1. The purpose of these experiments was to determine the accuracy with which human subjects could discriminate proprioceptive (nonretinotopic) targets during movement. The targets were located at either a specific angle in joint space, or a specified angular distance from an initial joint angle. 2. In these experiments the right elbows of normal human subjects were passively extended from either predictable or unpredictable starting angles. The subjects were instructed to open the right hand to indicate that the elbow was passing through a target joint angle or a target angular distance. The subjects were not given visual information about the location of the elbow, so they had to rely on proprioceptive input to perform this task. The target (criterion joint angle or angular distance) was learned by the use of proprioception during 8-15 practice trials. 3. Four experiments were conducted. In three experiments the target was located at a constant joint angle, and in the fourth experiment the target was located at a constant angular distance from the starting angle. The starting angle in all four experiments was pseudorandomly varied from trial to trial. 4. On the basis of an analysis of constant errors, subjects were more accurate at discriminating angular distance than joint angle. The slope of the relationship between the starting position and the constant errors was dictated by the task requirement. 5. In the distance discrimination experiment, when the starting angle was more flexed than the intermediate (i.e., central) position, the subjects slightly overshot the target distance. Conversely, when the starting angle was more extended than the intermediate position, the subjects slightly undershot the target distance. 6. In the joint angle discrimination experiments, the opposite results were obtained. Subjects overshot the target when the starting position of the elbow was more extended than the intermediate starting position, and they undershot the target when the starting position was more flexed than the intermediate starting position. The amplitude of these systematic errors increased when the subjects were unaware that the initial angle of their elbow was variable. 7. It is concluded that, in kinesthetic tasks of this type, the discrimination of angular distance is more accurate than the discrimination of joint angle. We hypothesize that the nervous system extracts kinematic information related to both joint angle and angular distance from proprioceptors, and that the encoding and or decoding of angular distance is more accurate than that of absolute joint angle. The results of these experiments are discussed in the context of the controversy of distance versus position (angle), 'endpoint' control during targeted movements, and the perceptual discrimination of joint angles.
ASJC Scopus subject areas