Objective: Left atrial (LA) and left ventricular (LV) remodelling may lead to stroke. The aim of this study was to analyze LA function and LV strain in patients with embolic stroke of undetermined source (ESUS). Materials and Methods: This prospective study included 35 ESUS patients and 37 age and sex-matched controls. All participants underwent brain computed tomography (CT), conventional and diffusion-weighted magnetic resonance imaging (MRI), CT or MR angiography, 12 lead ECG, transthoracic echocardiography, and 48 hour Holter ECG monitoring. LA volume and function were determined by echocardiography. LA reservoir and LV strains were measured longitudinally by speckle-tracking method. CHA2DS2-VASc, The National Institutes of Health Stroke Scale (NIHSS), modified Rankin Scale (mRS) scores were calculated. Results: Major cardiovascular risk factors were similar between thetwogroups. The mean CHA2DS2-VASc score was 2.6 ± 1.2. NIHSS was 3.9 ± 3.0 and mRS was 1.3 ± 0.8. Atrial electromechanical coupling intervals and delays, LA emptying fraction and volumes were similar between the two groups. LA reservoir strain was lower than controls (25.2 ± 7.2% vs. 29.7 ± 8.8%, p=0.019). LV global longitudinal strain was lower than controls (-14.7 ± 4.2% vs -16.4 ± 3.9%, p=0.031). There was no correlation between LA, LV strains and the scores (CHA2DS2-VASc, NIHSS, mRS). Conclusions: ESUS patients had lower LA reservoir and LV longitudinal global strains than controls. Left atrial volume index, LA emptying fraction did not differ between the two groups. Echocardiographic quantification of LA and LV remodelling has great potential for secondary prevention from ESUS. Further studies are needed to confirm our findings.