Article Preview
TopIntroduction
The ability to ‘learn’ the environment before engaging in navigation is an area of interest for a variety of application domains ranging from military, to touring of real space for people with mobility problems (Egsegian et al., 1993, Wilson et al., 1996, Foreman et al., 2003). Although spatial training can be provided through maps and briefings of an environment, these methods only provide topological (survey) knowledge, which, whilst being more flexible, pays little attention to the details of routes and landmarks; and requires ‘mental construction’ of the environment, rather than direct experience and formation of procedural memory (Thorndyke, 1980; Golledge, 1991). Procedural learning has a distinct advantage over survey knowledge, as can be seen in the experiments of Thorndyke and Hayes-Roth (1982) where participants with procedural knowledge of an environment, estimated route distances significantly better than participants who had acquired just survey knowledge. Navigation therefore relies heavily on previously acquired visual information, e.g. the process of re-orientation during navigation in a previously visited environment (Montello, 2005), which relies on previously seen “visual references” in order to adjust bearings during navigation. Maps and other traditional navigational equipment cannot provide the same level of supporting information. VE training, therefore, promises to provide procedural knowledge through direct experience, and has caught the attention of a variety of researchers all attempting to discuss whether virtual training is more efficient than training through more traditional methods (Witmer et al., 1995; Goerger et al., 1998; Waller et al., 1998; Foreman et al., 2003), and even whether mobile GPS systems can effectively help people acquire spatial knowledge of their environment while on the move (Huang et al., 2012).
Learning in virtual environments relies on the ability of users to develop an understanding of space either through the use of relational memory, or by creating a cognitive map of the environment (Asthmeir et al., 1993; Cobb & d’Cruz, 1994; Silverman & Spiker, 1997; Clark & Wong, 2000; Riva & Gamberini, 2000; Kumaran & Maguire, 2005; Konkel & Cohen, 2009). There is some debate over which of the two methods is employed by humans during spatial learning. Cognitive maps, are in a sense ‘mental images’ of an environment, and, according to Kumaran and Maguire (2005) are more likely to be the dominant method of spatial memory, since the human hippocampus shows a bias towards spatial relations during navigational learning. Relational memory, on the other hand has been described as essentially a mental network of nodes representing memories that occurred at a given time. These memories could be visual, but could also be part of a larger array of information from other modalities. For example, when walking down a familiar street, you may be flooded with memories of things that happened lasts time you walked down that path, the music that was playing in a nearby cafe, the people you talked to, but also any associations this street may have to other streets (that may run parallel to this street for example). Therefore, relational memory rejects the idea of a mental map model, but instead argues that maps are formed by a series of memories from various modalities (Eichenbaum et al., 1999).