Is it possible to learn and transfer spatial information from virtual to real worlds?

Abstract

In the present study spatial behavior was assessed by utilization of a desktop virtual environment and a locomotor maze task. In the first phase of the experiment, two groups of healthy middle-aged participants had to learn and remember five out of 20 target locations either in a "real" locomotor maze or an equivalent VR-version of this maze. The group with the VR-training was also confronted with the task in the real maze after achieving a learning criterion. Though acquisition rates were widely equivalent in the VR- and locomotor groups, VR participants had more problems learning the maze in the very first learning trials. Good transfer was achieved from the virtual to the real version of the maze by this group and they were significantly better in the acquisition phase of the locomotor task than the group that had not received VR-training. In the second phase of the experiment -the probe trials- when the cue configuration was changed the group with the VR-training seemed to have specific problems. A considerable number of participants of this group were not able to transfer information.