自我运动路径辨别：图像稳定的影响

Johns Hopkins University School of Medicine, Wilmer Eye Institute, Baltimore, Maryland, USA.

J Vestib Res. 1995 Nov-Dec;5(6):411-20.

Path-deviation thresholds were measured as the effects of eye movements in the retinal flow were minimized through image stabilization. Thresholds obtained with image stabilization were compared to those obtained with unstabilized viewing to determine whether the elimination of eye movements from the retinal flow improves self-motion judgments. The results showed that, at slow forward speeds, eliminating the retinal effects of eye movements did not improve path-discrimination performance; subjects required more of an angular deviation to discriminate a circular from a straight motion path with image stabilization than with unstabilized viewing. In an effort to understand the results, eye movements were measured in unstabilized viewing conditions, and the measured eye velocities were used to estimate the retinal-image motion. The results showed that, for slow forward speeds, eye movements increased the average retinal speed, independent of the circular flow direction. At fast forward speeds, there was no significant increase in the average retinal-image speed due to eye movements. A parsimonious explanation for the decreased performance with image stabilization at the slow forward speed is that retinal-image motion was too slow to optimally stimulate the visual motion sensors.

通过图像稳定技术将视网膜血流中眼动的影响降至最低，以此来测量路径偏差阈值。将通过图像稳定技术获得的阈值与未进行稳定观看时获得的阈值进行比较，以确定消除视网膜血流中的眼动是否能改善自我运动判断。结果表明，在缓慢向前移动速度下，消除眼动的视网膜效应并不能提高路径辨别性能；与未进行稳定观看相比，受试者在使用图像稳定技术时需要更大的角度偏差才能区分圆形运动路径和直线运动路径。为了理解这些结果，在未进行稳定观看的条件下测量了眼动，并使用测量到的眼动速度来估计视网膜图像运动。结果表明，在缓慢向前移动速度下，眼动会增加平均视网膜速度，且与圆形血流方向无关。在快速向前移动速度下，眼动并未使平均视网膜图像速度显著增加。对于在缓慢向前移动速度下使用图像稳定技术时性能下降的一个简洁解释是，视网膜图像运动太慢，无法最佳地刺激视觉运动传感器。