AIM: The Montreal Cognitive Assessment (MoCA), developed by Dr Nasreddine (Nasreddine et al. 2005), is a brief cognitive screening tool for detecting older people with mild cognitive impairment (MCI). We examined the reliability and validity of the Japanese version of the MoCA (MoCA-J) in older Japanese subjects. METHODS: Subjects were recruited from the outpatient memory clinic of Tokyo Metropolitan Geriatric Hospital or community-based medical health check-ups in 2008. The MoCA-J, the Mini-Mental State Examination (MMSE), the revised version of Hasegawa's Dementia Scale (HDS-R), Clinical Dementia Rating (CDR) scale, and routine neuropsychological batteries were conducted on 96 older subjects. Mild Alzheimer's disease (AD) was found in 30 subjects and MCI in 30, with 36 normal controls. RESULTS: The Cronbach's alpha of MoCA-J as an index of internal consistency was 0.74. The test-retest reliability of MoCA, using intraclass correlation coefficient between the scores at baseline survey and follow-up survey 8 weeks later was 0.88 (P < 0.001). MoCA-J score was highly correlated with MMSE (r = 0.83, P < 0.001), HDS-R (r = 0.79, P < 0.001) and CDR (r = -0.79, P < 0.001) scores. The areas under receiver-operator curves (AUC) for predicting MCI and AD groups by the MoCA-J were 0.95 (95% confidence interval [CI] = 0.90-1.00) and 0.99 (95% CI = 0.00-1.00), respectively. The corresponding values for MMSE and HDS-R were 0.85 (95% CI = 0.75-0.95) and 0.97 (95% CI = 0.00-1.00), and 0.86 (95% CI = 0.76-0.95) and 0.97 (95% CI = 0.00-1.00), respectively. Using a cut-off point of 25/26, the MoCA-J demonstrated a sensitivity of 93.0% and a specificity of 87.0% in screening MCI. CONCLUSION: The MoCA-J could be a useful cognitive test for screening MCI, and could be recommended in a primary clinical setting and for geriatric health screening in the community.