We examined central nervous system [CNS] stress responses among deprived and continuing heavy marijuana users and nonusers. Participants (N = 210; 46.7% female; Mage = 21.99; 91.4% White, 94.3% Non-Hispanic) were heavy marijuana users (N = 134) and nonusers (N = 76). Heavy users were randomly assigned to a 3-day marijuana deprivation condition (N = 68) or to continue using regularly (N = 66). Participants completed two threat-of-shock stressor tasks that manipulated stressor predictability by varying shock probability or timing. We measured central stress responses via startle potentiation (stressor conditions minus matched no-stressor condition). We examined two group contrasts (heavy use: all heavy users vs. nonusers; deprivation: deprived vs. continuing heavy users) on startle potentiation overall and moderated by stressor predictability (unpredictable vs. predictable). Deprivation did not affect startle potentiation overall (timing task: p = .184; probability task: p = .328) or differently by stressor predictability (timing task: p = .147; probability task: p = .678). Heavy use did not affect startle potentiation overall (timing task: p = .213; probability task: p = .843) or differently by stressor predictability (timing task: p = .655; probability task: p = .273). Posthoc analyses showed mixed evidence of general startle reactivity × deprivation interaction on startle potentiation overall (timing task: p = .019; probability task: p = .056) and differently by stressor predictability (probability task: p = .024; timing task: p = .364). A history of marijuana use or acute deprivation did not alter central stress responses despite prominent theoretical expectations. This study adds to growing research on central stress responses in individuals with a history of drug use and begins to parse moderating roles of individual differences and stressor characteristics. (PsycInfo Database Record (c) 2022 APA, all rights reserved).