Within the two parts, ROIs were defined as the 125 voxels (5 × 5 × 5) around the most active voxel (Fig. 2). Due to the proximity of the primary motor and the primary sensory cortex, we cannot exclude that some of the measured fMRI activity originated from the primary sensory cortex. Hence, we refer to this region as primary sensorimotor cortex. Figure 2 Subdivision of the primary sensorimotor cortex hand area in the more medial inferior and posterior (green) and more lateral superior and anterior (blue) part of all subjects.

Left picture side corresponds to right hemisphere, Inhibitors,research,lifescience,medical right picture side to left … Coding was as follows: for right-handers, right hand was coded as dominant hand and left hand was coded as nondominant hand, whereas Inhibitors,research,lifescience,medical left hemisphere was coded as dominant hemisphere and right hemisphere was coded as nondominant hemisphere. For left-handers it was the other way around, left hand was coded as dominant hand and right hand was coded as nondominant hand, whereas right hemisphere was coded as dominant hemisphere and left hemisphere was

coded as nondominant Inhibitors,research,lifescience,medical hemisphere. For each subject, the time course of the BOLD signal of the 11 experimental conditions was extracted separately for the four ROIs (dominant and nondominant hand area divided in two subregions each). The time courses were averaged over the four blocks of each experimental condition and over all voxels of the respective ROI. The mean signal of the 2-sec preceding the finger movements was assigned as baseline and the percentage Inhibitors,research,lifescience,medical of signal change to baseline was extracted and averaged from second six to 18 of every block for each experimental condition and ROI separately. Data were analyzed separately for both handedness groups with four mixed models, namely one for moving finger of the dominant hand, Inhibitors,research,lifescience,medical one for the nondominant hand, one for both hands under undivided concentration, and one for both

hands under divided concentration. In all analyses of the functional data, the random effect was click here subject and the fixed effects were hemisphere, subregion, attention level, and the interaction terms between the fixed effects. The fixed effects of the full models were tested with F-tests. The post hoc tests comparing two subconditions only were done with t-tests. In the case of missing data from an experimental condition (due Terminal deoxynucleotidyl transferase to technical issues), we excluded subjects from the subanalysis (right-handers nondominant hand, n = 1; both hand undivided attention, n = 2; left-handers both hands undivided attention, n = 2). Mixed-model calculations for the ROI analyses were performed with the nlme package (Pinheiro et al. 2012) in R 2.14.0 (R Development Core Team 2011). Reported significance levels were corrected for eight independent tests, to correct for the four models calculated in both handedness groups.