Abstract
Background: Spatial association fields (SAFs) describe perceived grouping strength between texture elements as a function of their joint position and orientation/direction. However, SAFs have only been measured in the “sustained” regime with cues such as static or constant-motion elements. Here, I investigated properties of SAFs in the “transient” regime using harmonic motion and harmonic contrast.
Methods: Stimuli consisted of two texture elements, each of which was composed of a static grating and a dynamic (5-Hz counterphasing) grating. Manipulating the spatial phase of the static grating produced time-varying oscillations either in motion or in contrast while sparing properties of the dynamic grating. Observers performed a two-interval forced-choice task and reported the element pair with greatest harmonic coherence (i.e., maximum synchrony). Coherence thresholds were measured at the 75%-correct level for all combinations of 3 inter-element distances, 3 orientation/direction angles for the first element, and 8 orientation/direction angles for the second element.
Results: When both elements underwent harmonic motion, the SAF was highly direction selective and tuned for direction similarity regardless of relative element position. When both elements underwent harmonic contrast, the SAF was nonselective for orientation. For both harmonic types, grouping strength decreased with inter-element distance. When elements of a pair underwent different harmonic types, observers were strictly unable to judge element coherence. This finding is surprising given that dynamic gratings are physically unaffected by harmonic type.
Conclusions: Unlike grouping across sustained cues, grouping across transient cues requires exquisite sensitivity to the relative timing of events across the visual field. Correspondingly, SAFs for transient and sustained cues have markedly different properties that may reflect distinct underlying neural mechanisms. The inability of observers to integrate across harmonic motion and harmonic contrast supports the notion of distinct grouping pathways in the transient-cue domain.
This work was supported by COBRE grant P20 RR20151-02 from the National Institute of Health (NEI) to the Center for Visual Neuroscience at North Dakota State University.