Harris J M, Watamaniuk S N
Smith-Kettlewell Eye Research Institute, San Francisco, CA 94115, USA.
Vision Res. 1996 Jul;36(14):2149-57. doi: 10.1016/0042-6989(95)00278-2.
Luminance-defined and stereo-defined (cyclopean) motion share some common properties, suggesting that the two forms of motion may be detected by similar mechanisms. For luminance-defined motion there are at least two levels of processing: direction is detected and then speed is thought to be extracted by a specialized processing mechanism at a higher level. Here, we tested whether there is also a specialized speed processing mechanism for stereo-defined motion. Speed discrimination thresholds were compared for stimuli containing only stereo-defined motion, and stimuli that contained both stereo-defined and luminance-defined motion. When the stimulus contained luminance-defined motion, increment thresholds were around 0.05-0.1. For stereo-defined motion, increment thresholds were never better than 0.3. By careful analysis, it was possible to test what cues were being used to solve the speed discrimination task. Results were consistent with observers responding to distance cues rather than to speed for stereo-defined motion, suggesting that there is no specialized mechanism for processing the speed of stereo-defined motion.
由亮度定义的运动和由立体视觉定义的（双眼单视的）运动具有一些共同特性，这表明这两种运动形式可能由相似的机制检测。对于由亮度定义的运动，至少有两个处理层次：先检测方向，然后由更高层次的专门处理机制提取速度。在此，我们测试了对于由立体视觉定义的运动是否也存在专门的速度处理机制。比较了仅包含由立体视觉定义的运动的刺激与同时包含由立体视觉定义的运动和由亮度定义的运动的刺激的速度辨别阈值。当刺激包含由亮度定义的运动时，增量阈值约为0.05 - 0.1。对于由立体视觉定义的运动，增量阈值从未优于0.3。通过仔细分析，有可能测试用于解决速度辨别任务的线索。结果表明，观察者对由立体视觉定义的运动是根据距离线索而非速度做出反应，这表明不存在处理由立体视觉定义的运动速度的专门机制。
