Dresp Birgitta, Langley Osmund Keith
Centre d'Ecologie Fonctionnelle et Evolutive, UMR 5175 CNRS, Université de Montpellier, 1919 Route de Mende, 34293 Montpellier, France.
Vision Res. 2005 Feb;45(3):275-84. doi: 10.1016/j.visres.2004.08.018.
It is widely assumed that the integration of orientation contrast across spatial gaps within the long-range regime is not selective to the contrast sign of the individual stimuli. Probabilistic models of perceptual integration, however, suggest that long-range spatial integration should be, if not selective, at least sensitive to local contrast signs. To clarify this issue, we tested predictions of a model based on conditional probabilistic weights of identical and opposite contrast signs in a simple spatial configuration of two co-linear lines. Contrast detection thresholds of the target line presented either by itself (control condition) or simultaneously with the co-linear inducer (test condition) were measured. The contrast sign of targets and inducers was varied so that all four possible combinations of signs were produced in the test conditions: (1) dark target with dark inducer, (2) dark target with bright inducer, (3) bright target with bright inducer and (4) bright target with dark inducer. The contrast intensity (Weber ratio) of dark and bright inducers was identical. The coaxial distance between target and inducer was constant in each of two experiments, testing for two distances that corresponded to an angular separation within the long-range domain of spatial integration as defined previously. It is found that targets and inducers with identical contrast signs produce significantly stronger facilitating effects on detection than stimuli with opposite signs. The data closely match predictions consistent with those of a probabilistic model of line contrast integration across spatial gaps and contrast signs within the long-range regime.
人们普遍认为,在远距离范围内跨空间间隙的方向对比度整合对单个刺激的对比度符号没有选择性。然而,感知整合的概率模型表明,远距离空间整合即使没有选择性,至少也应对局部对比度符号敏感。为了澄清这个问题,我们在两条共线的简单空间配置中,测试了一个基于相同和相反对比度符号的条件概率权重模型的预测。测量了单独呈现的目标线(对照条件)或与共线诱导物同时呈现时(测试条件)的对比度检测阈值。目标和诱导物的对比度符号有所变化,以便在测试条件下产生所有四种可能的符号组合:(1) 暗目标与暗诱导物,(2) 暗目标与亮诱导物,(3) 亮目标与亮诱导物,以及 (4) 亮目标与暗诱导物。暗诱导物和亮诱导物的对比度强度(韦伯比率)相同。在两个实验中,目标和诱导物之间的同轴距离是恒定的,测试了两个对应于先前定义的远距离空间整合范围内角度分离的距离。结果发现,具有相同对比度符号的目标和诱导物对检测产生的促进作用明显强于具有相反符号的刺激。这些数据与远距离范围内跨空间间隙和对比度符号的线对比度整合概率模型的预测密切匹配。