Key Components of Spatial Visualization Capacity

Spatial visualization is ubiquitous in human cognition. People visualize the spatial aspects of things in a variety of situations, ranging from mentally repositioning furniture to solving complex scientific and engineering problems. However, as anyone who has listened to complex directions over the phone can attest, our capacity to visualize spatial information is limited. Here we describe an ACT-R model (the spatial field model) of the underlying processes that limit visualization capacity, and apply this model to a measurement task called path visualization (Lyon, 2004). Like some other visualization tasks, (e.g., Brooks, 1968; Attneave & Curlee, 1983; Kerr, 1993) path visualization requires participants to imagine a series of movements (up, right, down...) in a two-dimensional or three-dimensional array of locations. Unlike these other tasks, path visualization requires participants to decide whether each new path segment intersects with any part of the existing path. This requirement to detect intersections forces a visualization strategy, rather than, for example, a verbal rehearsal strategy. As the path gets longer, load increases, visualization capacity is exceeded, and accuracy declines. The spatial field model predicts this decline very accurately by positing three key components of visualization capacity. These components are implemented using ACTR’s declarative memory processes. The model test reported here uses no free parameters, because (1) the values for all but two parameters are standard in ACT-R memory models, and (2) the values for the two remaining parameters were set previously using data from a different set of participants.