The Self-Reference System (SRS)

How do we know the locations and bearings of landmarks that lie outside our field of view as we navigate a familiar large-scale environment? Such knowledge underlies our ability to set a course to a goal, to change course in the face of unexpected obstacles, to anticipate landmarks along the way, and to plan the most efficient route between multiple goals. Following from animal models of spatial navigation, Jeanne Sholl has developed a functional model of human spatial memory that subscribes to the view that memory of environmentally scaled space has evolved in the service of navigation. The model distinguishes between a self-reference system, which codes allocentric and egocentric spatial relations from the body’s point of view, and an object-to-object system, which codes metric inter-landmark relations allocentrically. Retrieval of spatial knowledge from the allocentric representational system is mediated by the self-reference system. The model further distinguishes between self-reference system functioning at a perceptual-motor level and at a representational level. At the perceptual-motor level, the self-reference system codes the spatial coordinates of visible landmarks, and at the representational level, it retrieves the spatial coordinates of landmarks stored in the object-to-object system. Normally the two levels of self-reference system functioning are tightly synchronized, but they must be decoupled under certain conditions of spatial retrieval. The model is supported by behavioral findings consistent with (1) egocentrically mediated retrieval of allocentric spatial relations from both an actual and imagined point of view, (2) differential accessibility of spatial relations within the self-reference and object-to-object systems, (3) disrupted perspective taking when perceptual-motor and representational self-reference systems are misaligned.

Sense of Direction and Allocentric Heading Recall

Internal sun compasses, polarized light compasses and magnetic compasses have been discovered in ants, bees, and other insects (e.g., Gallistel, 1990) . Migratory birds use star compasses to compute their heading and mammals such as rats, mice, chinchillas, and monkeys have internal compass systems that use the self-motion signals produced by body rotation as well as input from visual azimuth cues to compute and update allocentric heading (e.g., Sharp, Blair, & Cho, 2001) . Despite the wealth of knowledge about how non-human animals compute allocentric heading, little is known about how it is computed by humans. Our lab is conducing research that applies what is known about the functioning of a head-direction (HD) system in other mammalian species to predict the behavioral effects expected if a similar system were present in humans. In our view, the efficacy of the putative human allocentric-heading system functioning is experienced subjectively as a sense of direction. Our early research on this topic show large individual differences in people’s ability to recover allocentric heading from pictured scenes of an overlearned environment that are highly related to individual differences self-reported sense of direction.

Episodic Memory and Path Integration

Endel Tulving has theorized that the episodic memory system enables people to mentally travel back in time to re-experience earlier events in their lives. Based on findings from navigational tasks in which animals appear to recover spatial information from earlier moments in time, Whishaw and Wallace (2003) have recently proposed that the path-integration system is an evolutionary precursor to the episodic-memory system. Path integration is a type of navigation in which a moving organism keeps track of its current location relative to its point of origin. Based on laboratory experiments with rats, Whishaw and Wallace argue that when path integration is interrupted, a rat can mentally return to earlier locations in its current trajectory to reengage the path-integration process. Thus, they suggest that the capacity to mentally time-travel was initially adapted to solve navigational problems and subsequently recruited by the human language system to recover non-spatial memories at earlier moments in time. A new line of research in our laboratory uses an interference paradigm to test the hypothesis that path integration and episodic memory share a common cognitive substrate in human adults.

Working Memory

How are resources shared/divided within a working memory system? Currently we are studying working memory (WM) in the context of multiple-systems models. Our current projects involve crossing processing and recall tasks of same and different domains (spatial and verbal) to create complex WM tasks. By testing the relative demands of various processing tasks on concurrent recall task we hope to better understand how WM divides and shares resources. In future studies we plan to explore the relationship of WM with LTM stores involving the object-to-object representational system.