This study investigated the influence of different modalities of target information (visual, kinesthetic) on the accuracy, kinematics, and interjoint coordination of pointing movements to remembered targets. The targets were presented by a robot arm in five locations in three-dimensional (3D) space, either as a point of light in a dark room ("visual" condition), or kinesthetically. Relative pointing accuracy in the visual compared with kinesthetic conditions was influenced by the target location: pointing errors were the largest for the visual targets most eccentric relative to the subject's head. In addition, for the two most lateral targets, the final arm positions were, on average, closer to the center than the targets in the visual condition and farther from the center than the targets in the kinesthetic conditions. This result suggests that the pattern of errors in the visual condition described elsewhere ("range effect") may derive from visual processing rather than motor planning and implementation. Two modes of kinesthetic target presentation were utilized. During "passive" kinesthetic presentation of the target, the experimenter moved the subject's relaxed arm. Alternately, in "active" kinesthetic presentation of the target, the subject actively (with minimal help from the experimenter) moved his arm. No visual feedback was allowed in either kinesthetic condition. The variability in the final fingertip position was significantly smaller in the active condition than in the passive condition. In contrast, variability in the final values of arm orientation angles did not differ significantly in the active and passive conditions. This apparent contradiction may be resolved by the fact that, for the given target location, the influence of the deviation of these angles in the given trial from their average values on the position of the fingertip tended to be mutually compensated, and this tendency was stronger in the active condition. Our analysis of the correlations among the arm orientation angles and of the relationship between the initial and final arm configurations suggests that the kinesthetic conditions enabled the implementation of a mixture of strategies for achieving accuracy. The first strategy is to use a specific memory of an adequate arm configuration (that assumed during target presentation), such that accuracy is achieved by using this memory as a template. The second strategy is to use synergistically coordinating joint angles, such that accuracy is achieved by focusing on a specific endpoint that can be reached by a range of equivalent arm positions. The latter strategy was better utilized in the active condition. In conclusion, our results indicate that human subjects can use diverse sensory information to achieve comparable final accuracy, but that the details of the strategies employed differ with the kind of information available.

Errors in pointing to actual and remembered targets presented in three-dimensional (3D) space in a dark room were studied under various conditions of visual feedback. During their movements, subjects either had no vision of their arms or of the target, vision of the target but not of their arms, vision of a light-emitting diode (LED) on their moving index fingertip but not of the target, or vision of an LED on their moving index fingertip and of the target. Errors depended critically upon feedback condition. 3D errors were largest for movements to remembered targets without visual feedback, diminished with vision of the moving fingertip, and diminished further with vision of the target and vision of the finger and the target. Moreover, the different conditions differentially influenced the radial distance, azimuth, and elevation errors, indicating that subjects control motion along all three axes relatively independently. The pattern of errors suggest that the neural systems that mediate processing of actual versus remembered targets may have different capacities for integrating visual and proprioceptive information in order to program spatially directed arm movements.

A three-dimensional tracking system was used to examine whether subjects with Parkinson's disease (PD) would show characteristic performance deficits in an unconstrained pointing task. Five targets were presented in a pyramidal array in space to 11 individuals with mild to moderate PD and 8 age-matched controls. After the target was indicated, subjects closed their eyes and pointed to the remembered target locations without vision. Despite the absence of visual feedback during movement, PD subjects were as accurate overall as controls. However, PD subjects showed greater variable errors, more irregular trajectories, and a vertical endpoint bias in which their endpoints were significantly lower than controls. They also showed deficiencies in the compensatory organization of joint rotations to ensure consistency in azimuthal (horizontal) positioning of the arm endpoint. We concluded that, under appropriate task conditions, PD subjects may not show overall deficits in accuracy even when making targeted movements at normal speed without visual feedback. Nevertheless, our findings indicate that there are certain dimensions of performance which are selectively altered in Parkinson's disease even when overall performance is normal.

No abstract

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 without ongoing corrections. In the present study, this hypothesis was tested for pointing movements in 3D space to kinesthetically defined remembered targets. Passive versus active acquisition of kinesthetic information was contrasted. Pointing errors, movement kinematics, and joint-angle coordination 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 visual feedback was allowed during the target presentation or the movement. Variability in the final position of the arm endpoint did not increase with speed in either the active or the passive condition. Variability in the final values of the arm-orientation angles determining the position of the forearm and of the upper arm in space was also speed invariant. This invariance occurred despite the fact that angular velocities increased by a factor of two for all the angles involved. The speed-invariant variability supports the hypothesis that there is an absence of ongoing corrections for movements to remembered targets: in the case of a slower movement, where there is more time for movement correction, the final arm endpoint variability did not decrease. In contrast to variability in the final endpoint position, the variability in the peak tangential acceleration increased significantly with 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 movement deceleration according to the muscle-torque variability at the initial stage of the movement. The final endpoint position was on average farther from 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 feedforward or feedback mechanism with an internal delay.