Individuals who have suffered a significant loss of motor function as a result, for example, of spinal cord injury are now able to regain a degree of sensorimotor control through the use of a brain-machine interface (BMI). A BMI decodes intact neural signals to extract voluntary motor commands that reflect the individual’s movement intentions, and uses the processed signal to control an external device or the individual’s own body.

BMIs have been described as a (novel) way to “translate thought into action.” Accordingly, we should regard BMI-enabled control of a robotic (or biological) limb and a standard case of intentional bodily movement as equally providing examples of action, for in both cases we have bodily movement that is causally related to the individual’s belief, desires, and movement goals.

This description can be challenged, however, by reflection upon the nature of the causal relationship and the role of bodily awareness. According to an influential line of thought, in order for an action to occur it is not sufficient that the intention merely causes the bodily movement; rather, action requires that the cause be of the “right type” such that the movement is directly related and appropriately sensitive to the individual’s intentions. The argument can be made, however, that present BMIs does not meet this condition: first, indirect BMIs that co-opt neural signals that are not intrinsically related to the intended movement lack this direct relationship; second, although direct BMIs that decode movement parameters possess this relationship, the lack of sensorimotor feedback suggests that the causal connection is “deviant,” for the BMI serves as an unreliable causal intermediary.