Multimodal Imaging of Object Categorization

16 right-handed, healthy volunteers with normal or corrected-to-normal vision (mean ± std. = 25.87 ± 5.38, 10 female) participated in the experiment. The study was conducted according to the Declaration of Helsinki and is approved by the local ethics committee (Institutional Review Board of the Massachusetts Institute of Technology). 15 participants completed two MRI and MEG sessions, and one participant participated in the MEG experiment only. The sample size is comparable to that used in previous fMRI and MEG studies. All participants provided written consent for each of the sessions. During the experiment participants sawee images of 92 different objects presented at the center of the screen (2.9 degrees visual angle, 500ms duration) overlaid with a dark gray fixation cross. We chose this particular dataset for two reasons. First, it allowed assessment of distinctions at three levels: superordinate-, ordinate, and subordinate category. Second, it enabled direct comparison of our MEG and fMRI results with previous experiments utilizing the same date set in monkey electrophysiology and human MRI^{3 (link),26 (link)}. The presentation parameters were adapted to the specific requirements of each acquisition technique (Supplementary Fig. 1). In detail, for each MEG session, participants completed 10 to 15 runs, each having duration 420s. Each image wais presented twice in each MEG run in random order, with a trial onset asynchrony (TOA) of 1.5 or 2s. Participants were instructed to press a button and blink their eyes in response to a paperclip that was shown randomly every 3 to 5 trials (average 4). For each fMRI session, participants completed 10 to 14 runs, each having duration 384s. Each image wais presented once in each run in random order, with the restriction of not displaying the same condition on consecutive trials. 30 null trials with no stimulus presentation were randomly interspersed, during which the fixation cross turneds darker for 100ms and participants reported the change with a button press. TOA was 3s or 6s in the presence of a null trial.