PURPOSE To determine the feasibility of the use of temperature-sensitive magnetic resonance (MR) imaging for the detection of local temperature elevations at the focus of a low-power ultrasound beam in the brain. MATERIALS AND METHODS The brains in 28 rabbits were sonicated at acoustic power levels of 3.5-17.5 W. Four to five different locations were sonicated at different acoustic power levels in each rabbit. MR images were obtained 2 hours, 48 hours, 10 days, and 23 days after the sonications, depending on when the animals were sacrificed. Histologic evaluation of whole brain was performed. RESULTS Forty of 43 (93%) of the lowest-power (3.5-W) sonications were visible on temperature-sensitive MR images and did not result in any short- or long-term histologic or MR imaging evidence of tissue damage. A contrast-to-noise ratio of approximately 6 and a temperature elevation of 7 degrees-8 degrees C were observed. CONCLUSION Temperature elevations induced by means of focused ultrasound exposures that do not cause damage in the in vivo rabbit brain can be detected at temperature-sensitive MR imaging.