Format

Send to

Choose Destination
Front Hum Neurosci. 2016 May 20;10:238. doi: 10.3389/fnhum.2016.00238. eCollection 2016.

Slowing of Hippocampal Activity Correlates with Cognitive Decline in Early Onset Alzheimer's Disease. An MEG Study with Virtual Electrodes.

Author information

1
Alzheimer Center and Department of Neurology, Neuroscience Campus Amsterdam, VU University Medical Center Amsterdam, Netherlands.
2
Department of Clinical Neurophysiology and Magnetoencephalography Center, Neuroscience Campus Amsterdam, VU University Medical Center Amsterdam, Netherlands.
3
Alzheimer Center and Department of Neurology, Neuroscience Campus Amsterdam, VU University Medical CenterAmsterdam, Netherlands; Department of Epidemiology and Biostatistics, Neuroscience Campus Amsterdam, VU University Medical CenterAmsterdam, Netherlands.
4
Department of Clinical Neurophysiology and Magnetoencephalography Center, Neuroscience Campus Amsterdam, VU University Medical CenterAmsterdam, Netherlands; Nutricia Advanced Medical Nutrition, Nutricia ResearchUtrecht, Netherlands.

Abstract

Pathology in Alzheimer's disease (AD) starts in the entorhinal cortex and hippocampus. Because of their deep location, activity from these areas is difficult to record with conventional electro- or magnetoencephalography (EEG/MEG). The purpose of this study was to explore hippocampal activity in AD patients and healthy controls using "virtual MEG electrodes". We used resting-state MEG recordings from 27 early onset AD patients [age 60.6 ± 5.4, 12 females, mini-mental state examination (MMSE) range: 19-28] and 26 cognitively healthy age- and gender-matched controls (age 61.8 ± 5.5, 14 females). Activity was reconstructed using beamformer-based virtual electrodes for 78 cortical regions and 6 hippocampal regions. Group differences in peak frequency and relative power in six frequency bands were identified using permutation testing. For the patients, spearman correlations between the MMSE scores and peak frequency or relative power were calculated. Moreover, receiver operator characteristic curves were plotted to estimate the diagnostic accuracy. We found a lower hippocampal peak frequency in AD compared to controls, which, in the patients, correlated positively with MMSE [r(25) = 0.61; p < 0.01] whereas hippocampal relative theta power correlated negatively with MMSE [r(25) = -0.54; p < 0.01]. Cortical peak frequency was also lower in AD in association areas. Furthermore, cortical peak frequency correlated positively with MMSE [r(25) = 0.43; p < 0.05]. In line with this finding, relative theta power was higher in AD across the cortex, and relative alpha and beta power was lower in more circumscribed areas. The average cortical relative theta power was the best discriminator between AD and controls (sensitivity 82%; specificity 81%). Using beamformer-based virtual electrodes, we were able to detect hippocampal activity in AD. In AD, this hippocampal activity is slowed, and correlates better with cognition than the (slowed) activity in cortical areas. On the other hand, the average cortical relative power in the theta band was shown to be the best diagnostic discriminator. We postulate that this novel approach using virtual electrodes can be used in future research to quantify functional interactions between the hippocampi and cortical areas.

KEYWORDS:

Alzheimer’s disease; MEG; beamformer; hippocampus; peak frequency; relative power; source-space; virtual electrode

Supplemental Content

Full text links

Icon for Frontiers Media SA Icon for PubMed Central
Loading ...
Support Center