Modeling spatial and temporal aspects of visual backward masking.

Abstract

Visual backward masking is a versatile tool for understanding principles and limitations of visual information processing in the human brain. However, the mechanisms underlying masking are still poorly understood. In the current contribution, the authors show that a structurally simple mathematical model can explain many spatial and temporal effects in visual masking, such as spatial layout effects on pattern masking and B-type masking. Specifically, the authors show that lateral excitation and inhibition on different length scales, in combination with the typical time scales, are capable of producing a rich, dynamic behavior that explains this multitude of masking phenomena in a single, biophysically motivated model.

Grouping in the shine-through effect

“…Interestingly, a three-element grating yields better results than does a five-element grating, even though the edges of the three-element grating are closer to the vernier. Computer simulations show that this result can be explained well by the spatial width of the lateral interactions (Hermens et al, 2008). Hence, the first experiment showed fast and spatially complex processing in accordance with the first hypothesis.…”

“…On a neural description level, we showed with computer simulations that a first step of this grouping may indeed be realized by dynamic lateral interactions (Hermens et al, 2008;. This processing may well occur at a very early neural processing stage, such as in the primary visual cortex V1, and does not require any specialized grouping modules.…”

“…Computer simulations showed indeed that an enhanced activity corresponding to the representation of edges can dynamically suppress the activity corresponding to that for the vernier when an edge is close to the vernier (Figure 1, lowest row;Hermens, Luksys, Gerstner, Herzog, & Ernst, 2008;Herzog, Ernst, Etzold, & Eurich, 2003; see also Zhaoping, 2003). For extended homogeneous gratings, shine-through occurs, because the edges are remote and the inner grating elements are suppressed.…”

Object substitution masking and its relationship with visual crowding

“…Given this the masking effect observed tends to reflect both the temporal properties of the mask as well as the spatial locus of distractor elements. In this respect our results are consistent with other claims regarding the close interdependence of spatial and temporal factors in masking, and in visual processing more generally (Enns, 2004;Herzog, 2007;Hermens, Luksys, Gerstner, Herzog & Ernst, 2008;Ghose, Hermens & Herzog, 2012;Lleras & Moore, 2003;Lev & Polat, 2015;Yeshurun, Rashal & Tkacz-Domb, 2015). …”

Spatial grouping determines temporal integration.

“…Note that for this excellent data fit, we used only two free parameters (see appendix). The remaining model parameters correspond to those used before (Hermens et al, 2008;. Figure 4 illustrates how the neural network model explains three important conditions in Experiment 1.…”