Memory effects from cognitive control: A stage-specific account of desirable difficulty

Abstract

This thesis investigates predictions from prominent conflict theories of cognitive control that information experienced under high conflict conditions should be better encoded. More specifically, recent research suggests that selectively attending to relevant stimuli while ignoring conflicting stimuli can lead to better memory. These ideas have been broadly discussed in the desirable difficulty literature – described by instances where increasing difficulty during initial task performance leads to better later memory. As a growing number of studies have attempted to produce these effects with mixed success, calls for more focused investigations into the underlying mechanisms have been made. This encoding benefit for high-control-demand or high-difficulty situations has been broadly conceptualized as a task-general property, where all activated representations should be better encoded. The goal of this thesis was to investigate whether memory-enhancing effects of difficulty manipulations depend on inducing additional cognitive control at particular information processing stages. This thesis documents some of the first work showing that the within-task locus of conflict and attentional control is critical to whether later memory benefits are seen – conflict/control focused on semantic item representation produces better memory, but conflict/control focused away from item representations at response selection gives no memory benefit. These findings and theory are then extended to physiological measures of pupil dilation and sequential (Grattron-like) conflict/control situations. This thesis proposes a stage-specific conflict-encoding model which complements and extends current leading theories of conflict-driven cognitive control.