POINTING IN 3D SPACE TO REMEMBERED TARGETS - I - KINESTHETIC VERSUS VISUAL TARGET PRESENTATION

This study investigated the influence of different modalities of targe t information (visual, kinesthetic) on the accuracy, kinematics, and i nterjoint coordination of pointing movements to remembered targets. Th e targets were presented by a robot arm in five locations in three-dim ensional (3D) space, either as a point of light in a dark room (''visu al'' condition), or kinesthetically. Relative pointing accuracy in the visual compared with kinesthetic conditions was influenced by the tar get location: pointing errors were the largest for the visual targets most eccentric relative to the subject's head. In addition, for the tw o most lateral targets, the final arm positions were, on average, clos er to the center than the targets in the visual condition and farther from the center than the targets in the kinesthetic conditions. This r esult suggests that the pattern of errors in the visual condition desc ribed elsewhere (''range effect'') may derive from visual processing r ather than motes planning and implementation. Two modes of kinesthetic target presentation were utilized. During ''passive'' kinesthetic pre sentation 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 conditi on. The variability in the final fingertip position was significantly smaller in the active condition than in the passive condition. In cont rast, variability in the final values of arm orientation angles did no t differ significantly in the active and passive conditions. This appa rent contradiction may be resolved by the fact that, for the given tar get location, the influence of the deviation of these angles in the gi ven trial from their average values on the position of the fingertip t ended to be mutually compensated, and this tendency was stronger in th e active condition. Our analysis of the correlations among the arm ori entation angles and of the relationship between the initial and final arm configurations suggests that the kinesthetic conditions enabled th e implementation of a mixture of strategies for achieving accuracy. Th e first strategy is to use a specific memory of an adequate arm config uration (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 ut ilized in the active condition. In conclusion, our results indicate th at human subjects can use diverse sensory information to achieve compa rable final accuracy, but that the details of the strategies employed differ with the kind of information available.