Lower region and figure-ground assignment

Figure-ground assignment is an important visual process because figures underlie many visuomotor processes; humans recognize, attend, and act upon figures, not backgrounds. There are many visual cues that lead to figure-ground organization, notably the classic gestalt cues of area, symmetry, and convexity.

In this line of research we are exploring a new cue to figure-ground assignment called 'lower region.' Regions in the lower part of a stimulus array appear more figure-like than corresponding regions in the upper visual field. (See an example.) In a series of experiments, we have demonstrated that

  • 1. this lower-region preference is not due to luminance factors or eye movements,
  • 2. lower-field regions are perceived as figure for longer durations than upper-field regions and undergo fewer figure-ground reversals than control displays,
  • 3. voluntary visuospatial attention alone does not cause this lower-region figural preference, and
  • 4. "lower region" is defined relative to the stimulus display.

We discuss our results in terms of the environmental cues that this figure-ground cue may reflect.

Continuing work on this project uses a match-to-sample visual working memory task to study the lower region cue. We have found that lower region is a stronger cue for figural assignment than some of the original gestalt cues, such as convexity.

Finally, we have been exploring the reference frame of figure-ground assignment using the lower region cue. Lower region, and perhaps figure-ground generally, appears to be processed in a viewer-centered representation: When viewers observe lower-region displays with a tilted head or view these displays upside down, the lower-region preference follows the head orientation. We have a paper on this reference frame effect available here.