Beacon training in a water maze can facilitate and compete with subsequent room cue learning in rats.

Abstract

In Stage 1 of 4 experiments in which rats completed a water-maze blocking procedure, experimental groups were trained to use a predictive beacon (hanging above, connected to, or displaced from the platform) to find a submerged escape platform in the presence of predictive or irrelevant background cues and in the presence or absence of irrelevant landmarks. In Stage 2, a fixed beacon, landmarks, and background cues all predicted the platform location. A Room Test (landmarks and background cues only) showed that Stage 1 training with a fixed hanging beacon or the moving displaced beacon facilitated Stage 2 learning of predictive room cues for experimental relative to control subjects. In contrast, Stage 1 training with a moving pole beacon interfered with Stage 2 learning about predictive room cues relative to controls, whereas training with a fixed pole or moving hanging beacon had no effect. We conclude that multiple spatial learning processes influence locating an escape platform in the water maze.

Absence of overshadowing between a landmark and geometric cues in a distinctively shaped environment: A test of Miller and Shettleworth (2007).

“…They further demonstrated that this shortcoming could be overcome by attaching the wedge to the platform with a rod. Timberlake et al (2007) suggested that attaching the wedge to the platform made it more salient, and thus increased its effectiveness for blocking (see Hall, Mackintosh, Goodall, & Martello, 1977). Applying this reasoning to the present studies makes it possible to explain why the sphere failed to overshadow the geometric cues, but if the sphere was of low salience then it should have been overshadowed by the geometric cues.…”

“…If subjects are first trained to locate a goal with reference to a set of relatively weak cues then they may learn rapidly about the significance of cues that are subsequently introduced because they will be introduced against a stable spatial framework. Timberlake et al (2007) further acknowledge that this process can operate in opposition to associative learning about the cues that would be governed by, say, the Rescorla--Wagner (1972) equation. In the case of the landmark in the present studies, therefore, the geometric cues may have disrupted its influence through overshadowing, but compensated for this effect by providing a stable…”

“…A rather different explanation for potentiation in spatial learning has been put forward by Timberlake et al (2007), who proposed that satisfactory spatial learning depends upon the development of a stable framework of background cues. If subjects are first trained to locate a goal with reference to a set of relatively weak cues then they may learn rapidly about the significance of cues that are subsequently introduced because they will be introduced against a stable spatial framework.…”

Revaluation of geometric cues reduces landmark discrimination via within-compound associations

“…This same argument can be applied to other accounts of spatial learning that suppose learning based on one cue--type or frame of reference to be independent of another (e.g., Doeller & Burgess, 2008;Wang & Spelke, 2002. Austen et al have previously also argued that both template--matching and framework stability (Timberlake, Sinning, & Leffel, 2007) fail to account for the kinds of revaluation effects observed in the current paper. It may be argued that the revaluation effect observed in this study is not due to the formation of within--compound associations between the landmark and geometry cues, but instead is a consequence of the revaluation training for group INCON resulting in interference between the newly correct geometric location and 11 the previously correct landmark cue.…”

The Influence of Sex and Temperament on Spatial Learning in Domestic Dogs

“…In addition to relative information about objects in the environment, other mechanisms may also be necessary for spatial learning, including the use of landmark features, environmental geometry, and path integration (Timberlake, Sinning, & Leffel, 2007). Previous research has suggested that dogs are able to navigate the radial arm maze (RAM), but have a low spatial learning capacity when completing the radial arm maze task (e.g., MacPherson & Roberts, 2010).…”