Unitization of route knowledge

Abstract

There are many theories that explain how route knowledge is acquired. We examined here if the sequence of elements that are part of a route can become integrated into a single unit, to the extent that the processing of individual transitions may only be relevant in the context of this entire unit. In Experiments 1 and 2, participants learned a route for ten blocks. Subsequently, at test they were intermittently exposed to the same training route along with a novel route which contained partial overlap with the original training route. Results show that the very same stimulus, appearing in the very same location, requiring the very same response (e.g., left turn), was responded to significantly faster in the context of the original training route than in the novel route. In Experiment 3, we employed a modified paradigm containing landmarks and two matched routes which were both substantially longer and contained a greater degree of overlap than the routes in Experiments 1 and 2. Results were replicated, namely, the same overlapping route segment, common to both routes, was performed significantly slower when appearing in the context of a novel than the original route. Furthermore, the difference between the overlapping segments was similar to the difference observed for the non-overlapping segments, i.e., an old route segment in the context of a novel route was processed as if it were an entirely novel segment. We discuss the results in relation to binding, chunking, and transfer effects, as well as potential practical implications.

Appendix 1

Simple algorithms for modeling sequence learning: We ask what kind of simple algorithms could, in principle, describe performance in our experiments and, specifically, the key finding that the overlapping stimuli were responded to differently in the context of a practiced sequence than in isolation. These algorithms are clearly not cognitive models, but still they may be useful in that they illustrate the algorithmic complexity of the obtained results. For example, imagine one needs to program the order of operations for a robot from 1 to n. This can be done in several ways, A–D.

In this (A) situation after 1, 2 has to appear. Even when a robot performs Action 2 after (say) 6 rather than after the Action 1, it knows to proceed to Action 3. In B, if 2 appears, then 3 may not necessarily appear, rather, only if 1 and 2 appear in sequence will 3 follow.

In C and D situations, Action 3 must appear after Action 2 that follows Action 1. In Situation C for example, when the robot performs Action 2 after (say) Action 6 rather than after the first action, it does not know that it has to continue to Action 3. If the robot in situation D performs Action 6 and then 3, it will correctly infer Action 4. Yet even in such a case the robot does not seem able to reproduce the obtained behavioral results, as the overlapping segment is performed differently in the original and novel routes. The very same route sequence is performed differently by the cognitive system according to the route context it appears in.

One of the possibilities that arise from this study is that during training, there is a transition from declarative memory of separate connections between the locations from 1 to n, that is as in A, to procedural and automatic execution where Action 1 leads to Action 2 which leads to Action 3 which leads to 4 as in B, C and D. If one preforms the route in an automatic manner as a unit, but at some point transfers to a different route that partly overlaps with the old route, performance must revert again to declarative memory of separate connections between the locations.