Abstract
Objects that appear close in space or time can interfere with the perception of each other — the crowding effect. Much of our understanding of the crowding effect is based on the scenario when the target of interest and its flankers are static and appear simultaneously. Here, we propose a mechanistic model based on feed-forward interactions between sustained and transient channels to account for the crowding effects observed when the spatial or temporal properties of the target and flankers differ. The premise of the model is that the appearance of a visual stimulus would generate signals in the sustained and transient channels with different latencies, and crowding arises when the sustained signals from the target, which presumably code for target identity, are inhibited by the sustained and/or transient signals from the flankers. Thus, the crowding effect would depend on the relative timing of the onset of target and flankers. Further, our model includes separate pathways with different latencies for static and moving stimuli. We tested this model by obtaining simulated crowding functions for several conditions for which we had empirical data from human observers: (1) static spatial condition - target and flankers were static and appeared simultaneously; (2) static temporal condition - target and flankers were static and were presented with a target-flanker onset asynchrony; (3) motion condition - static target with moving flankers; (4) static unmatched condition - target and flankers were static and were presented for unmatched durations. In general, simulated results from the model (peak crowding magnitude, target-flanker onset asynchrony at peak crowding, and the extent of crowding) matched the empirical results very well, suggesting that interactions between sustained and transient signals and differential latencies of the pathways used to generate these signals can explain the crowding effects under various spatio-temporal conditions between the target and flankers.