Introduction: The diagnosis of Alzheimer's Disease (AD) is made mostly on clinical grounds. Electroencephalography (EEG) has been investigated as a possible surrogate biological marker for AD. A number of studies indicate that EEG is a sensitive measure of AD but not very specific. The aim of this study is to investigate how to make EEG more clinically useful for early diagnosis of AD by using a cholinergic antagonist, scopolamine.

Methods: In this study we use advanced signal processing and pattern recognition algorithms together with administration of scopolamine to challenge the cholinergic system, which deteriorates significantly in the earliest phases of the disease. We hypothesize that the use of scopolamine might enhance the sensitivity of EEG in AD. A clinical trial of 10 AD subjects, 10 patients with Mild Cognitive Impairment and 10 age matched healthy individuals was conducted.

Results: A statistical pattern classifier was used to find features to discriminate between AD and healthy controls. We obtained complete separation of the groups using our statistical pattern classifier approach, which takes all features into account. In the last 18 months 3 of the MCI patients have progressed to AD, their initial pattern, indistinguishable from the AD group.

Conclusions: We demonstrate that this approach discriminates between the two groups and thus suggests that it may lead to improved early diagnosis of AD. Furthermore, this approach may be helpful in discriminating persons with MCI progressing to AD from those who will not progress.