This study investigated whether manipulated visual feedback can affect neural activations while observing interactions in a virtual environment. Twelve healthy, right-handed subjects performed sequential finger flexions during functional MRI scanning. An MRI-compatible data glove was used to measure joint angles and to animate the movement of virtual hands in real time. Visual feedback of hand movement was manipulated by either reducing the angle of flexion in VR hand model relative to real movement (hypometric condition) or animate the movement of another finger that does not correspond to real movement (mismatched condition). The data showed no differences in movement kinematics across feedback conditions. Different types of feedback had different effects on the activity in the sensorimotor circuitry. The motor cortex contralateral to the moving hand was more active in the veridical condition (perfect correspondence between feedback and movement) than in the hypometric and mismatched conditions. The ipsilateral motor cortex was recruited more in the hypometric condition, while bilateral dorsal premotor areas and left ventral premotor areas were recruited more in the mismatched condition when compared to the veridical condition. These results imply the possibility to design virtual reality based rehabilitative paradigms that would facilitate remapping in specific brain areas of patients with neurologically-based movement disorders like stroke.