Observational Fear Learning in Rats: Role of Trait Anxiety and Ultrasonic Vocalization

Abstract

Rats can acquire fear by observing conspecifics that express fear in the presence of conditioned fear stimuli. This process is called observational fear learning and is based on the social transmission of the demonstrator rat's emotion and the induction of an empathy-like or anxiety state in the observer. The aim of the present study was to investigate the role of trait anxiety and ultrasonic vocalization in observational fear learning. Two experiments with male Wistar rats were performed. In the first experiment, trait anxiety was assessed in a light-dark box test before the rats were submitted to the observational fear learning procedure. In the second experiment, ultrasonic vocalization was recorded throughout the whole observational fear learning procedure, and 22 kHz and 50 kHz calls were analyzed. The results of our study show that trait anxiety differently affects direct fear learning and observational fear learning. Direct fear learning was more pronounced with higher trait anxiety, while observational fear learning was the best with a medium-level of trait anxiety. There were no indications in the present study that ultrasonic vocalization, especially emission of 22 kHz calls, but also 50 kHz calls, are critical for observational fear learning.

Keywords: anxiety; observational fear learning; rat; ultrasonic vocalization.

Figure 1

Role of trait anxiety in observational fear learning. (a) Behavioral protocol (for more details, see Materials and Methods): After a light–dark box test, rats were divided into three groups (“demonstrator” (DEM), “observer” (OBS), “naive” (NAIVE)) with similar anxiety levels. DEM rats were fear-conditioned. One day later, the DEM and OBS pairs were exposed to the conditioning boxes. Last, all rats were individually tested for conditioned fear. (b) Mean percent time spent in the light compartment of the light–dark box. DEM, OBS, and NAIVE rats had similar anxiety levels. (c) DEM rats expressed more freezing in the retention test than rats of the two other groups. Importantly, OBS rats had significantly higher freezing scores than NAIVE rats, indicating observational fear learning in OBS rats. (d) Freezing scores as a function of trait anxiety (more% time spent in the light compartment indicate less anxiety). In DEM rats, higher freezing scores could be observed with higher trait anxiety. In OBS rats, a bell-shaped correlation was found, freezing was highest with a medium levels of trait anxiety. These correlations were confirmed by grouping the animals into low, medium, and high anxiety. (e) High anxious DEM rats had significantly higher freezing scores than low-anxiety DEM rats. (f) Medium anxious OBS rats had significantly higher freezing scores than low- and high-anxiety OBS rats. (g) Trait anxiety did not affect freezing scores in NAIVE rats. (h) The difference of trait anxiety within a DEM-OBS pair did not influence freezing scores in the retention test. In this analysis, the DEM-OBS pairs were separated into subgroups in which DEM rats being more anxious than the OBS rat (“DEM more anxious”), DEM and OBS rats that were similarly anxious (“none more anxious”), and OBS rats that were more anxious than the DEM rats (“OBS more anxious”). Abbreviations: DEM, demonstrator rats; OBS, observer rats; NAIVE, naive rats. Numbers in the bars indicate group and subgroup sizes. *** p < 0.001, ** p < 0.01, * p < 0.05; comparisons as indicated, after significant main effects in ANOVA.

Figure 2

(a) Behavioral protocol of experiment 2 was identical to experiment 1, but no light–dark test was performed, and ultrasonic vocalization was recorded throughout the experiment. (b) In the retention test, DEM rats showed more freezing than OBS and NAIVE rats. OBS rats had significantly higher freezing scores than NAIVE rats, indicating observational fear learning in OBS rats. Abbreviations: DEM, demonstrator rats; OBS, observer rats; NAIVE, naive rats. Numbers in the bars indicate group sizes. *** p < 0.001, * p < 0.05; comparisons as indicated, after significant main effects in ANOVA.

Figure 3

(a) Number of rats emitting 22 kHz calls, (b) latency of the first 22 kHz call, and (c) number of 22 kHz calls in the different phases of the experiment. 22 kHz calls were almost exclusively emitted by DEM rats. There was no difference in the number of rats emitting calls and in the number of calls. However, latency was longest in the fear-conditioning session and significantly longer during the observational fear learning session than during the retention tests. (d) Freezing behavior of DEM and OBS rats in the retention session as a function of the number of 22 kHz calls in the observational fear learning session. No correlations were found. (e) Next, the observational fear learning session were grouped into sessions with low, medium, and high number of 22 kHz calls. DEM rats of OFL sessions with low number of 22 kHz calls showed lower freezing behavior than the DEM rats from the other session. (f) Freezing behavior of OBS rats was not affected by the number of 22 kHz calls in the observational fear learning session. Y-axis scale and units in panels (e,f) are the same as in panel (d). Abbreviations: DEM, demonstrator rats; FC, fear-conditioning session; OBS, observer rats; OFL, observational fear learning session; Test, retention test. Numbers in or below the bars or boxes indicate group and subgroup sizes. * p < 0.05; comparisons as indicated, after significant main effects in Kruskal–Wallis test or ANOVA, respectively.

Figure 4

Parameters of the 22 kHz calls. Histogram depicting (a) the distribution of the call duration and (b) peak frequencies of the calls in the different phases (mean ± SEM). (c–f) Examples of calls with very short (c–e) or exceptionally long (f) call duration. Abbreviations: FC, fear-conditioning session; OFL, observational fear learning session; Test, retention test.

Figure 5

(a) Number of 50 kHz calls in the different phases of the experiment. Most 50 kHz calls were emitted in the observational fear learning session in which two rats were present. In the retention test, NAIVE rats emitted more 50 kHz calls than DEM and OBS rats. (b) 50 kHz call category profiles in the different sessions. The different colors of the bar sections indicate the different 50 kHz call categories. The numbers in the sectors represent the percent number of calls in a given category, related to the total number of calls in the respective session (indicated on the bottom). The letters after the number indicate whether the proportion of a call category differs across the sessions (χ² test). Different letters symbolize significantly different proportions; equal letters stand for non-significant proportions. (c) Representative examples of the different 50 kHz call categories. (d) Freezing behavior of DEM and OBS rats in the retention session as a function of the number of 50 kHz calls in the observational fear learning session. No correlations were found. (e,f) Therefore, the observational fear learning sessions were grouped into sessions with low, medium, and high number of 50 kHz calls. No effects of 50 kHz calls on the freezing behavior of DEM and OBS rats in the retention test were found. Y-axis scale and units in panels (e,f) are the same as in panel (d). Abbreviations: DEM, demonstrator rats; FC, fear conditioning session; OBS, observer rats; OFL, observational fear learning session; NAIVE, naive rats. Numbers in the bars or above the boxes indicate group and subgroup sizes. *** p < 0.001, * p < 0.05; comparisons as indicated, after significant main effects in Kruskal–Wallis test.