A memory for goals model of sequence errors

A model of routine sequence actions is developed based on the Memory for Goals framework. The model assumes that sequential action is guided by episodic control codes generated for each step, and that these codes decay with time and can be primed by contextual retrieval cues. These control codes serve a place-keeping function that allows the system to infer the correct next action after performance is interrupted. According to the model, perseveration (repeat) errors occur because an older episodic trace intrudes due to noise in the system. Anticipation (skip) errors occur because of failures in reality monitoring, in which the model believes that it has completed a step it has not. The model predicts that perseveration errors should occur more frequently than anticipation errors, and that perseveration errors should occur in a graded fashion away from the current step. Across two different experiments, these predictions were supported at both a qualitative and a quantitative level.