Visual spatial interactions between dotted line segments.

-This study examines the interactions that occur between dotted line segments as a function of their separation and orientation using a dotted target detection paradigm. The results of the experiment indicate that it is the separation between interacting pairs of lines rather than shifts in retinal locus that account for the observed decline in detectability as the separation increases. However, factors of global organization or figural goodness seem also to exert a powerful influence on the detectability of the pattern. The results are interpreted in terms of a central, rather than a peripheral, model. One of the most often enunciated generalizations of visual perception is that, compared to the presentation of a single part, multiple parts of a stimulus presented to different retinal loci interact in ways that can either enhance or reduce the perception of the whole stimulus. Many types of spatial interaction have been discovered, including reciprocal inhibition and spatial summation. Contour enhancement is usually attributed to the former (see Ratliff, 1965, for a summary), decreases in detection threshold, to the latter (see Kristofferson and Dember, 1958, for an example). The present study inquires into a form of spatial interaction that seems to fit into neither of these categories but is closer to the organizational interactions considered in the older Gestalt literature. In this paper, we are dealing with a form of spatial interaction that occurs over long distances and that is also sensitive to the cognitive quality, or figural goodness, of the pattern. To explore this new form of spatial interaction, we determine the detectability of two colinear dotted-line segments as a function of their separation and orientation with respect to each other. The task requires that these target patterns be detected in a visual noise field composed of a number of randomly positioned dots-a paradigm that has proved to be quite powerful in assaying the perceptual effects of geometrysensitive pattern recognition tasks of this sort A full description of the logic of this procedure is given by Uttal (1975). The results of the present study indicate that, although an increase in separation between two dotted-line segments results in a de-crease in detection when the two lines remain colinear, there is a more complex effect of separation when factors that we may call in the Gestalt tradition, ‘Yigural quality” are varied simultaneously. In the latter situation, some of the separation effects can be overridden, indicating that detection performance seems to be highly dependent on the global organization of the fi_yre. This result has substantial implications for theones of pat-