conditioning (e.g., avoidance, punishment, or conditioned suppression; Aparicio, 2010; Knox, 2016) often requires the establishment of steady performance prior to testing the effects of other experimental manipulations (e.g., the effects of a given substance on performance). The step-down task reported herein offers an advantage over other experimental procedures (e.g., lever pressing in an operant chamber) such that steady performance could be obtained in a few short-duration sessions (6–8 sessions, 10 min each).

Although suppression of platform descents (step-down responses) was clearer and steadier across the 1.0 mA sessions, the rate of these responses often did not reach complete stability by the end of a given treatment condition, which could last up to 15 consecutive sessions. It is unclear why the gerbils did not completely stop descending from the platform and why the number of these responses continued to fluctuate after several sessions. However, Galvani, Riddel & Foster’s (1975) observation that the frequency of platform descents was higher in gerbils than in rats, warrants further research to extend these findings, which could result in a resumption of comparative efforts that were practically abandoned several decades ago. For example, it seems promising to explore patterns of step-down and exploratory responses (including probing, which Galvani, Riddel & Foster, 1975 only observed in gerbils) and analyze them in terms of the defensive behavior of these two species (e.g., unlike rats, which typically become almost permanently immobile, gerbils show facilitation of general activity while undergoing fear conditioning; Galvani, Riddel & Foster, 1975).

Similar to the case of acquisition, behavioral effects related to the extinction and reacquisition of the step-down task were clearer and steadier under 1.0 mA foot-shock conditions. Extinction conditions (i.e., foot-shock cessation) typically produced gradual decrements in freezing and on-platform time across sessions, whereas in the case of step-down responses, this procedure instead resulted in more sudden than gradual increments. Reacquisition of the task (i.e., performance during post-extinction treatment conditions) typically consisted of the rapid reestablishment of these three responses to levels observed during previous treatment conditions. Further studies could (a) explore differences in the pattern of immobility and on-platform time allocation versus platform descents during extinction conditions—i.e., changes in step-down responses were sudden, whereas variations in freezing and time allocation to the platform were typically gradual; and (b) examine the neurobiological and comparative aspects of the emotional-memory effects that were observed during the extinction and reacquisition conditions of the step-down protocol. In this regard, our step-down protocol shows promise for implementation in pharmacological and behavioral neuroscience research of complex emotional-memory phenomena (e.g., as an animal model of persistent fear memories characteristic of posttraumatic stress disorder—Perrine et al., 2016; Wurtz et al., 2016).

Random-time versus response-dependent foot shocks

A comparison of the adjustment of gerbils to random-time versus response-contingent shocks indicated that random-time shocks did not produce substantial increments in time allocation to the platform that were characteristic of response-dependent shocks. Experiment 2’s data showed that changing from random-time to response-dependent shocks affected (a) the length of each visit to the platform, and not the rate of step-down responses per se, which remained at the same level during both conditions; and (b) the patterns of exploratory behavior of gerbils, in terms of a decrease in digging and rearing, and increments in probing and jumping.

Although no differences in freezing times were observed across conditions of Experiment 2, it is worth noting that the levels remained close to those observed during the 1.0 mA treatment conditions of Experiment 1 (approximately 40%; see Fig. 2). This outcome suggests that exposure to both types of shock delivery (random-time and response-dependent) increase freezing to levels above the characteristic of no-shock conditions (i.e., baseline), which during Experiment 1 typically remained at approximately 20%. Evidently, further research is needed to test this interpretation because no freezing baseline data were collected during Experiment 2.

Similarly, the low rates of step-down responses across both conditions in Experiment 2 resembled those of the subjects that were first exposed to 1.0 mA shocks during Experiment 1 (see S3 and S4 in Fig. 3), i.e., at approximately 1.5/min. This finding, and the lack of differences across conditions in Experiment 2, suggested that both types of shock delivery suppressed step-down responses to levels below no-shock conditions (baselines), which during Experiment 1 typically remained close to 3 per min (see Fig. 3). Again, this interpretation requires further systematic testing because no baseline data were collected during Experiment 2.

Exploratory and defensive behavior

The exploratory behavior of gerbils that accompanied traditionally measured adjustment to the step-down task—freezing and on-platform time and frequency of platform descents—consisted of the suppression of digging and rearing responses, and increments in probing and jumping behavior. The patterning of freezing, platform-descents, and exploratory behavior may be related to the active defensive system of the gerbil (Ellard, 1993; Ellard, 1996; Ellard & Chapman, 1991), including its adjustment to the frequency and temporal distribution of aversive events. Further detailed research examining the patterning of probing behavior seems particularly promising considering its potential for comparative research, since it has not been reported in other rodent models and could possibly be interpreted as a form of risk assessment.