BACKGROUND: Cognitive dysfunctioning (CDF) is an important issue in stroke, interfering with recovery and social dysfunctioning. We aimed to investigate the clinical and imaging correlates of CDF in patients with a first-ever subacute ischemic stroke and no dementia. METHODS: We evaluated CDF 15 days after stroke in a prospective cohort of consecutive patients with a Mini Mental State Examination score > or =23 using a comprehensive neuropsychological battery. CDF was ranked into 3 categories according to Z scores calculated for each test and adjusted for age and education. CDF was analyzed in relation to stroke features. Imaging was assessed using MRI. An ordinal regression procedure was used to determine the clinical correlates of CDF and to compute probabilities. RESULTS: Cognitive evaluation was achieved in 177 consecutive patients (age 50.0 +/- 16.0 years). In bivariate analysis, CDF was associated with age, low level of education, depression, neurological deficit at day 15, stroke subtype, arterial territory and leukoaraiosis but not with stroke volume or location. The predictors of CDF were NIHSS score at day 15 (OR = 1.35; 95% CI = 1.05-1.73), middle cerebral artery infarct (OR = 2.96; 95% CI = 1.30-6.73), depression interacting with left stroke side (OR = 1.09; 95% CI = 1.03-1.15), and female gender interacting with high level of education (OR = 0.209; 95% CI = 0.085-0.514). CONCLUSIONS: Stroke features correlate with CDF in nondemented patients. These simple criteria may help to predict CDF at bedside in the subacute phase after stroke and to recommend a neuropsychological evaluation for patients' management. Modeling CDF soon after stroke using simple neurological criteria may be a useful tool for designing clinical trials.