アイテム詳細

要旨

Using attention, humans can control to what extent which sensory information is granted access to higher brain areas. Since perception is often the result of a combination of different sensory cues, the question arises whether attention can also change the relative weights of the relevant cues in this sensor fusion process and thus influence the resulting percept. This question was studied on the example of body yaw rotations. The perception of the angle of body rotation is influenced by both visual and vestibular/proprioceptive cues. We measured how subjects reproduced rotations in a cue conflict situation. Subjects were rotated sitting on a Stewart motion platform with a projection screen (86° x 63°). The visual stimulus was a random-dot starfield providing optic flow information during rotation. The stars had a limited lifetime to prevent subjects from using them as reference points. Active-noise-cancellation headphones were used to prevent an uncontrolled influence of auditory cues. For each trial, first a concurrent rotation (between 7.5° and 18° in 6 steps) of the platform and the visual display was presented. The subjects were then instructed to either return or repeat the presented rotation actively by using a joystick. During their active reproduction of the turn, we applied a gain factor between the visual and vestibular rotation (rotation speeds of the visual movement relative to the platform movement varied between 0.71 and 1.42 in 5 steps). Using these gain factors allowed us to analyze the weights of the two cues in the subjects’ responses, and to investigate the influence of the rotation angle and the gain factors on the cue weights. In a second and third experiment, the subjects were additionally instructed to pay attention specifically to the visual movement or to the platform movement. Subjects were able to reproduce the angles with standard deviations of 5° - 15°, but tended to over-estimate small rotation angles by 1.5°. However, they did not notice that conflicting rotation angles were presented for visual and vestibular cues. In all three experiments subjects used a stronger weight for the visual cue than for the vestibular/proprioceptive cue. In experiment 2 and 3 no significant influence of attention on the perception of the rotations was found. Apparently, in human body rotation perception, the weights in the sensor fusion process can not be voluntarily changed by attention processes.