B a c k g r o u n d : Spatial analysis of EEG is a phenomenal assessment and not so informative for phase space anddynamic aspect of EEG data. In contrast, nonlinear EEG analysis attempts to characterize the dynamics of neural networksin the brain. We have analyzed the features of EEG nonlinearly in subjects with white matter change on brainMRI and compared the results with cognitive function in each.Methods : Digital EEG data were taken for 30 seconds in 9 subjects with white matter degeneration and in 5 healthynormal controls without white matter change on MRI. Then we analyzed them nonlinearly to calculate the correlationdimension(D2) using the MATLAB software. The cognitive function was assessed by 3MS(modified mini-mental stateexamination). The severity of white matter change was assessed by Scheltens scale.Results : The mean D2 value of normal control was greater than that of white matter degeneration group. The D2s ofsome channels were correlative with 3MS and degree of white matter degeneration significantly.Conclusions : nonlinear analysis of EEG can be used as one of adjuvant functional studies for prediction of cognitiveimpairment in subjects with white matter degeneration and subcortical white matter change can be influential on cognitivefunction and correlation dimension of EEG.