Quantitative magnetoencephalography (qMEG) was used to investigate differences in the 2 to 60 Hz spectral power, between Alzheimer disease (AD) patients and control subjects. Twenty-two AD patients and 21 age-matched control subjects participated in this study. MEG signal analysis comprised the division of the entire 2 to 60 Hz spectrum in 2 Hz-width subbands. Both the relative power and the contribution of each subband to the correct classification of AD patients and controls were calculated. The relative power in 2 bands comprised between 2 to 4 Hz and 16 to 28 Hz was selected by a restrictive multiple-comparison test, among the entire 2 to 60 Hz spectrum. Using 2 to 4 Hz values it is possible to choose a classification rule with an estimate sensitivity and specificity given by 68% and 76% respectively. Alternatively, when 16 to 28 Hz values are used, it is possible to obtain a better classification rule with an estimate sensitivity and specificity given by 81% and 80%, respectively. To the best of our knowledge, this is the first electroencephalography or MEG study where a so exhaustive analysis of the magneto-electric spectrum has been performed. This study supports the notion that more attention should be devoted to the study of beta band in AD.