Fractal analysis has played an important role in various fields such as physics, biology and medicine. Recently, multifractal analysis based on generalized concepts of fractals has been applied to biological tissues composed of complex structures. Deep white matter hyperintensity (DWMH) on brain magnetic resonance imaging (MRI) is more often observed in patients with geriatric depression than in healthy elderly subjects, and its clinical significance is receiving attention. We applied multifractal analysis to white matter images on brain T2-weighted MRI in 62 patients with geriatric depression (50-75 years). The local fractal dimensions, alpha(max) and alpha(min), which serve as indices of complexity, and their difference, alpha(max) - alpha(min), were closely correlated with the macroscopic grading according to Fazekas classification, suggesting that multifractal analysis is useful for quantitative evaluation of DWMH on MRI.