Embodiment of Spatial Reference Frames and Individual Differences in Reference Frame Proclivity

Spatial cognitive processes can be based on distinct reference frames centered on the body (egocentric reference frame) or centered on aspects of the surrounding environment independent of the position and orientation of the cognizing subject (allocentric reference frame). Updating of spatial information based on an egocentric reference frame is often believed to be highly automatic while the computation of spatial information based on an allocentric reference frame is assumed to be effortful and dependent on prior egocentric space processing. In this position paper I will review theoretical and empirical work that challenges the view of such a hierarchical organization of spatial reference frames and propose an embodied view of spatial reference frame computation. This perspective is based on three interdependent aspects of spatial reference frames. First, based on the existence of neural structures that allow for an automatic computation of both, egocentric and allocentric spatial representations, a functional equivalence of distinct reference frames is proposed. Second, based on the assumption of efficient computation and parallel accessibility of distinct reference frames individual proclivities are proposed to develop based on environmental and socio-cultural influences. Finally, it is proposed that ontogenetic differences are manifest in anatomical changes associated with the dominant use of different reference frames. These changes influence the microgenesis of spatial knowledge and influence behavior in other spatial cognitive tasks. In conclusion, an embodied framework of spatial reference frames strongly suggests consideration of individual reference frame proclivities to gain further insights into the complex architecture of human spatial cognition. More importantly, neuroscientific approaches describing cortical networks and associated brain dynamics have to allow participants to actively move and behave in their environment to allow for investigating the cognitive processes and brain dynamics underlying embodied spatial Page 2 of 40 URL: http://mc.manuscriptcentral.com/hspcc Email: [email protected] and [email protected] Spatial Cognition and Computation: An Interdisciplinary Journal 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 or Peer Rview O ly EMBODIMENT OF SPATIAL REFERENCE FRAMES 3 cognition. Introduction Spatial orientation is a complex cognitive function that allows for a wide variety of interactions with our environment. These range from exploring new territories without getting lost to performing fast and automated spatial maneuvers based on stimulus-response associations in well-learned surroundings. During spatial orienting humans use a multitude of information from several senses including but not limited to vision, audition, the vestibular system, and kinesthesis. All these different sources inform human navigators about their movement in the outside world as well as chang s in position and orientation with respect to aspects of the environment. This information is further integrated and coordinated with action plans and the behavior of other social agents. While this appears to be an effortless endeavor for human and animal navigators the complexity of the underlying spatial cognitive processes becomes obvious when orientation fails because of impairments of the underlying neural substrate (Aguirre & D'Esposito, 1999), because of ambiguity in the environment or bad wayfinding instruction (Waller, Montello, Richardson, & Hegarty, 2002), or when completely different spatial strategies lead to comparable orientation performance (Gramann, Muller, Eick, & Schonebeck, 2005). One of the most commonly observed differences in spatial cognitive behavior might be the use of maps (Lobben, 2004). While one group of navigators physically turns the map to align it with their actual heading in the real world a second group mentally rotates themselves to bring their actual heading in agreement with the north-south orientation of the map. Both strategies can be successful but differ fundamentally with respect to the underlying cognitive processes. Here, I will review research on individual differences in spatial orientation based on the use of distinct reference frames. The review focuses on the neural mechanisms underlying the Page 3 of 40 URL: http://mc.manuscriptcentral.com/hspcc Email: [email protected] and [email protected] Spatial Cognition and Computation: An Interdisciplinary Journal 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 or Peer Rview O ly EMBODIMENT OF SPATIAL REFERENCE FRAMES 4 computation of distinct reference frames in humans and other species, the cultural and geographical influences on individual proclivities to use distinct reference frames, and on ontogenetic differences of reference frame proclivities that are accompanied by and expressed in cortical changes. While this review is by no means exhaustive and does not cover sex differences in spatial cognition it makes a strong point for a more careful consideration of individual differences in spatial cognitive processing in general, and specifically, for a cautious interpretation of results from imaging studies describing the neural basis of spatial cognitive processes. Spatial Representations and Underlying Reference Frames In the example of the city map above two different spatial strategies can be applied to solve the same problem, i.e. updating of one’s own position and orientation with respect to a symbolic representation of the environment versus updating the representation of an environment with respect to one’s own position and orientation. The important difference between these two strategies is the reference frame that is used to update the navigators’ current position and orientation. A spatial reference frame can be defined as a means to represent entities in space (Klatzky, 1998) while the term strategy is defined as the use of distinct reference frames, or the use of a combination of distinct reference frames for cognitive ends. There are at least two distinct spatial reference frames that can be distinguished based on the origin of the underlying coordinate system and the information that is stored in the resultant spatial representation: an egocentric reference frame and an allocentric reference frame. An egocentric reference frame has an origin based on the navigator’s body and represents entities in the environment with respect to the current position and orientation of the navigator. In contrast, an allocentric reference frame Page 4 of 40 URL: http://mc.manuscriptcentral.com/hspcc Email: [email protected] and [email protected] Spatial Cognition and Computation: An Interdisciplinary Journal 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 or Peer Rview O ly EMBODIMENT OF SPATIAL REFERENCE FRAMES 5 has an origin outside the navigator and represents entities in space independent of the momentary position and orientation of the navigator (for extensive discussion see Klatzky, 1998). While this