Excess theta electroencephalographic (EEG) activity has been described as an accurate predictor for cognitive decline at least 7 years before symptom presentation. To test whether this predictor for cognitive decline correlates with structural changes in the brains of healthy elderly individuals, we compared the magnetic resonance structural images of healthy individuals with excess of theta activity [group with a risk for cognitive decline, risk group (RG); n=14] with healthy controls with normal EEG activity (control group; n=14). Neuropsychological and epidemiological analyses showed significant differences in only two features: more years of education and better performance in the visuospatial process task in the control group. Voxel-based morphometry results were not conclusive, but showed tendencies toward larger volumes in the prefrontal and parietal lobes, and smaller volumes in the right temporal lobe, right occipital lobe, and left cerebellum for the RG; these tendencies are in agreement with those proposed by the posterior–anterior shift in an aging model. Cortical-thickness analyses yielded a significant correlation between cortical thickness and years of education in the prefrontal and inferior-temporal regions, and larger cortical thickness in the RG, independent of age and years of education, in the right superior temporal region. These results suggest changes in the cortical thickness of structures related to memory and visuospatial functions in healthy, cognitively normal individuals before the appearance of cognitive decline. Thus, the performance of healthy elderly individuals with EEG risk may only be slightly different from normal because of compensation mechanisms allowing them to fulfill daily-life tasks, masking structural changes during preclinical neurocognitive disorders.