Alzheimer’s disease has very specific etiology that can typically only be confirmed postmortem. Are there ways to identify it in the dynamical features of brain activity?
Alzheimer’s disease (AD), a neurodegenerative disorder characterized by a decline in cognitive functioning, in particular memory loss, is the most common cause of dementia with an estimated 30 million people affected worldwide [1,2]. At a neurobiological level it is characterized by aggregations of beta-amyloid (Aβ) protein into plaques, the accumulation of tau protein neurofibrillary tangles and progressive neurodegeneration. One recent question of interest is how these structural changes translate into changes in brain activity. Can it be reliably measured in the EEG to provide biomarkers of disease onset and progression, allowing clinicians to make an early diagnosis and intervention?
Biomarkers for early intervention.
For most of its history, AD has been diagnosed solely through clinical observation and cognitive testing, with a confirmatory diagnosis only performed on postmortem examination. However, the neurobiological changes associated with AD, and a potential precursor, Mild Cognitive Impairment (MCI), often appear many years (or even decades) before any visible clinical signs in the patient.
The advent of neuroimaging and the development of new biomarkers offer clinicians the opportunity to do this [3,4]. However, the challenges of developing either structurally or functionally relevant AD biomarkers which provide accurate and reliable indicators of disease onset, progression and outcome, or which assist in drug development, are considerable.
Examples of currently accepted biomarkers involve measuring levels of brain chemicals related to amyloid or tau (e.g. in the cerebrospinal fluid, CSF), or through estimates of metabolic activity (e.g. with Positron Emission Tomography, PET). For example, CSF levels of amyloid-beta (Aβ42) and phosphorylated tau (p-Tau) are thought to reflect AD pathology. In addition, the formation of plaques and tangles disrupt the balance of excitatory and inhibitory activity in the brain, and also result in synaptic dysfunction, at least in mouse models , both of which affect brain dynamics. This provides an opportunity for studying the progression of AD with techniques such as resting-state EEG.
LORETA and Alzheimer’s Disease.
Multiple studies have attempted to examine changes in resting-state EEG dynamics, and to relate these to other markers of AD . For example, in one recent small-scale study, resting-state EEG was used to explore whether there was a relationship between cortical hypometabolism – something commonly observed in AD – and cortical EEG rhythms . To do this they measured cortical hypometabolism using fluorodeoxyglucose-PET and recorded resting EEG in 19 AD patients and compared this against 40 healthy controls and analyzed the results using LORETA. The EEG results showed higher levels of source localized delta band activity that correlated (r=0.579, p=0.009, N=19) with measures of cortical hypometabolism (other bands were not statistically different). This suggests that, in AD patients, delta activity at rest may be related to the PET biomarker of cortical hypometabolism. However, since the healthy patients did not agree to a PET scan, it limits the validity of this conclusion. Also, such conclusion is confounded by similar results relating to a host of other mental health disorders and may simply be representative of a disorder in general, but not AD specifically.
See related post: EEG Frequency Bands Across Mental Health Disorders
Grand average across subjects of the normalized LORETA solutions. From 
CSF markers and Alzheimer’s disease
Another larger-scale study explored the relationship between EEG measures and CSF biomarkers . In this study they compared patients with subjective cognitive decline (n=210) (i.e. they reported subjective complaints but had no significant cognitive deficit or clinical symptoms) against those with MCI (n=230) or AD (n=197). They analyzed resting-state EEG data using two different metrics – global field power (GFP) and global field synchronization (GFS). GFP is a reference-free method that reduces multichannel recordings to a single measure corresponding to the generalized EEG amplitude, resulting in a global measure of scalp potential field strength whilst GFS is a measure of global functional connectivity which resembles the global amount of instantaneous phase locked synchronization of oscillating neuronal networks across the scalp. Linear regression models showed that decreased levels of Aβ42 in the CSF significantly correlated with increased theta (β coefficient=0.514, p<0.001) and delta (β coefficient=0.304, p=0.001) GFP. In addition, decreased levels of Aβ42 in the CSF were significantly associated with decreased GFS alpha (β coefficient=0.024, p<0.001). and beta (β coefficient=0.013, p<0.001). These latter correlations were present in individuals with subjective cognitive decline, suggesting that GFS may be a potential pre-clinical marker of early AD.
