The ability of functional magnetic resonance (MR) imaging to detect a selective sensorimotor cortex activation in healthy subjects and the feasibility of motor activation in patients with lesions around the central sulcus were investigated. Twenty-five healthy volunteers performed 100 motor activation trials, using a variety of motor tasks, which were monitored by several image analysis methods. The functional images were obtained using a 1.5-tesla standard MR imaging system magnet with blood oxygenation level-dependent contrast. Four patients were assessed using functional MR imaging and invasive cortical mapping. Rolandic cortex activation was observed in 98% of the trials performed on healthy subjects in which no head motion occurred. Nevertheless, the cortical response was not selective in a task-rest analysis due to concurrent activation of neighboring regions. Across-task comparison analyses were useful in cancelling nonrelevant activity in most cases (86%). In the patient group, the region identified as the sensorimotor cortex by invasive means corresponded accurately to the area that was activated in functional MR imaging. Present data support the feasibility of detecting selective activation of the rolandic cortex, even in the clinical setting, leading the authors to suggest the usefulness of this widely available technique in surgical planning.