Tolman’s spatial learning setup. (A) During pre-training, rats entered into a circular arena passing by A and B. Once in the arena, the only possible path is to follow C, a pathway with 18 inches walls that obstructed the rat’s vision. Following the path from D to G (surrounded by transparent walls), the rat’s goal, G, was a feeding box. A light was placed in such a way that the path from D to G was illuminated. The red line indicates the only path possible for the rat. (B) Following training in maze (A), rats were tested in the sunburst (maze B), the same location as during training but now equipped with 18 possible paths directly opening from the circular arena. During this test, the original training path was blocked. Out of the 18 possible arms (choices), only one, the sixth arm, pointed at the specific position of the feeding box. Out of 56 previously trained rats, 20 rats chose this shortest route, indicated by red. Figure modified from Tolman et al. (1946).

Spatial learning in goldfish. (A) The experimental setup was a cross-shaped tank. Fish were started either on the right or left arm (50/50). A cue card was associated with the rewarded arm during the entire training. Three types of probe tests were performed on a different group of fish that experienced the same training. Type A: Cue card, as well as room information, was available. Type B: Only the cue card was visible, room information was hidden thanks to a curtain. Type C: Cue card was removed, fish could only access the room information. (B) The bars present the percentage of choice for the location or the cue card during the probe tests. In this experiment, intact fish were used (control), as well as fish for which the telencephalon was ablated. In the probe test, similarly to zebrafish, control goldfish were able to use either the cue card or the location information. However, the telencephalon ablated fish were only able to use the cue card information. These results suggest a role of the telencephalon in configural memory retention, but not in elemental memory retention. The figure is adapted from figures in López et al. (2000a). Asterisks indicate significant (p ≤ 0.05) differences in the percentage of choices of the various arms.

Spatial learning in zebrafish. (A) The experimental tank was composed of a large square tank with four stimulus tanks placed in it adjacent to each sidewall. Only one of the stimulus tanks contained five conspecifics during training and was highlighted with a red plastic cue card and always keep at the same location. The other stimulus tanks were empty and did not have a red cue card behind them. The areas with different shading served as a template for quantification of the location of the fish during training and probe test. (B) During the probe test, in absence of conspecifics as a reward, fish from the paired group showed a significant preference for the area marked with the red card when the cue card was presented. However, in absence of a cue card, fish showed a preference for the original location of the conspecific tank during training. These results demonstrate that zebrafish were able to use both elemental as well as configural learning strategies to simultaneously learn both the association between CS-US and location-US. Figure modified from Karnik and Gerlai (2012).