F66. Event-related potentials on working memory tasks for early diagnosis of Alzheimer’s disease and mild cognitive impairment

Introduction Alzheimer’s disease (AD) affects above 24 million people in the world. Petersen et al. developed the concept of mild cognitive impairment (MCI), thus describing a state of pre-dementia characterized by intermediate symptoms between normal aging and dementia caused by the onset of AD. Electroencephalographic (EEG) records in patients with AD/MCI evidence neuronal degeneration and functional/ cognitive impairment long before loss of brain tissue can be detected by structural neuroimaging methods, thus allowing an early diagnosis. A number of studies show that AD and MCI patients demonstrate abnormalities in EEG rhythms. Such alterations can be observed in both resting and task-related EEG protocols. In this study we conducted an Event-Related Potential (ERP) analysis from 32-channel EEG recorded when participants from three groups (healthy elderly, MCI and AD) performed a three-level N-back working memory (WM) task, with the purpose of finding out whether or not an early diagnosis of AD/MCI can be established based on such measures. Methods A total of 61 participants were enrolled, of whom 15 had Alzheimer’s disease (AD), 20 had mild cognitive impairment (MCI) and 26 were healthy elderly (HE) individuals. Control group participants (HE) were recruited at Concordia University in Montreal, Canada. Patients (MCI and AD) were recruited at the Jewish General Hospital (JGH) Memory Clinic, a tertiary care referral center of McGill University. Ethical approval was obtained from the involved institutions and all participants provided written consent. All subjects were submitted to a three-level N-back task in ascending WM load (0-back, 1-back, 2-back). Each condition consisted of 100 trials, 40% of which were match trials. The stimuli (digits from 1 to 9) were presented during 600 ms with an inter-stimulus interval of 1400 ms. Participants responded by pressing the left or right button on a keyboard with the index finger of each hand. EEG signals were recorded at 500 Hz sampling rate using a 32-channel Neuroscan device. Of the 32 channels available, two were reserved to monitor horizontal and vertical eyeball movements, and a third one was attached to the right earlobe and used as an additional reference. ERPs of the 29-channel EEG for each of the six N-back tasks (match and non-match responses of the three-level WM load) were separately obtained and submitted to statistical analysis. The multiple comparison tests used were based on the Cluster correction. Results A considerable amount of significant differences between patient and control groups were found in the ERP during execution of the WM tasks, predominantly in fronto-centro-parietal electrodes. Our results indicate that, compared to the control group, both patient groups (MCI and AD) show a decrease in the amplitude of the P450 component, which is related to WM update. Conclusion Although tests on larger databases are required to confirm these results, they are promising for achieving an early ERP-based diagnosis of MCI and AD.