The Mental Representation in Mental Rotation

What is rotated in mental rotation? The implicitly or explicitly most widely accepted assumption is that the rotated representation is a visual mental image. We here provide converging evidence that instead mental rotation is a process specialized on a certain type of spatial information. As a basis, we here develop a general theory on how to manipulate and empirically examine representational content. One technique to examine the content of the representation in mental rotation is to measure the influence of stimulus characteristics on rotational speed. Experiment 1a and 1b show that the rotational speed of university students (10 men, 10 women and 10 men, 11 women, respectively) is influenced exclusively by the amount of represented orientation-dependent spatial-relational information but not by orientation-independent spatial-relational information, visual complexity, or the number of stimulus parts. Obviously, only explicit orientation-dependent spatial-relational information in an abstract, nonvisual form is rotated. As information in mental-rotation tasks is initially presented visually, a nonvisual representation during rotation implies that at some point during processing information is recoded. Experiment 2 provides more direct evidence for this recoding. While university students (12 men, 12 women) performed our mental-rotation task, we recorded their EEG in order to extract slow potentials, which are sensitive to working-memory load. During initial stimulus processing, slow potentials were sensitive to the amount of orientation-independent information or to the visual complexity of the stimuli. During rotation, in contrast, slow potentials were sensitive to the amount of orientation-dependent information only. This change in slow potential behavior constitutes evidence for the hypothesized recoding of the content of the mental representation from a visual into a nonvisual form. We further assumed that, in order to be accessible for the process of mental rotation, orientation-dependent information must be represented in those brain areas that are also responsible for mental rotation proper. Indeed, in an fMRI study on university students (12 men, 12 women) the very same set of brain areas was specifically activated by both the amount of mental rotation and of orientation-dependent information. The amount of orientation-independent information/visual complexity, in contrast, influenced activation in a different set of brain areas. Together, all activated areas constitute the so-called mental rotation network. In sum, the present work provides a general theory and several techniques to examine mental representations and employs these techniques to identify the content, timing, and neuronal source of the mental representation in mental rotation.