Abstract: Mental rotation involves the spatial transformation of the mental image of a physical stimulus. The following two mental rotation tasks have been employed in the literature: (1) left–right or normal-mirror judgment task and (2) the same–different judgment task. The time taken in making these judgments as well as the errors made often increase with the angular difference in orientation between the stimuli, or the difference between imagined and normalized orientation.
Further, mental rotation can be classified into two categories based on the reference frame adopted: egocentric and allocentric. The stimuli representing bodies or body parts such as hands are often considered as a special class of stimuli; this is because when they are used as stimuli in mental rotation tasks, they tend to induce egocentric mental rotation. However, there is no conclusive evidence demonstrating that material of body parts in fact draws forth egocentric mental rotation; more importantly, there has been no evidence demonstrating that the images of body parts can also elicit allocentric mental rotation.
In the current study, using two experimental paradigms (left–right judgment task and same–different judgment task), we explored the possibility of the dissociation of these two processing mechanisms using a special stimulus (hand), which may induce a mental transformation of either the viewer’s own body (hand) or the visual display. The stimuli comprised images of the back of the human hands created by a 3D graphics software. Each picture of the hand was presented at an orientation rotated in the medial (in-rotation) or lateral (out-rotation) direction from the upright orientation.
In experiment 1, each participant completed a left or right hand judgment task (LR task) when the picture of either a left or right hand was presented. Three factors were manipulated: hands (left hand vs. right hand), the direction of rotation (in-rotation vs. out-rotation), and the magnitude of rotation (45°, 90°, and 135°). All possible combinations of the different levels of the hands, direction of rotation, and degree of rotation were tested in a within-subject 2 x 2 x 3 completely randomized factorial design. In experiment 2, each participant completed a same or different judgment task (SD task) when the pictures of two hands were presented. In addition to the 3 factors used in experiment 1, there was an additional factor: the match (match vs. no-match). For both experiments, the dependent variables were the reaction time and the errors in the judgment.
For both experiments, the results indicated that the degree of rotation did have an effect, i.e., the performance decreased as the degree of rotation increased. In addition, for the LR task, for the same magnitude of rotation, the hand when rotated medially (in-rotation) was recognized more quickly and accurately than when rotated laterally (out-rotation). Such an effect of the direction of rotation (we termed it the “in-rotation effect”) was not seen for the SD task. This suggests that the processing of mental rotation in the LR task was limited by the biomechanical constraints of the corresponding physical rotation. Thus, the subjects might have used the egocentric frame of reference in mental rotation for the LR task but the allocentric frame of reference in the SD task. Therefore, the hand stimuli can elicit either egocentric or allocentric mental rotation depending on the experimental task.
These results suggest that participants use different spatial transformation mechanisms in LR (egocentric) and SD (object-centric) tasks. It appears that both the material of the body parts and the paradigms of mental rotation determine the reference frame that the participants adopt; further, we can state that the in-rotation effect might serve as an indicator of the dissociation of egocentric and allocentric mental rotation. Further, the result provided a good method for clarifying the disputation of unitary and multiple-systems models in egocentric and allocentric mental rotation.

Key words: mental rotation, in-rotation effect, angle effect, egocentric reference frame, allocentric reference frame