A purpose of present study is to provide basic information evaluating the utility of Magnetic Resonance imaging as a biological marker estimating manganese effects to central nervous system among welders, which is conducted by comparing urinary and blood manganese concentrations and signal intensities of brain MR images between exposed group and non-exposed group, evaluating the objectivity of subjective grading estimated by correlations between Pallidal signal intensity index (P. I) and subjective grades among exposed group, and comparing the difference of signal intensities according to presence of neurologic symptoms, signs and exposure variables among the exposed group. The exposed group is composed of 11 welders complaining severe symptoms or showing neurological signs, and the non-exposed group is composed of 5 patients who admitted a hospital. Urinary manganese concentrations and signal intensities in T1-weighted MR images among exposed group were higher than those of the non-exposed group significantly, which exhibits that increased signal intensities in T1-weighted MR image represent the effect of manganese exposure. P. Is among the exposed group revealed relatively high correlations with subjective grades ( gamma =0.63, p=0.037) , which suggests the objectivity of subjective grade. Signal intensity in globus pallidus was a suitable single variable representing the effect of manganese accumulation in C.N.S system appropriately, which was verified as follows ; Increased signal intensities among the exposed group had the highest frequency and intensity in the globus pallidus, and the P.I. had a relatively high correlation coefficient ( gamma 0.62, p=0.044) with total score of subjective grades. Signal intensity with subjective grading in globus pallidus represented very high correlation gamma =0.97, p=0.00) with total score of subjective grades, and had a similar correlation coefficient with many variables. It is hard to argue that signal intensities are markers representing pathologic change in C.N.S system or can be used as a diagnostic tool for manganese intoxication, because signal intensities had no difference between the exposed group and the non-exposed group according to presence of neurological signs.