Item

Abstract

Ego-motion perception is a crucial ability of humans, particularly with regard to control of posture. Therefore the role of multisensory integration, which handles the information from multiple senses, seems to be a basic necessity. Currently, most experimental work on visualvestibular interaction is based on modification of one or more sensory signals. Cognitive variables
like previous knowledge, expectancy or attention are largely disregarded in this research area. A few recent research, however, suggests that especially attention has a significant impact on the way information from different senses is combined [1,2]. To investigate the influence of attention in an ego-motion perception task, we employed a
realistic, physics-based helicopter simulation on a Stewart motion platform. Participants performed a ‘hover task’, in which they were required to obtain and maintain a pre-defined orientation around the earth vertical by simultaneously controlling four (yaw, pitch, roll, height) of the six helicopter’s degrees-of-freedom. In a between-group design, participants were instructed to
focus either on the visual input (projection screen) or on the vestibular input (motion platform). During the hover task, subjects were distracted by task-irrelevant disturbances (visual lateral offset, vestibular lateral offset, or combined lateral offset). Given some inherent ambiguity between translational and rotational movements, we expected visually attending subjects to respond
strongest to the task-irrelevant disturbances during a visual offset. Similarly, vestibular instruction was expected to lead to a stronger deflection under a mere platform offset. 12 naive subjects and 11 experienced helicopter pilots participated in this experiment.
The results did not reveal any systematic influence of the mode of instruction on the control behavior. Further analysis showed that pilots and non-pilots exhibited different patterns of responses under the different conditions. Non-pilots were distracted only by visual disturbances. In contrast the pilots responded only to the combined offset (visual together with vestibular).
Furthermore, female subjects benefited from the vestibular whereas the male subjects profited by the visual focus of attention, resulting in a better hover performance.
The findings indicate that lateral disturbances are in part erroneously interpreted as rotational movements during helicopter flying. The specific response patterns suggest that nonpilots were using a rather visually based control strategy, whereas pilots relied on the integrated,
visual and vestibular, information—the later finding possibly reflecting the pilots’ previous knowledge of helicopter dynamics. The apparent failure to influence the control behavior by instructionally enforced focus of attention will be discussed in terms of different concepts
of attention and in terms of potential differences in the task demands of open-loop compared to close-loop studies.