The regulator of G-protein signaling 4 (RGS4) has been suggested as a candidate gene for schizophrenia. However, following an initial positive report, subsequent association studies between RGS4 and schizophrenia have yielded inconclusive results. Also, few studies have investigated the association of RGS4 polymorphisms with the phenotypic subgroups of schizophrenia. To further clarify the role of RGS4 in this disease, we performed a case-control study (504 cases and 531 controls of Han Chinese descent) to examine the association of RGS4 with schizophrenia and with clinical and neurocognitive profiles. The four markers (SNPs 1, 4, 7, and 18) implicated in the original association study were genotyped. We detected significant association of four-marker haplotypes with schizophrenia (UNPHASED: global P = 0.037; PHASE: global P = 0.048). The haplotype G-G-G-G, which was implicated in at least three previous studies, was the major risk haplotype (UNPHASED: P = 0.019; PHASE: P = 0.010). Regarding the clinical phenotypes, the Wechsler Adult Intelligence Test (WAIS) information subtest score was associated with SNP4 genotypes (P = 0.001). PANSS total and global psychopathology scores were also associated with SNP4, but may not reliably reflect the general severity of disease as the scores may be affected by confounders like medication response. Our study provides further support for a role of RGS4 in the pathogenesis of schizophrenia. We identified G-G-G-G as the risk haplotype in our Chinese sample. The association with information subtest score suggests an effect of RGS4 on premorbid functioning, which may be related to neurodevelopmental processes. Further independent studies are required to verify our findings. © 2007 Wiley-Liss, Inc.