These two studies provide a snapshot into the direction of research and progress that is being made in the development of potential resting-state EEG biomarkers which track the progression of AD (other research focuses on task related ERP biomarkers which isn’t discussed here). However, it is unlikely that a single biomarker will be sufficient in adequately predicting the onset, progression and outcome of AD. One longitudinal study which has tried to address this monitored 86 patients initially diagnosed with MCI over a period of 2 years . During this time 25 of the patients developed AD allowing them to search for a marker indicating the likelihood of a patient converting from MCI to AD. They measured multiple different biomarkers and found that several EEG biomarkers based around the alpha and beta range were associated with the conversion from MCI to AD. Rather than focusing on just one of these, they found that by integrating 6 of them together they were able to develop a diagnostic tool that predicted AD progression with a sensitivity of 88% and specificity of 82%. This was compared to a sensitivity of 64% and specificity of 62% when only a single biomarker was used.
The 6 Biomarkers of Interest. From 
EEG offers an opportunity to support the early identification of Alzheimer’s disease and so far there are promising directions. However as with many EEG biomarkers, this search is also hindered by inconsistencies in the methodological approach across studies . More significantly, there is a substantial challenge of identifying markers that are specific to AD and not general to all cognitive and mental health function, one that may be overcome by studying the EEG in multiple forms of Dementia together in combination with multiple other types of markers.
see related post The Remarkable Inconsistency of EEG Frequency Band Defintions
 McKhann, G., Knopman, D., Chertkow, H., Hyman, B., Jack, C., & Kawas, C. et al. (2011). The diagnosis of dementia due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimer’s & Dementia, 7(3), 263-269. doi: 10.1016/j.jalz.2011.03.005
 Holtzman, D., Morris, J., & Goate, A. (2011). Alzheimer’s Disease: The Challenge of the Second Century. Science Translational Medicine, 3(77), 77sr1-77sr1. doi: 10.1126/scitranslmed.3002369
 Maestú, F., Cuesta, P., Hasan, O., Fernandéz, A., Funke, M., & Schulz, P. (2019). The Importance of the Validation of M/EEG With Current Biomarkers in Alzheimer’s Disease. Frontiers In Human Neuroscience, 13. doi: 10.3389/fnhum.2019.00017
 Wurtman, R. (2015). Biomarkers in the diagnosis and management of Alzheimer’s disease. Metabolism, 64(3), S47-S50. doi: 10.1016/j.metabol.2014.10.034
 Selkoe, D. (2002). Alzheimer’s Disease Is a Synaptic Failure. Science, 298(5594), 789-791. doi: 10.1126/science.1074069
 Cassani, R., Estarellas, M., San-Martin, R., Fraga, F., & Falk, T. (2018). Systematic Review on Resting-State EEG for Alzheimer’s Disease Diagnosis and Progression Assessment. Disease Markers, 2018, 1-26. doi: 10.1155/2018/5174815
 Babiloni, C., Del Percio, C., Caroli, A., Salvatore, E., Nicolai, E., & Marzano, N. et al. (2016). Cortical sources of resting state EEG rhythms are related to brain hypometabolism in subjects with Alzheimer’s disease: an EEG-PET study. Neurobiology Of Aging, 48, 122-134. doi: 10.1016/j.neurobiolaging.2016.08.021
 Smailovic, U., Koenig, T., Kåreholt, I., Andersson, T., Kramberger, M., Winblad, B., & Jelic, V. (2018). Quantitative EEG power and synchronization correlate with Alzheimer’s disease CSF biomarkers. Neurobiology Of Aging, 63, 88-95. doi: 10.1016/j.neurobiolaging.2017.11.005
 Poil, S., de Haan, W., van der Flier, W., Mansvelder, H., Scheltens, P., & Linkenkaer-Hansen, K. (2013). Integrative EEG biomarkers predict progression to Alzheimer’s disease at the MCI stage. Frontiers In Aging Neuroscience, 5. doi: 10.3389/fnagi.2013.00058