Encoding-related neural activity before word onset predicting later recollection in the high-reward condition. A, Group-averaged ERP waveforms elicited by high-reward cues at the three midline electrode sites (equivalent to Fz, Cz, and Pz of the international 10/10 system). Waveforms are overlaid according to whether the following word was recollected (given a remember judgment), confidently judged as familiar (given a confident old judgment), or forgotten (given a new judgment) in the subsequent recognition test. The waveforms are based on the subset of participants with at least 13 trials in each category. Positive values are plotted upward. Activity preceding later recollected words differs from activity preceding confidently familiar and forgotten words. B, Range-scaled two-dimensional voltage spline maps illustrating the positive-going, widespread scalp distribution of the recollection-related subsequent memory effect before word onset. Shown are the amplitude differences between ERPs for later recollected and forgotten words in the 300–600, 600–1100, and 1100–2000 ms periods in the cue-word interval. Note that these maps were computed across all participants with sufficient numbers of recollected and forgotten trials to maximize the signal-to-noise ratio, not just the subset shown in A.