Despite the validity of clinical criteria for identifying kids at risk for psychosis, they are not predictive enough to guide important treatment decisions. Individuals at clinical high risk for psychosis (CHR-P) may benefit from improved outcome prediction if biomarkers can be identified. This study aimed to determine if young people with CHR-P syndrome have diminished amplitudes of an event-related potential known to be deficient in people with schizophrenia and to determine if this deficit is associated with adverse outcomes after controlling for the effects of antipsychotic medication use. Methods, Subjects, and Location As part of the North American Prodrome Longitudinal Study-2 (NAPLS-2), mismatch negativity (MMN) data were obtained from 8 university-based outpatient research programs. The initial MMN data were gathered between June 2009 and April 2013. There was a 24-month follow-up on the clinical results. Individuals with CHR-P syndrome and matched healthy controls were included. Patients with CHR-P syndrome who were monitored for 24 months were divided into 2 groups: those who developed psychosis (called “converters”) and those who did not (called “non converters”). The time frame for this analysis is December 2019 through December 2021. During a passive auditory oddball paradigm, electroencephalography readings were taken. The MMN was evaluated in response to duration-, pitch-, and duration-plus-pitch double-deviant tones. There were 247 women in the CHR-P group (n=580; mean [SD] age, 19.24 [4.39] years), while there were 114 women in the healthy control group (n=241; mean [SD] age, 20.33 [4.74] years). As a whole, 450 people (77.6%) in the CHR-P group weren’t using antipsychotics at the start of the study. Future CHR-P converters to psychosis (n=77, unmedicated n=54) had lower MMN amplitudes at baseline than non converters (n = 238, unmedicated n=190) in the complete sample (d=0.27) and the unmedicated subsample (d=0.33). Double-deviant MMN was lower in unmedicated converters compared with non converters (d = 0.43), demonstrating an interaction between baseline medication status, group, and deviant type. Additionally, deficiencies in double-deviant MMN were most strongly related to quicker conversion to psychosis within the unmedicated subsample (hazard ratio, 1.40 [95% CI, 1.03-1.90]; P=.03], which lasted beyond positive symptom severity). Associations were moderate, but this study indicated that MMN amplitude deficits were sensitive to subsequent psychosis conversion among patients at risk of CHR-P, especially those not taking antipsychotic medication at baseline. While MMN has little utility as a standalone biomarker of psychosis onset, it has the potential to add fresh risk information to multivariate prediction algorithms and provide a translational neurophysiological target for novel therapy development in a subgroup of at-risk patients.