Learning to selectively attend from context-specific attentional histories: A demonstration and some constraints

Multiple lines of evidence from the attention and performance literature show that attention filtering can be controlled by higher level voluntary processes and lower-level cue-driven processes (for recent reviews see Bugg, 2012; Bugg & Crump, 2012; Egner, 2008). The experiments were designed to test a general hypothesis that cue-driven control learns from context-specific histories of prior acts of selective attention. Several web-based flanker studies were conducted via Amazon Mechanical Turk. Attention filtering demands were induced by a secondary one-back memory task after each trial prompting recall of the last target or distractor letter. Blocking recall demands produced larger flanker effects for the distractor than target recall conditions. Mixing recall demands and associating them with particular stimulus-cues (location, colour, letter, and font) sometimes showed rapid, contextual control of flanker interference, and sometimes did not. The results show that subtle methodological parameters can influence whether or not contextual control is observed. More generally, the results show that contextual control phenomena can be influenced by other sources of control, including other cue-driven sources competing for control.