The biological mechanisms underlying the experience of psychosis are not yet known, impeding progress towards adequate treatments for people with psychotic disorders. Biomarker research could provide clues to the mechanism of psychosis, but current research approaches have largely failed to uncover strong, translational biomarkers of illness despite rapid technological developments. Emotional processing abnormalities are key components of psychosis phenomenology that remain poorly understood. Scene processing research using electroencephalography provides well-studied markers of neural emotional processing which could be useful biomarkers of socioemotional impairments in psychosis. This dissertation tests event-related potentials (ERPs) elicited by emotional scenes as biomarkers of psychotic illness and addresses 3 crucial concepts missing from the current literature: within-disorder heterogeneity, across-disorder homogeneity, and approaches to psychosis subtyping that do not rely on symptom categories. Studies leverage a large, well-characterized sample of individuals with psychosis for comprehensive analysis. Results indicate that research based on current diagnostic categorization greatly limits emotional biomarker discovery, though certain symptom dimensions such as cognitive ability and social functioning can be highly informative. Distinct emotional biomarkers of psychosis are identified and related to medications, symptomology, and psychosis subtype. Results invite further research into the relationship between emotional processing, cognition, and neural sources of activity to improve the treatment of socioemotional psychosis symptoms.