Motor imagery of catching the ball, as compared with baseline, led to an increase in BOLD activity in cortical sensorimotor areas of the left

hemisphere and the right posterior cerebellum (Table 1). The cortical areas involved were the left supplementary motor area (SMA; Fig. 3A), the left IFG (Fig. 3B), the left posterior insula, the left postcentral gyrus, and the left IPL (Fig. 3B). In addition, the left anterior superior prefrontal cortex, the ventral ACC and the right inferior temporal cortex were activated (Table 1). To explore the BOLD changes found in the motor imagery condition in comparison with the action and observation conditions, regional analyses were performed across the following regions of interest: left ACC, left IFG, left SMA, and left IPL. We found a significantly higher degree of activation in the left SMA during TGF-beta inhibitor motor imagery than during active catching [T = −3.44, degrees of freedom (df) = 16, P = 0.003, Cohen's d = 0.8] and observation of catching [T = 3.57, df = 15, P = 0.003 (Fig. 4); pairwise t-tests with Bonferroni correction α = 0.003

and additional effect size Cohen's d]. The same pattern was observed for the left IFG (motor imagery vs. catching, T = −2.51, df = 16, P = 0.023, Cohen’s d = 0.6; motor imagery vs. observation, T = 2.26, df = 15, P = 0.039; Fig. 4) and left IPL signaling pathway (motor imagery vs. catching, T = −1.93, df = 16, P = 0.071, Cohen’s d = 0.5; motor imagery vs. observation, T = 1.84, df = 15, P = 0.086; Carbachol Fig. 4), although the medium effect as indicated by Cohen’s d was not statistically significant. Note that, in the left IFG and left IPL, there was no change in BOLD activity in the catching trial. No differences in the degree of activation were found when active catching and the observation of catching were compared within all regions of interest defined. In the current

fMRI study, as a first step to explore the neural correlates of RGS, we investigated in healthy volunteers whether actual or imagined catching of moving balls modulated the activity in candidate areas of the human mirror neuron system in frontal and parietal cortical areas. In order to address this question, we adapted the RGS to the fMRI environment, and compared active, passive and imaginary task conditions within a VR world. Similarly to the clinically used RGS, the MRI-adapted version simulated natural activities while maintaining action control by pressing of buttons to steer the avatar. In agreement with the working hypothesis behind the RGS, we observed the activation of a number of brain areas in the imagination condition, including the left SMA, the left IFG, the left posterior insula, the left postcentral gyrus, the left IPL, and the right cerebellum. These areas constitute a widespread circuit of sensorimotor areas including key cortical areas of the human mirror neuron system (Gallese et al., 1996; Iacoboni & Mazziotta, 2007; Sale & Franceschini, 2012).