Abstract
People are able to predict the sensory consequences of their motor commands, such that they can compensate for delays in sensorimotor loops. It has been shown that integrating different sources of sensory information improves the precision of the prediction, but we lose precision when we adapt to sensory delays. There are two possible explanations for this: 1) The temporal conflict of the sensory modalities results in the loss of integration; 2) The delay forces people to rely more on the prediction of the consequence. To explore these options, subjects had to anticipate conflicting sensory information in a temporal-coincidence task. A target on a screen moved from left to right towards a reference line (at different speeds/distances). Subjects were instructed to press a button when the target came in contact with the line. As a consequence of the press, the target disappeared (visual feedback) and a sound was presented (auditory feedback). In the following conditions, either the visual consequence (VD condition), auditory consequence (AD condition) or both were delayed (VAD condition). Every condition consisted of a block of incremental delay (135 trials, 1ms/trial) followed by a block of fixed delay (135ms). The results show that the adaptation to delay is present in the conditions were the visual consequence is delayed (VD, VAD), but not when only the auditory consequence is delayed (AD). This is consistent with subjects using only the visual feedback to evaluate the success on the task. As expected, subjects show a decreased precision when adapting, compared to the conditions without adaptation. Importantly, the precision is not better in the VAD condition, suggesting that the accessory beep is not integrated. We propose that the increased variability is due to subjects having to rely on the prediction of the consequence whose uncertainty increases with the delay.
Meeting abstract presented at VSS 2016