Three-dimensional spatial learning in a virtual space station node

Abstract

Astronauts find it difficult to recognize their orientation while facing any of the viewing directions in 6-ported space station node modules. Our previous experiments tested the spatial memory of human subjects in 1-G in an analogous cubic virtual environment and showed that humans are able to learn to orient when instructed to imagine different body orientations while facing in two different directions. Can subjects do the task when facing in all 6 directions? Does training help? Does spatial memory depend on the direction of remembered targets relative to the body? Does performance depend on the subject's ability to rotate himself mentally and use imagery? How long is ability retained after training? 3D spatial learning was studied in two virtual cubic chambers, in which a picture of an animal was drawn on each wall. Through trial-by-trial exposures to a virtual chamber, subjects (n=24) had to memorize the spatial relationships among the 6 pictures around them and learn to predict the direction to a specific picture if they were facing any wall in any roll orientation. After learning in one chamber, the procedure was repeated in a second. Before being tested, subjects received computer-based instructions and practice. Half of subjects were taught to remember logical picture groupings (strategy), while the remaining (control) subjects were free to do the task as they saw fit. Subjects' retention of configurational knowledge (both chambers) and spatial ability (second chamber only, without feedback) were re-tested 1, 7, and 30 days after initial training. Response time (RT) and percent correct (% correct) learning curves were measured on all four days, while configurational knowledge was tested on the last three. All subjects ultimately learned to do the task within 36 trials in either test environment, but performed faster in the second environment than in the first (especially the strategy-trained group). The strategy group showed superior % correct and RT for above/behind targets and generally better configurational knowledge. Retention of configurational knowledge and spatial ability for both groups was good over 30 days. The subjects who reported using mental imagery (n=8) had higher scores on figure rotation tests and % correct for left/right targets. Performances by the control group on the experimental tasks were significantly correlated with those on conventional tests of field independence and 2/3D figure rotation ability. Strategy training helped those who had poorer mental rotation skills, and those who could not use mental imagery. Supported by NASA Cooperative Agreement NCC9-58 with the National Space Biomedical Research Institute, USA.