Abstract
Crowding, in which flankers impede the identification of an adjacent target, represents a significant deficit in form vision in the periphery. Crowding persists even after having compensated for the lower spatial acuity in the periphery. Little consensus exists among the theories of crowding. Palomares et al. (ARVO99) showed that while white noise was an effective masker when superimposed on a letter target, it was ineffective as a spatially separated flanker, challenging any contrast-masking theory for crowding. We (He & Tjan, VSS04) found evidence that “pink” noise, obtained by phase-scrambling a letter, was an effective flanker, which defies theories that rely on similarities between high-level features and imprecision in spatial attention. The purpose of the current study is to determine if pink letter noise are as effective as letter flankers by measuring the threshold contrast vs. spatial separation (TvS) function. Threshold contrast for identifying a letter target at 5-deg inferior visual field, of size 2.5 times the Ss' acuity, was measured as a function of spatial separation between the target and flankers. The flankers (at 20% rms contrast) were white noise, pink noise, or letters. The TvS functions for letter flankers and pink noise were virtually identical, both showed a threshold elevation of 0.6 log units (factor of 4) at a separation of 0.8 x-height which declined rapidly with a log-log slope of −1.4 to 0.05 log units at a separation of 2 x-height. In contrast, white noise elevated threshold by only 0.3 log units at a separation of 0.8 x-height, and its TvS function declined gradually with a log-log slope of −0.63. In a separate experiment, we showed that when white noise or pink noise was superimposed on a letter target, they led to the same threshold elevation. Our results show that a minimum requirement for an effective flanker is that it has a similar spatial-frequency distribution as the target but need not share any high-level features.