Sensorimotor signals underlying space perception: An investigation based on self-touch.

Perception of space has puzzled scientists since antiquity, and is among the foundational questions of scientific psychology. Classical "local sign" theories assert that perception of spatial extent ultimately derives from efferent signals specifying the intensity of motor commands. Everyday cases of self-touch, such as stroking the left forearm with the right index fingertip, provide an important platform for studying spatial perception, because of the tight correlation between motor and tactile extents. Nevertheless, if the motor and sensory information in self-touch were artificially decoupled, these classical theories would clearly predict that motor signals - especially if self-generated rather than passive - should influence spatial perceptual judgements, but not vice versa. We tested this hypothesis by quantifying the contribution of tactile, kinaesthetic, and motor information to judgements of spatial extent. In a self-touch paradigm involving two coupled robots in master-slave configuration, voluntary movements of the right-hand produced simultaneous tactile stroking on the left forearm. Crucially, the coupling between robots was manipulated so that tactile stimulation could be shorter, equal, or longer in extent than the movement that caused it. Participants judged either the extent of the movement, or the extent of the tactile stroke. By controlling sensorimotor gains in this way, we quantified how motor signals influence tactile spatial perception, and vice versa. Perception of tactile extent was strongly biased by the amplitude of the movement performed. Importantly, touch also affected the perceived extent of movement. Finally, the effect of movement on touch was significantly stronger when movements were actively-generated compared to when the participant's right hand was passively moved by the experimenter. Overall, these results suggest that motor signals indeed dominate the construction of spatial percepts, at least when the normal tight correlation between motor and sensory signals is broken. Importantly, however, this dominance is not total, as classical theory might suggest.