Construction of the third dimension in mental imagery

Abstract

How do we create mental images of three-dimensional objects? Three clear alternatives are apparent: all object surfaces may be imaged simultaneously; parts may be imaged individually but in a random sequence; or parts may be imaged sequentially but constrained to unfold in a particular order. A set of experiments employing a new, objective methodology for measuring relative image generation time converged to demonstrate that images of three-dimensional objects evolve a part at a time. But more critically, the time required to image a given surface of the object was found to be proportional to its distance from the viewer, even when eye movements were restricted and object surfaces nearer the viewer subtended visual angles smaller than those subtended by more distant surfaces. Thus, mental images of three-dimentional objects were generated in a near-to-far sequence. Results were discussed in terms of computational and ecological considerations.