We demonstrated that mild cognitive impairment (MCI) participants of the ADNI database (N=640) can be discriminated into 3 coherent and neuropsychologically-defined subgroups. Our clustering approach revealed an amnestic MCI, a mixed MCI and a false positive subgroup. Furthermore, we investigated the neurobiological foundation of these automatically extracted MCI subgroups. Classification modelling exposed that specific predictive features can be used to differentiate amnestic and mixed MCI from healthy controls: CSF Aβ1-42 concentration for the former and CSF Aβ1-42 concentration, tau concentration as well as cortical atrophies (especially in the temporal and occipital lobes) for the latter. In contrast, false positive participants exhibited an identical profile to healthy participants in terms of cognitive performance, brain structure and CSF biomarker levels. Our comprehensive data-analytics strategy provide further evidence that multimodal neuropsychological subtyping is both clinically and neurobiologically meaningful.