Altering the visuomotor gain. Evidence that motor plans deal with vector quantities.

Two experiments investigated the effects of providing nonveridical knowledge of the results (KR) in a visuomanual task in which participants pointed to briefly (200 ms) presented targets without seeing their hand. By showing after each trial the movement endpoint displaced radially with respect to its true position, we were able to alter progressively the gain of the visuomanual loop. In experiment 1, the KR was provided only for transversal movements and for one target distance, but the effect generalized to all directions and all distances. Moreover, it also generalized to the other hand that had never been biased. In experiment 2, nonveridical KR was supplied for movements along the two major diagonals which require sharply different muscle synergies. The transfer to other directions and to the other hand was equally substantial. It is argued that the results support the vector coding hypothesis, which holds that the input to the motor execution stage is supplied by specifying independently the amplitude and the direction of the vector from the initial to the final position in an extrinsic frame of reference. We also discuss the possible brain structures involved in the biasing action of the KR.