bioRxiv2017
Knowing where north is provides a navigator with invaluable information for learning and recalling a space, particularly in places with limited navigational cues, like complex indoor environments. Although north is effectively used by orienteers, pilots, and military personnel, very little is known about whether non-expert populations can or will use north to create an accurate representation of an indoor space. In the current study, we taught people two non-overlapping routes through a complex indoor environment, with which they were not familiar - a university hospital with few windows and several turns. Along one route, they wore a vibrotactile compass on their arm, which vibrated continuously indicating the direction of north. Along the other route, they were only told where north was at the start of the route. At the beginning, the end, and back at the beginning of each route, participants pointed to well-known landmarks in the surrounding city and campus (external landmarks), and newly-learned landmarks in the hospital (internal landmarks). We found improved performance with the compass only for external landmarks, driven by peoples use of the availability of north to orient these judgments. No such improved orientation occurred for the internal landmarks. These findings reveal the utility of vibrotactile compasses for learning new indoor spaces. We speculate that such cues help users map new spaces onto familiar spaces or to familiar reference frames.Keywords: Spatial navigation; vibrotactile compass; reference frames; spatial memory; pointing; real world navigation.\n\nGetting lost in complex indoor spaces, like hospitals, airports, or subways, can be dangerous, distressing, and costly. Unlike outdoor spaces, which typically offer long sightlines and distal landmarks, indoor spaces are often confined, undifferentiated, and labyrinthine. Individual building layouts can be hard to learn if knowledge about direction with respect to the larger world is not easily attained. Confusing layouts, signs, and an inability to see outdoor landmarks compound these problems. Even in places with adequate signs or sight of outdoor landmarks, successful navigation by blind people and people with visual impairments is a substantial problem (Schinazi, 2008; Schinazi, Thrash, & Chebat, 2016). In the current study, we investigate whether providing non-visual directional information aids sighted people in learning an unfamiliar, complex indoor space. Specifically, we tested how people use a commercially-available vibrotactile compass, which vibrates in the direction of north. Unlike visual compasses, people with visual impairments can use vibrotactile compasses. A non-visual aid might also limit divided attention, which complicates the use of visual aids (Gardony, Brunye, & Taylor, 2015).\n\nWe tested two competing hypotheses in the current study. These hypotheses stem from a multifarious theoretical framework of navigational ability (Wolbers & Hegarty, 2010). 1. General: People would have access to more spatial information, and thus improve generally. In this case, providing a specific spatial cue decreases the error in the sensory signals, increasing the fidelity of the resulting spatial representations. We would expect improved performance across all aspects of navigation behavior. 2. Specific: People would have access to a specific spatial cue, and thus improve on tasks for which that cue is immediately helpful. In this case, providing a specific spatial cue only improves access to specific information, but in a way that does not create a more accurate general spatial representation. We would expect improved performance only on aspects of navigation behavior that knowledge of the cue is directly relevant. Addressing these questions offers insights into how spatial information is acquired in general, and provides insight into how vibrotactile compasses might aid blind or visually impaired navigators.\n\nWhy might people get lost in complex indoor spaces? One reason is that place-based strategies used by humans to navigate are thwarted by indoor spaces. Research on humans and rats reveals two neural systems, which undergird two navigational strategies (Hartley, Maguire,Spiers, & Burgess, 2003; Marchette, Bakker, & Shelton, 2011; McDonald & White, 1994;Morris, Garrud, Rawlins, & OKeefe, 1982; Munn, 1950; Packard & McGaugh, 1996; Restle, 1957; Tolman, Ritchie, & Kalish, 1946). First, a place-based strategy, implemented by the hippocampus, represents space as a cognitive map. A cognitive map is a flexible spatial representation, which affords the opportunity to infer novel shortcuts. Second, a response strategy, implemented by the caudate nucleus, applies a stimulus-response approach to navigation whereby choice points are identified, and responses are recalled (e.g., turn left at the bank, then right at the big tree). The response strategy is relatively inflexible, and generates associations between scenes and actions. In indoor environments, cues that support a place strategy, like distal landmarks and visual discriminability (Restle, 1957) are absent. Instead, place strategies rely on a process of path integration - tracking ones movement away from a starting location by attending to translations and rotations. Unfortunately, unlike rodents, people are poor path integrators (Foo, Duchon, Warren, & Tarr, 2007; Loomis et al., 1993). Thus, in indoor spaces, if people are more likely to adhere to response-based strategies, they are likely to gain less spatial knowledge about the overall environment.\n\nA second reason people easily get lost in complex indoor spaces arises from research on reference frames. A large body of evidence (Diwadkar & McNamara, 1997; McNamara, Rump, & Werner, 2003; Meilinger, Riecke, & Bulthoff, 2014; Meilinger, Frankenstein, Watanabe, Bulthoff, & Holscher, 2015; Mou & McNamara, 2002; Mou, McNamara, & Zhang, 2013; Mou & Wang, 2015; Roskos-Ewoldsen, McNamara, Shelton, & Carr, 1998; Shelton & McNamara,2004) supports the idea that spatial knowledge is stored in a preferred reference frame. A reference frame is a spatial representation in which objects or other representations of space are contained, or with respect to which they are ordered, oriented, located, or thought to move; a preferred reference frame is an orientation in which a spatial layout is most easily recalled. A reference frame is considered global if it is common across multiple areas of space (e.g., global north). In familiar, large-scale (i.e., city-sized) spaces, reference frames can be aligned with north (Frankenstein, Mohler, Bulthoff, & Meilinger, 2011), or a salient organizing feature (like a street through a campus; Marchette, Yerramsetti, Burns, & Shelton, 2011; Yerramsetti, Marchette, & Shelton, 2013). Indoor environments, however, provide less access to global reference frames. In the absence of global cues (like distal landmarks), people are likely to organize space into several local reference frames, instead of one global reference frame (Meilinger et al., 2014). In the extreme, using local reference frames means each segment of a route is disconnected from the last, and results in a 1-dimensional representation of space (i.e., an ordered sequence of places, with no 2-dimensional spatial relations specified; Ishikawa & Montello, 2006). Being able to integrate across areas of a learned route to take novel shortcuts, for example, requires some spatial knowledge about how to join local reference frames to each other to form a global reference frame - a difficult and demanding cognitive process (Weisberg & Newcombe, 2016; Weisberg, Schinazi, Newcombe, Shipley, & Epstein, 2014).\n\nIf people have direct sensory access to a global spatial direction, like a cardinal direction, they might map an unfamiliar indoor space onto the larger external environment, and construct a more accurate spatial representation of this indoor space. In this way, place-based strategies and global reference frames become more useful in an unfamiliar indoor space. Recent work has begun to characterize the use of vibrotactile compasses in particular. Konig and colleagues (Karcher, Fenzlaff, Hartmann, Nagel, & Konig, 2012; Kaspar, Konig, Schwandt, & Konig, 2014; Konig et al., 2016) developed a feelSpace belt, which provides tactile information about magnetic north. Their work illustrates that, with extensive training (seven weeks), the vibrotactile compass can improve basic homing tasks in sighted individuals (Konig et al., 2016).However, it is unknown how such directional information is incorporated into spatial representations, particularly in complex indoor spaces. Moreover, we were interested in whether this directional information could be incorporated and used to learn a new environment with only rudimentary training in the use of the device.
animal behavior and cognition↗