Decreased EEG synchronization in Alzheimer’s disease and mild cognitive impairment

Abstract

The hypothesis of a functional disconnection of neuro-cognitive networks in patients with mild cognitive impairment (MCI) and Alzheimer Dementia was investigated using baseline resting EEG data. EEG databases from New York (264 subjects) and Stockholm (155 subjects), including healthy controls and patients with varying degrees of cognitive decline or Alzheimer Dementia were analyzed using Global Field Synchronization (GFS), a novel measure of global EEG synchronization. GFS reflects the global amount of phase-locked activity at a given frequency by a single number; it is independent of the recording reference and of implicit source models. Patients showed decreased GFS values in Alpha, Beta, and Gamma frequency bands, and increased GFS values in the Delta band, confirming the hypothesized disconnection syndrome. The results are discussed within the framework of current knowledge about the functional significance of the affected frequency bands.

Introduction

Selective distribution of characteristic neurofibrillary pathology in Alzheimer’s disease (AD) follows the projection of anatomico-functional circuits of the brain [2], [3]. During the course of the disease, the degenerative process spreads in a characteristic manner starting in the entorhinal region and progressing through the hippocampus and the limbic structures towards the neocortical association areas [2]. The disconnection is caused by the degeneration of the large cortical pyramidal neurons [14], [16] and assumingly occurs at the level of cortico-cortical projections. These cortico-cortical projections are crucial for higher order information processing, since they permit the integration of multimodal information through the coherent firing of extended neural networks. Therefore, Morrison et al. [27] suggested a “disconnection syndrome” as a model for the AD symptomatology. Since coherent firing of a sufficiently large number of neurons produces a detectable voltage field on the scalp, it is plausible that EEG data can be used to study changes in synchronization of neural circuits in AD that possibly occur at an early stage of the disease.

Within this context, a novel method called Global Field Synchronization (GFS) has recently been introduced to measure functional synchronization in frequency-domain EEG data [22]. For the entire set of EEG channels and for a given frequency, a single GFS value is obtained representing the percentage of EEG activity that oscillated with a common phase. High GFS values in a given frequency indicate, therefore, that a large part of the recorded brain activity at that frequency is phase locked. The presence of phase locked, oscillatory brain activity has been related to short-term binding of neural networks and may play an important role for higher order information processing and cognition [13], [33]. Since resting EEG usually is not assumed to originate from very focal sources, the amount of synchronization measured by GFS can reasonably be assumed to be mediated through a network of cortico-cortical connections. Contrary to coherence, where the stability of the phase relation between two pre-selected electrodes across analysis epochs is assessed and usually taken as indicator of cooperativity between the underlying regions, GFS makes no assumption about the spatial location of the activity. The computation of GFS assumes (in agreement with the existing volume conductor models of the head) that most types of measurable brain activity generate signals that are distributed over many electrodes on the scalp. If these signals show a predominant phase across all electrodes, this indicates that the (possibly many) local brain processes are interacting such that they appear to constitute a phase-locked, single and global meta-process. Compared to coherence that is local in space and global in time (i.e. requires selected pairs of electrodes recorded against a third, also pre-selected electrode, and a series of analysis epochs), GFS is global in space and local in time (i.e. it considers, during a single analysis epoch, all electrodes simultaneously and independent of the reference electrode). Therefore, since GFS does not require any assumptions about the location of the active processes, and since it can be assessed within relatively short (1–2 s) time periods, it may be well suited to study a hypothesized general disruption of short-lasting functional binding between extended neural networks. Additionally, the frequency information provided by the measure may be used to gain further information about the functional significance of the assumingly disrupted processes. Although animal work has shown that Gamma (40 Hz) activity plays an important role for feature binding [8] EEG research has indicated that other frequency bands are also crucially involved in human information processing [9]. A previous study on GFS comparing acutely schizophrenic patients with healthy controls has shown that patients had reduced GFS values in the Theta band, which has been related to the existing hypothesis of a disconnection of working memory processes in schizophrenia [22].

The aim of the present study was to explore and extend the hypothesis of functional disconnection in AD and mild cognitive impairment (MCI) as reflected in GFS. The study was based on two independent, large samples encompassing healthy subjects, individuals with mild cognitive impairment and AD patients in different stages of the disease.