Observing a trained demonstrator influences associative appetitive learning in rats

Abstract

The ability to acquire information about the environment through social observation or instruction is an essential form of learning in humans and other animals. Here, we assessed the ability of rats to acquire an association between a light stimulus and the presentation of a reward that is either hidden (sucrose solution) or visible (food pellet) via observation of a trained demonstrator. Subsequent training of observers on the light-reward association indicated that while observation alone was not sufficient for observers to acquire the association, contact with the reward location was higher in observers that were paired with a demonstrator. However, this was only true when the light cue predicted a sucrose reward. Additionally, we found that in the visible reward condition, levels of demonstrator orienting and food cup contact during the observation period tended to be positively correlated with the corresponding behaviour of their observer. This relationship was only seen during later sessions of observer training. Together, these results suggest that while our models were not sufficient to induce associative learning through observation alone, demonstrator behaviour during observation did influence how their paired observer's behavioural response to the cue evolved over the course of direct individual training.

Keywords: appetitive social learning; associative learning; observational learning; social learning.

Figure 1.

Experiment 1 behavioural design and results. A graphical depiction of (a) the treatment of the rats in the various conditions, (b) the timeframe of each experimental phase, and (c) how the stimulus presentation was divided up for scoring. Mean (+/− s.e.m.) consumption of the sucrose solution over the course of (d) demonstrator and (e) (line graph) observer training suggested successful acquisition of the CS–US association. Additionally, (d) (bar graph) a comparison of drinking averaged across all sessions by sex shows significantly higher sucrose consumption in female demonstrators as compared with male demonstrators. (f–h) Contact with the faceplate through which the sucrose solution was made available was examined between T-Obs rats and UT-Obs rats across sessions during the (f) pre-CS, (g) (line graph) baseline (i.e. pre-CS) adjusted CS1, and (h) baseline adjusted CS2 timepoints. (g) (bar graph) An overall effect of experimental condition was detected on faceplate contact during the CS1 period, with T-Obs rats spending a significantly higher amount of time in contact with the faceplate than UT-Obs rats across sessions. No overall effect of condition was detected on (i) the average frequency of sipper contact per trial, (j) the average latency to first sipper contact per trial, or (k) the number of trials in which no sipper contact was made in a session. * p < 0.05, ** p < 0.01.

Figure 2.

Experiment 2 behavioural design and results. (a) A graphical depiction of the treatment of the rats in each in the conditions in experiment 2 and (b) the timeframe of each experimental phase and food restriction. An examination of food cup contact found that (c) during the pre-CS period, an overall reduced time of food cup contact was observed in D-Obs rats (bar graph) which did not differ across sessions (line graph). By contrast, no differences in food cup contact were observed between D-Obs and ND-Obs rats during the (d) CS1 and (e) CS2 period. Similarly, no significant difference between the experimental conditions was seen in orienting to the cue during the (f) pre-CS and (g) CS1 period. However, during the (h) CS2 period, D-Obs rats displayed slightly higher orienting at sessions 2 and 4, with the latter session not quite reaching significance. *p < 0.05, +p < 0.1.

Figure 3.

Experiment 2 correlations between demonstrator and observer behaviours. A correlation heat map of observer behaviour across training and the behaviour of their paired demonstrator during the observation phase. Correlations were calculated using averaged demonstrator behaviour across observation sessions and observer behaviour at each observer training session. *p < 0.05, ** p < 0.01, *** p < 0.001.

Figure 4.

Demonstrator/observer behaviour correlations across sessions. The above figures display the correlation coefficients (left) and associated p-values (right) for each session and for each combination of observer/demonstrator behaviours that displayed a trend across sessions. Behaviour combinations are indicated by the title above each correlation/p-value plot combination. Dashed black lines (left) indicate a r = 0, while the dotted black line (right) and the dashed red line indicate a p-value point of 0.1 and 0.05, respectively.