PurposeTo investigate the contributions of motor impairments to limitations in upper-limb function three months after stroke. DesignCross-sectional, exploratory study. MethodsDependent variable was upper-limb function, measured by the Motor Assessment Scale (MAS), which scores range from 0 to 18. Independent variables included measures of strength, dexterity, spasticity, and contracture of the paretic upper limb. Multiple linear regression analysis was employed to identify which of the independent variables could explain the MAS scores (p<0.05). Analysis was performed with the whole sample and with a sub-group of participants, who had high function (MAS≥12). ResultsSixty-nine individuals participated. Out of them, 63 had high upper-limb function. Regression analysis with the whole sample revealed that strength and dexterity were retained in the models. Together they explained 64% of the variance of the MAS scores (p<0.001), being strength the greater relative contributor. When the regression analysis included only participants with high upper-limb function, dexterity had a greater relative contribution, than strength. Together they explained 52 % of the variance (p< 0.001). ConclusionsStrength was the main contributor to upper-limb function in individuals three months after stroke. However, in individuals, who already had higher upper-limb function, dexterity showed to be the major contributor.