Alzheimer’s Disease (AD) is characterized by progressive amnesia resulting in global dementia. The most consistent electrophysiological findings involve a prolonged latency and decreased amplitude of P300 in oddball paradigms. However, these findings remain relatively unnoticable in the early stages of the disease. In this study, P3b and P3a potentials in auditory oddball and novelty paradigms were measured from 14 healthy volunteers and 22 patients at the early to medium stages of AD (early, GDS<=4 and CDR < 2, medium, GDS>4 or CDR >=2) with the aim to find a more sensitive sign of the early AD. All patients, who were referred from the Department of Behavioral Neurology and Movement Disorders, were evaluated with MMSE, and Global Deterioration (GDS) and Clinical Dementia Rating Scales (CDR). In the oddball paradigm the differences in P3b amplitudes and latencies between the AD patients and the control group were not statistically significant. P3a amplitudes to novel events showed a significant overall latency prolongation, but no significant amplitude reduction. However, the amplitude of the P3b generated by target stimuli of the novelty paradigm showed a highly significant decrease in the parietal leads. These findings suggests that there is a delay in the processing of distracting novel stimuli in the novelty paradigm, and that this causes a difficulty in focusing the attention to target stimuli. In this study, the oddball P3b amplitude and latency changes reported in the literature have not been observed possibly due to the subjects’ being in an early stage of the disease. However, the P3b potentials obtained in the novelty paradigm are more sensitive in detecting early stage AD subjects than classical oddball paradigm.