||Previous studies show increased antibody levels to bovine casein in some individuals with schizophrenia. The immunogenicity of specific domains of bovine casein varies among people with milk sensitivities and thus could vary among different neuropsychiatric disorders. Using ELISAs and immunoblotting, we characterized IgG class antibody specificity to whole bovine casein and to the as, ß, and ¿ subunits in individuals with recent onset psychosis (n = 95), long-term schizophrenia (n = 103), and non-psychiatric controls (n = 65). In both patient groups, we found elevated IgG to casein proteins, particularly to whole casein and the as subunit (p = 0.0001). Odds ratios of casein seroprevalence for recent onset psychosis (age-, gender-, race-, smoking-adjusted) were significant for whole casein (8.13, p = 0.0001), and the as (7.89, p = 0.0001), ß (5.23, p = 0.001) and ¿ (5.70, p = 0.0001) subunits. Odds ratios for long-term schizophrenia were significant for whole casein (7.85, p = 0.0001), and the as (4.78, p = 0.003) and ¿ (4.92, p = 0.004) subunits. Within the recent onset group, odds ratios were particularly significant for a subgroup of people with psychotic disorders that included major depressive disorders (8.22–16.48, p = 0.0001). In a different recent onset subgroup (schizophrenia-spectrum disorders), PANSS scores for negative symptoms were correlated with casein antibody levels for the as and ¿ subunits (p = 0.001–0.01). Immunoblotting patterns also exhibited group specificity, with ¿ predominant in recent onset and as in schizophrenia (Fisher's Exact Test, p = 0.001). The elevated IgG and unique patterns of antibody specificity to bovine casein among diagnostic groups provide a rationale for clinical trials to evaluate efficacies of dietary modifications in individuals with neuropsychiatric diseases.