Hanes Douglas A
Legacy Research Center, 1225 NE 2nd Ave, Portland, OR 97232, USA.
J Vestib Res. 2007;17(5-6):251-70.
It has been shown experimentally that under certain combinations of sensory stimuli, human subjects can perceive one of several distinct illusions about their overall orientation in or movement through space. In at least some cases, the structure of such multistable illusory perceptions of orientation can be efficiently described by perceptual transformations that act on a current orientation estimate to yield an updated perceptual construct. Repeated application of identified generating transformations yields a limited set of predicted illusions for a given sensory environment. This approach is especially valuable for perceptual data that exhibits discretely differing classes of illusions between subjects or trials. In a previous study, application of a semigroup of perceptual centering transformations has succeeded in reproducing and simplifying data from an experiment in which subjects experiencing visual vection reported a range of illusions about the orientations of their gaze, head, and torso to gravity. After reviewing previously obtained results on perceptual centering, this article generalizes the approach, presenting the mathematics required to characterize perceptual transformations. The developed framework should be widely applicable in the understanding of perceptual illusions, particularly when these are guided by alignment with preferred constructs. Secondly, the article reveals the nontrivial mathematical process of perceptual semigroup formation and evaluation, deducing the complete description of the semigroup constructed in the previous study. Perceptual centering transformations identified in terrestrial experiments may predict illusions to be expected in spaceflight. For example, our results indicate that under certain conditions, many astronauts will misperceive a visual rotation axis to be centered in front of the head or even the torso.
实验表明，在特定的感觉刺激组合下，人类受试者会对自身在空间中的整体方位或移动产生几种不同的错觉。至少在某些情况下，这种多稳态方位错觉的结构可以通过感知变换有效地描述，这些变换作用于当前的方位估计，以产生更新后的感知结构。对于给定的感觉环境，重复应用已识别的生成变换会产生一组有限的预测错觉。这种方法对于在受试者或试验之间表现出离散不同类错觉的感知数据特别有价值。在先前的一项研究中，应用感知中心化变换半群成功地再现并简化了一项实验的数据，在该实验中，经历视觉运动错觉的受试者报告了一系列关于他们的注视、头部和躯干相对于重力的方位错觉。在回顾了先前关于感知中心化的结果后，本文对该方法进行了推广，介绍了表征感知变换所需的数学知识。所开发的框架应广泛适用于理解感知错觉，特别是当这些错觉由与首选结构的对齐引导时。其次，本文揭示了感知半群形成和评估的非平凡数学过程，推导了先前研究中构建的半群的完整描述。在地面实验中识别出的感知中心化变换可能预测太空飞行中预期出现的错觉。例如，我们的结果表明，在某些条件下，许多宇航员会错误地将视觉旋转轴感知为位于头部甚至躯干前方的中心位置。
