Building on our previous implementation of a word serial position task (Kayser et al., 2006), the WM paradigm consisted of two serial position tests (27 trials per block, 4 blocks per task, 216 total trials) with the same word or face stimuli used in the RM paradigm. In each trial, a series of three words (or faces) was presented (500 ms, 1.5-s stimulus onset asynchrony) and, after a 3-s delay, one of the items was repeated (probe). Participants indicated the position of the probe item during the encoding series by pressing one of three buttons after a fixed 1.5-s delay.
