ABSTRACTBackground:Many stroke research trials do not include assessment of cognitive function. A Very Early Rehabilitation Trial (AVERT) is an international multicenter study that includes the Montreal Cognitive Assessment (MoCA) as an outcome. At the Malaysian AVERT site, completion of the MoCA has been limited by low English proficiency in some participants. We aimed to develop a Bahasa Malaysia (BM) version of the MoCA and to validate it in a stroke population.Methods:The original English version of the MoCA was translated into BM and then back-translated to ensure accuracy. Feasibility testing in a group of stroke patients prompted minor changes to the BM MoCA. In the validation phase, a larger group of bilingual stroke patients completed both the original English MoCA and the finalized BM MoCA, with presentation order counter-balanced.Results:Forty stroke patients participated, with a mean age of 57.2 (SD = 10.3). Agreement between BM MoCA and English MoCA was strong (intra-class correlation coefficient = 0.81, 95% CI 0.68–0.90). Scores on BM MoCA were slightly higher than scores on English MoCA (median absolute difference = 2.0, IQR 0–3.5), and this difference was present regardless of which version was completed first.Conclusions:The existence of a validated BM version of the MoCA will be of major benefit to clinicians and researchers in Malaysia and the wider South-east Asian region, where the Malay language is used by over 200 million people.