We investigated whether activation in somatosensory areas was affected by the discordance of visual information between intended and executed actions. Subjects received visual feedback about thumb movement that was unintended in the Asymmetric task of the Mirror condition or expected in the Symmetric task of the Mirror condition and NoMirror condition. In the conflict caused by the unintended visual feedback, activation in the SIIc was significantly higher in the Mirror condition than NoMirror condition. The results of this study are in line with our previous findings showing cross-modal interaction between somatosensory and visual modalities in the SII when looking at unintended visual information on a moving hand provided during movement execution
Viewing the body influences the speed of tactile reaction time
 and improves tactile acuity
. Longo et al.
 reported that the short-latency component of somatosensory evoked potentials (SEPs) generated in the SI was higher on viewing the body than viewing an object. Furthermore, in addition to viewing the body, the observation of moving body parts also modulates the SI activation
[15–18], whereas our results showed no changes of SI activity on viewing the unintended visual feedback of body movement. It is well-known that the neural activity in the SI is strongly inhibited during voluntary movement
[29–34]. One explanation for the absence of modulation of SI activity with the unexpected visual information in our study may be that the inhibitory effect from motor-related areas canceled out the effect from visual information.
We found that the SIIc response in the Asymmetric task was significantly higher in the Mirror condition than NoMirror condition, while no change was observed in the Symmetric task. The Asymmetric task in both conditions was similar with regard to motor command and somatosensory feedback, and the only difference was the visual feedback. There was a possibility that a very subtle twitch following the stimulation caused the difference in visual feedback between the Mirror and NoMirror conditions. However, since the stimulus intensity was just above the motor threshold, the extension-flexion of the thumb was much larger than the small twitch during movement tasks. It seems reasonable to eliminate the notion that the difference in visual feedback caused by a minor twitch induces the enhancement in the SIIc.
The neural mechanism responsible for modulation of SIIc activities through non-veridical visual feedback is unclear. One possibility is that the incongruent visual information enhances the SIIc activation via an attentional effect. The SII is involved in cross-modal attentional links between the somatosensory and visual modalities
, and the sight of body parts influences somatosensory event-related potentials with tactile spatial attention
. A probable explanation for the modulated neural activations in the SIIc observed in the Mirror condition might be that the visual feedback of unintended phase movement caused by replacement of the subjects’ left hand with a mirror image of the right hand implicitly led to increased attentional demands for the somatosensory information. Another possibility is that the neural mechanism providing the visual information on the body part would be influenced by predictions of visual feedback based on the motor commands. It has been suggested that a copy of the motor signal, known as an efference copy, is created so that sensory signals generated from external stimuli can be distinguished from reafferent signals from body movement
[2, 35, 36]. Corollary discharges are produced only if the motor commands interact with unpredicted sensory inputs and inhibits the neural response to the self-generated sensory signals
. More activity in somatosensory areas was found when the unpredicted stimulus was externally delivered
[38–40]. There is considerable validity to notion that the prediction of visual feedback modulates the somatosensory areas. In the Asymmetric task in the Mirror condition, subjects faced the surprise of seeing that their hand was not responding as intended, and activation in the SIIc might be modulated by the effect of corollary discharge.
We assume that this modulation in the SIIc is involved in computing the sensory errors by comparing the actual hand’s location to the estimated location for controlling movement. There is another cortical area that is important to a forward model
. The cerebellum builds internal models that predict the sensory feedback of motor commands and correct motor commands through internal feedback. The cerebellum also has been proposed to combine information from motor efferent and sensory afferent signals
. However, we could not record any cerebellar responses because our whole-head MEG system did not fully cover the cerebellum.
There is evidence that humans are normally not conscious of sensory feedback from movement
, and are aware that their arms and legs belong to them through somatosensory and visual inputs. This feeling of self-attribution is impaired when the predicted sensory information estimated from motor intention does not match the actual sensory information. In our study, the Asymmetric task of the Mirror condition corresponded to this situation. Some subjects reported feeling that movement in the Asymmetric task of the Mirror condition was not controlled by themselves or the body did not belong to them. Studies in patients and recent neuroimaging results in healthy subjects suggest a prominent role for the posterior parietal cortex
 and insula
[44–47] in the self-awareness of limb actions, the sense of agency. Inui et al.
 reported simultaneous activation in the contralateral SII and insula peaking at 90 to 160 ms after electrical stimulation. We assumed that the late component peaking at 150 ms (M150) in the SIIc may involve the SII activity of the neighboring insula. However, we could not find significant differences in the insula and PC. Although subjects reported a disturbance of agency, we assumed that it was not enough to induce the difference in these areas. Further study will be needed to clarify the functional role of these areas in sensorimotor integration.