INTRODUCTION: Stroke is often associated with cognitive deterioration (CD) in varying degrees, and the risk factors associated to CD after a stroke vary from one study to another. AIMS: To determine the frequency of CD following cerebral infarction and to identify the factors that account for its presence. PATIENTS AND METHODS: A descriptive study was performed involving 126 stroke survivors who were hospitalised for rehabilitation. Patients were classified as having CD if they had scores below the cut-off point in one of the cognitive tests that were applied, i.e. the Folstein Minimental Test (< 24) or the Command-Condition Clock Test (< 7). The explanatory variables were: demographic variables (age, sex and years of schooling), comorbidity (arterial hypertension, diabetes mellitus and intelligence quotient), noxious habits (smoking and drinking), variables related to the lesion (extension, hemisphere and location), and others related to the consequences of the stroke (neurological status and depression). Bivariate and logistic regression analyses were performed to determine the role of these variables in CD. RESULTS: CD was observed in 51.6% of the patients. In the bivariate analysis, the extension of the infarction, age, years of schooling, degree of neurological compromise and depression were associated with the presence of CD. The logistic regression model showed that the factors explaining CD were severe neurological compromise (OR = 22.9; CI 95% = 4.2-125.2), having major depression (OR = 2.9; CI 95% = 1.14-7.8) and older age (OR = 0.94; CI 95% = 0.89-0.98). CONCLUSIONS: A little more than half of all stroke survivors, who are undergoing rehabilitation in hospital, have CD. The factors with the greatest explanatory power to account for CI were neurological status, depression and age.