conceptual dichotomy makes it easier to dissociate qualitatively different spatial processes, it is reasonable to assume that there are more than two representations active during spatial orientation. One prominent example of parallel spatial representations is the model of Redish and Touretzky(Redish & Touretzky, 1997; Touretzky & Redish, 1996). In their model animal navigation is explained by four interacting spatial representations. The local view representation provides the relationship of the animal to visible landmarks and is based on an egocentric reference frame. A second, egocentrically based representation, the path integrator, provides a direct trajectory back to the starting point of travel. The place code, an allocentric representation indicates the location of the animal in an environment based on place cells located in the hippocampus, and finally, the head direction representation stores information with respect to the movement direction of the animal in an allocentric reference frame. Integrating information from all four spatial representations with a goal selection subsystem allows the animal to successfully navigate in new and well-known environments (Redish & Touretzky, 1997). More importantly, while there is physiological evidence for the parallel computation of at least three different allocentric spatial reference frames in the mammalian brain (i.e., heading direction, place cells, and grid cells; McNaughton, Battaglia, Jensen, Moser, & Moser, 2006; Moser & Moser, 2008; Taube, 1998) research in the fields of Cognitive Psychology and Geography mainly focuses on the distinction between an egocentric ‘sensorimotor’ representation of space and an allocentric ‘map-like’ representation of space. Such an approach fosters conceptual reduction of representational systems and limits the development of theories explaining complex spatial behaviors in humans. One example for such a simplification is the Page 5 of 40 URL: http://mc.manuscriptcentral.com/hspcc Email: [email protected] and [email protected] Spatial Cognition and Computation: An Interdisciplinary Journal 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 or Peer Rview O ly EMBODIMENT OF SPATIAL REFERENCE FRAMES 6 approach to explain the genesis of spatial knowledge in humans. A well-established theory of spatial learning in children assumes an ontogenetic sequence from egocentric to allocentric representations of space implying a sequential development from basic and coarse to advanced and complex spatial representations (Piaget & Inhelder, 1967). During the first months when infants are more or less stationary in their environment they are assumed to use a sensorimotor representation (egocentric reference frame). Then, with onset of locomotion (crawling), a qualitative shift from sensorimotor to map-like (egocentric to allocentric coding) is proposed that helps infants to keep track of their position in the environment (Acredolo, 1990; Thelen & Smith, 1994). This concept of a developmental sequence from egocentric to allocentric spatial representations has been extended to a more general three-stage model of spatial knowledge acquisition in new environments (Hazen, Lockman, & Pick, 1978; Siegel & White, 1975). According to this framework of spatial microgenesis the first stage of spatial knowledge entails a representation of landmarks. Landmarks are prominent enduring features that define a specific spatial location of the perceiving navigator and are thus based on an egocentric reference frame. In the second stage, through repetitive travel, route knowledge develops connecting previously learned landmarks. The underlying reference frame is egocentric in nature providing a representation of sensorimotor sequences that allow the navigator to reach one landmark while being located at another (e.g., when stepping out of the train station, go straight ahead for two blocks and then turn left and you will see the cathedral). Finally, the connection of different routes into a two-dimensional map-like array defines survey knowledge, an allocentric spatial representation allowing for shortcuts and detours. This model explains a wide range of findings implying a hierarchy of spatial representations with a functionally superior (superiority Page 6 of 40 URL: http://mc.manuscriptcentral.com/hspcc Email: [email protected] and [email protected] Spatial Cognition and Computation: An Interdisciplinary Journal 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 or Peer Rview O ly EMBODIMENT OF SPATIAL REFERENCE FRAMES 7 assumption) allocentric representation preceded by and based on (dependence assumption) an egocentric spatial representation. Such a map-like allocentric representation is thought to allow more flexible spatial behaviors and is proposed in several models (e.g., Golledge & Spector, 1978; Siegel & White, 1975). However, there is good reason to question the assumptions of superiority and dependence of allocentric over egocentric spatial representations. Theoretical and empirical evidence suggests a parallel development of egocentric and allocentric spatial knowledge for both, the ontogenesis and microgenesis of spatial knowledge [insert Footnote 2 here](e.g., Graziano & Gross, 1994; Haith & Benson, 1998; Ishikawa & Montello, 2006; Mou, McNamara, Valiquette, & Rump, 2004; Newcombe & Huttenlocher, 2003). Experimental results support the assumption that allocentric spatial representations are simply one specific form of spatial representations that do not require effortful computation and are just as basic for spatial orienting as egocentric representations of space. Here, I will develop and discuss three general and interconnected hypotheses describing the embodiment of spatial reference frames. First, it is assumed that the mammalian brain provides a neural basis to compute several egocentric and allocentric reference frames in parallel. That is, the neural basis allowing for a computation of distinct spatial representations is softwired. Secondly, these genetically pre-specified neural circuits need specific inputs for maturation. Such inputs are to a certain extent coupled to the development of the navigators’ physical structure and thus the ability for movement in space. In addition, specific aspects of the environment (socio-cultural as well as geographical) will influence the reference frame used. Third and finally, successful use of a specific reference frame (or class of reference frames) over the time course of individual development is associated with a proclivity to use a specific Page 7 of 40 URL: http://mc.manuscriptcentral.com/hspcc Email: [email protected] and [email protected] Spatial Cognition and Computation: An Interdisciplinary Journal 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 or Peer Rview O ly EMBODIMENT OF SPATIAL REFERENCE FRAMES 8 reference frame. This proclivity for a specific reference frame is associated with structural changes in the cortical networks underlying the computation of the specific reference frame. In the following I will develop each of the three hypotheses, providing a theoretical as well as