--- - |2 Abstract Cognitive affective neuroscience tasks that are straightforward to administer, measure key constructs of interest, and can be used in different lab settings and with multiple psychophysiological methods can lead to a more complete understanding of experimental effects. The no‐threat, predictable threat, unpredictable threat (NPU‐threat) task assesses constructs of interest to both clinical and basic affective science literatures, is relatively brief to administer, and has been used across labs with a number of different measurements (e.g., startle eyeblink, fMRI, corrugator response, subjective ratings). ERPs provide another means of assessing neurobiological reactivity during the NPU‐threat task, but to date such measures have been underutilized. That is, no study has yet evaluated cue‐elicited ERPs in the NPU‐threat task. Here, cue‐elicited ERPs were assessed in 78 participants who completed a version of the NPU‐threat task previously shown to reliably moderate startle eyeblink amplitudes. Results showed larger P2 amplitudes for unpredictable versus predictable trials, increased P3s and late positive potentials for threatening versus no‐threat trials, as well as larger stimulus preceding negativities for threatening versus no‐threat trials (driven primarily by predictable threat cues). In line with prior work, we observed enhanced startle eyeblink for threatening versus no‐threat trials and for unpredictable compared to predictable threat interstimulus intervals. In addition, the probe‐elicited P3 was suppressed for predictable and unpredictable compared to no‐threat trials. Therefore, cue‐elicited ERPs, which can be recorded alongside other measures in the NPU‐threat task (e.g., startle), may provide useful indices of temporally distinct stages of predictable and unpredictable threat processing. - Psychophysiology, Volume 55, Issue 10, October 2018.