Aims: To develop a system of digital analysis of Magnetic Resonance Imaging images which allows a comparative measure of symmetric muscular areas. Material and methods: Images DICOM from de lumbar region of patient with low back pain and sciatica were used to study the multifidus muscles. The images were processed with Osiris and NIH image programs. The data obtained was analysed by program routines in Visual Basic, and with MS Excel. The image analysis was made with a region of interest, and its histogram. By this method we can distinguished muscle and fat, fixing a grey density level which can separate them and be able to measure areas of each one, proportions of both densities and to stablish the average value of the histogram. Conclusions: This method is useful for determination of differences in symmetric muscular masses, such as cross sectional areas, and quality of muscles