A high density ERP comparison of mental rotation and mental size transformation Suresh D. Muthukumaraswamy, * Blake W. Johnson, and Jeff P. Hamm Research Centre for Cognitive Neuroscience, Department of Psychology, University of Auckland, Auckland, New Zealand Accepted 24 April 2003 Abstract Tocomparementalrotationandmentalsizetransformation,128-channelEEGwasrecordedwhilesubjectsperformedbothtasks using random two-dimensional shapes as stimuli. Behavioural results showed significant linear effects of both size transformation and mental rotation on reaction times. Rotation ERPs showed experimental effects at two latencies: a bilateral component dis- tributed over posterior parietal electrodes at a latency of approximately 232–300ms and a second component at approximately 424– 492ms distributed over right anterior parietal electrodes. The latency and spatial distribution of this second effect is consistent with previous research indicating a functional connection between this component and mental rotation. ERPs for the size-transformation task showed an effect at 180–228ms distributed bilaterally over occipital–temporal electrodes. These results are consistent with previous hemodynamic imaging studies that show involvement of parietal cortex in mental rotation and also the involvement of BA 19 in size-transformation tasks. However, the superior temporal resolution of the present data indicates that BA 19 activation may occur at a latency that is more likely related to apparent motion than to the size-transformation operation per se. Ó 2003 Elsevier Science (USA). All rights reserved. Keywords: Mental rotation; Size transformation; Image transformation; ERPs; EEG; Brain mapping 1. Introduction Mental rotation and mental size transformation rep- resent two different examples of imagery. Both underlie the ability of the visual system to recognise visual pat- terns as being the same, despite the fact that any one pattern can create an almost infinite number of retinal images as a result of experimental manipulation. Bun- desen and Larsen (1975) investigated the effect of al- tering the size of objects on reaction time (RT) judgements of same or different. In a simultaneous matching-to-sample task, subjects were presented with two stimuli and were asked to respond as quickly as possiblewhetherthestimuluspairwasofthesameshape and orientation, regardless of whether they differed in size. It was found that RTs were a linear function of the size ratio of the two stimuli. This was true of both line stimuli and filled two-dimensional stimuli. It was pro- posedthattoperformsuchtasksoneofthestimulimust be transformed in size to that of the other stimulus and then a comparison made. Linear RT functions for size transformation have also been found for successive matching-to-sample tasks (Larsen, McIlhagga, & Bun- desen,1999).Inamentalrotationexperimentusingtwo- dimensional shapes, Cooper (1975) required her subjects to learn eight different shapes in pre-training. In the experimental conditions stimuli were presented in either their normal or mirror imaged forms at various orien- tations and subjects had to respond as to whether they were normal or mirror image stimuli. A linear slope for RT as a function of orientation from the upright was found for normal images. These experiments indicate that rotation and size transformation of visual patterns share several theoret- ically interesting properties. As well as being examples of visual imagery, they both seem to be performed using an analogue rather than digital process. The linear in- creases in reaction time with respect to either size ratio Brain and Cognition 52 (2003) 271–280 www.elsevier.com/locate/b&c * Corresponding author. Fax: +11-649-373-7450. E-mail address: smut008@psynov1.auckland.ac.nz (S.D. Muthu- kumaraswamy). 0278-2626/03/$ - see front matter Ó 2003 Elsevier Science (USA). All rights reserved. doi:10.1016/S0278-2626(03)00077-0