Mark W. Becker¹, Morgan R. Dodd¹, Andrew Rodriguez¹; ¹Michigan State University

There is controversy about whether feature-based attentional suppression can be applied proactively or only occurs retroactively. To address this debate, we had participants search for a Landolt C with a horizontal break either on the right or left among Cs with vertical breaks. Each C appeared on a colored square. Search arrays were preceded by a cue and participants were informed that the target C would never appear on a square of the cued color. There were three cuing conditions: a negative cue - ¼ of the squares in the array appeared in the cue’s color and the target never appeared in these squares; a neutral cue - the cue’s color did not appear in the array; and a “reduced set size” cue - the cue’s color did not appear in the array but the size of the array was reduced by ¼, providing a boundary condition for complete suppression by a negative cue. In addition, for one set of participants, the colored squares appeared two seconds before the C’s appeared within them, which allowed us to investigate retroactive suppression. For other subjects, the colored squares and Cs appeared simultaneously, which allowed us to investigate proactive suppression. Results showed a clear suppression benefit, but only in the retroactive timing condition. In a follow-up, we varied the delay between the onset of the color squares and the Cs across four conditions (0, .5, 1, or 2 seconds). We again showed evidence for suppression, but only in the 2 second condition. These results suggest that suppression can be used on a trial-by-trial basis, but only retroactively. In addition, the process of setting up effective retroactive suppression seems extremely slow – making it unlikely to be used to guide attention in most real-world situations.