These patients report that they perform intended actions, even though they are paralysed and unable to move (Berti et al., 2005). This anosognosia was interpreted as showing that normal awareness of action is driven partly by both intentional signals, and by monitoring reafferent signals generated during actual movement. Dorsal premotor lesions appeared

to impair the integration of actual reafferent information, leaving the patient with an experience of agency that relied only on their intentions, without any feedback from the affected limb’s lack of movement. One might therefore interpret the dorsal ABT 199 premotor cortex as binding the sensory effects of action with the intentional action that caused them. This interpretation is also consistent with our data: stronger activation of this area was associated with stronger binding between action and effect. Moreover, our activation was found in the left hemisphere, in a task where participants responded with their right hand. Intentional

binding may depend on both predictive processes (e.g., motor command signals, Blakemore et al., 2002; Wolpert and Ghahramani, 2000) and on post-hoc reconstruction find more (Dennett and Kinsbourne, 1992; Wegner, 2002). The prediction account suggests that compression of perceived time occurs because neural preparation for action already triggers anticipation of the effects of action. In contrast, reconstructive accounts suggest that the mind infers and constructs a narrative

in order to explain bodily movements or their external Benzatropine consequences after the fact. Recent behavioural studies suggest that intentional binding includes both predictive and reconstructive components (Moore and Haggard, 2008). The current design does not allow us to formally separate the predictive and reconstructive components of sense of agency. We speculate that the computations within BA6 that underlie the sense of agency may recapitulate the medio-lateral gradient for the generation of action. Predictive contributions to sense of agency would rely on intentions and motor plans, and would be housed more medially, while reconstructive contributions to sense of agency would rely on integration of external sensory feedback, and would be housed more laterally. Therefore, the fact that our intentional binding cluster effectively straddles the intermediate zone between medial and lateral subdivisions may reflect the combination of both predictive and reconstructive processes. The two processes cannot be dissociated using interval estimation, but could be distinguished in future studies using estimates of action timing, and varying the probability that an action produces a tone. We found no evidence that the angular gyrus was associated with our implicit temporal measures of sense of agency.