Abstract
The learning of sensorimotor mappings is often associated with after-effects (Helmholtz, 1909/1962). We investigated after-effects during the learning of hand gestures that required participants to control a simulated hand in virtual reality (Adamovich et al., 2009). METHODS: Finger movements (joint angles) were recorded with a data glove and were exploited for the online animation of a human hand model using a game engine. In the first phase of the experiment (baseline), the measured joint angles were mapped one-to-one onto the joint angles of the virtual hand. In the second phase (adaptation), this mapping was modified by flipping the joint angles of the index and the middle finger, and the ones of the ring and the little fingers. In the final phase (deadaptation), again the original natural mapping was reestablished. Participants had to reproduce hand gestures that were indicated by static pictures of hand postures. Training was terminated after 32 trials in the adaptation phase and after 16 trials in the deadaptation phase. RESULTS: In the adaptation phase, humans adapt to the unnatural sensorimotor mapping in a gradually improving manner, typically taking about 10-15 trials. The postural error and the reaction time are increased significantly (p < 0.001) at the beginning of the adaptation phase, and both measures returned to the baseline at the end of the learning. In the deadaptation phase both measures were increased typically only one or two trials before they returned to baseline p > 0.95 for both measures). CONCLUSIONS: The observed behavior seems not compatible with a standard sensorimotor adaptation effect, since the original sensorimotor mapping could be immediately established after adaptation within a single trial. Instead, after learning participants seem to recognize the appropriate sensorimotor mapping immediately, realizing it perfectly without observable after-effects. REFERENCES: Helmholtz, H. (1909/1962). Dover. Adamovich et al. (2009) Restor Neurol Neurosci, 27(3):209-23.
Meeting abstract presented at VSS 2015