Rationale : Methamphetamine (MA) dependence frequently co- occurs with HIV infection and increases the likelihood of poorer outcomes, including neurocognitive impairment. One aspect of cognition affected in both HIV and MA and may be a target for intervention is prospective memory (PM), or remembering to perform an intended action at a specific time. Self-generation is a potentially powerful approach to improve PM, hypothetically by supporting strategic aspects of encoding and cue monitoring that underlie PM impairment in HIV/MA individuals. This dissertation project sought to: 1) evaluate the efficacy of self- generation on PM in HIV/MA adults on basic, clinical, and naturalistic PM paradigms; 2) determine neurocognitive mechanisms of self-generation on PM in HIV/MA; and 3) explore potential correlates of self-generation on PM in HIV/MA. Design : Using a 3-group design (N=83), this study examined the effectiveness of self-generation in HIV+MA- and HIV+MA+ individuals compared to healthy comparison individuals (HIV-MA-) across three PM tasks of varying ecological validity. Participants were also administered a structured psychiatric interview, neuropsychological battery, and neuromedical examination. Results : Across tasks utilizing within-subjects designs (basic, clinical PM tasks), participants were more likely to respond to the PM cue at recall and recognition for self-generated cues relative to didactically-learned cues (ps<0.05; Hedge's gs=0.28-0.53); group and interaction factors were non- significant (ps>0.05). On the basic PM task, participants responded more quickly to the ongoing lexical decision- making task during the self-generated PM trials relative to the didactic trials (p=0.033; Hedge's g=0.18). Neurocognitively-impaired HIV+ participants experienced greater benefit from self-generation on the basic PM task relative to their neurocognitively-normal counterparts (p=0.008; Hedge's g difference=0.42). No main or interaction effects were observed on the naturalistic PM task, but longitudinal analyses revealed a trend-level effect of self-generation in the combined clinical sample. Across all tasks, no neurocognitive, psychiatric, HIV disease, or MA use characteristics were associated with self-generation benefit. Relevance : This study establishes potential effectiveness of self-generation to improve PM in healthy and HIV-infected adults, yielding effects in the small-to-medium range. Future research should seek to expand test mechanisms to health-related behaviors as well as the utility of a combined approach with manipulations designed to enhance PM cue monitoring