Improving the ability to detect Alzheimer's disease (AD) at the earliest stages is essential to effectively treat afflicted individuals. Electrophysiological signatures are a promising avenue for earlier diagnosis. In the present study, we investigated an ERP component associated with visual working memory capacity, the contralateral delay activity (CDA). Our participants were undiagnosed and supposedly healthy members of the community, but were defined to be at-risk for Mild Cognitive Impairment (MCI) based on performance from a brief, standardized neuropsychological test. We found that older adults at-risk for MCI had a reduced visual working memory capacity and reduced differentiation of the CDA. In a second experiment, we found that the P300, a well-characterized ERP component shown to be useful in determining conversion from MCI to AD, showed reduced amplitude in our at-risk group. Together, these findings suggest that electrophysiological signatures may be especially sensitive markers of the very earliest stages of AD.