Pointing in 3D space to remembered targets - II: Effects of movement speedtoward kinesthetically defined targets

The accuracy of visually guided pointing movements decreases with speed. We have shown that for movements to a visually defined remembered target, the variability of the final arm endpoint position does not depend on movement speed. We put forward a hypothesis that this observation can be explained by suggesting that movements directed at remembered targets are produced wi thout ongoing corrections. In the present study, this hypothesis was tested for pointing movements in 3D space to kinesthetically defined remembered t argets. Passive versus active acquisition of kinesthetic information was co ntrasted. Pointing errors, movement kinematics, and joint-angle coordinatio n were analyzed. The movements were performed at a slow speed (average peak tangential velocity of about 1.2 m/s) and at a fast speed (2.7 m/s), No vi sual feedback was allowed during the target presentation or the movement. V ariability in the final position of the arm endpoint did not increase with speed in either the active or the passive condition. Variability in the fin al values of the arm-orientation angles determining the position of the for earm and of the upper arm in space was also speed invariant. This invarianc e occurred despite the fact that angular velocities increased by a factor o f two for all the angles involved. The speed-invariant variability supports the hypothesis that there is an absence of ongoing corrections for movemen ts to remembered targets: in the case of a slower movement, where there is more time for movement correction, the final arm endpoint variability did n ot decrease. In contrast to variability in the final endpoint position, the variability in the peak tangential acceleration increased significantly wi th movement speed. This may imply that the nervous system adopts one of two strategies: either the final endpoint position is not encoded in terms of muscle torques or there is a special on-line mechanism that adjusts: moveme nt deceleration according to the muscle-torque variability at the initial s tage of the movement. The final endpoint position was on average farther fr om the shoulder than the target. Constant radial-distance errors were speed dependent in both the active and the passive conditions. In the fast speed conditions, the radial distance overshoots of the targets increased. This increase in radial-distance overshoot with movement speed can be explained by the hypothesis that the final arm position is not predetermined in these experimental conditions, but is defined during the movement by a Feedforwa rd or feedback mechanism with an internal delay.