No pain scale is available for stroke patients due to the presence of language or cognitive disorders. However, the Faces Pain Scale (FPS), which was initially developed for children, has been used with success in adults with cognitive impairments. The aim of this study is to test whether the FPS could be used in left or right hemispheric stroke patients (LHSP, RHSP). One hundred twenty-seven stoke patients and 21 controls were recruited in 2 rehabilitation units. Construct validity of FPS was assessed by rating and ranking facial expressions. FPS was correlated to a Visual Analog Scale (VAS) and to a Verbal Rating Scale (VRS) for the assessment of shoulder pain. Reliability was determined by test-retest procedures. Performances of RHSP in the ranking and rating procedures were very poor compared to LHSP and to controls. However, in the assessment of patients' shoulder pain, FPS scores were highly correlated with VAS and VRS in both stroke groups (r=0.65-0.82, p<10(-3)). FPS was more reliable in LHSP than in RHSP. It was preferred to VAS and VRS in LHSP, while in RHSP VAS was the preferred scale. The present study provides preliminary support for the validity and the reliability of FPS in LHSP. However, we do not recommend its sole use in stroke patients. Further studies are needed to determine whether FPS can be used in stroke patients for assessing changes in severity of pain over time.