Decreased Alpha Reactivity from Eyes-Closed to Eyes-Open in Non-Demented Older Adults with Alzheimer’s Disease: A Combined EEG and [ 18 F]florbetaben PET Study

Abstract

Background:

The degree of alpha attenuation from eyes-closed (EC) to eyes-open (EO) has been suggested as a neural marker of cognitive health, and its disruption has been reported in patients with clinically defined Alzheimer’s disease (AD) dementia.

Objective:

We tested if EC-to-EO alpha reactivity was related to cerebral amyloid-β (Aβ) deposition during the early stage of AD.

Methods:

Non-demented participants aged ≥55 years who visited the memory clinic between March 2018 and June 2019 (N = 143; 67.8% female; mean age±standard deviation, 74.0±7.6 years) were included in the analyses. Based on the [¹⁸F]florbetaben positron emission tomography assessment, the participants were divided into Aβ+ (N = 70) and Aβ- (N = 73) groups. EEG was recorded during the 7 min EC condition followed by a 3 min EO phase, and a Fourier transform spectral analysis was performed.

Results:

A significant three-way interaction was detected among Aβ positivity, eye condition, and the laterality factor on alpha-band power after adjusting for age, sex, educational years, global cognition, depression, medication use, and white matter hyperintensities on magnetic resonance imaging (F = 5.987, p = 0.016); EC-to-EO alpha reactivity in the left hemisphere was significantly reduced in Aβ+ subjects without dementia compared with the others (F = 3.984, p = 0.048).

Conclusion:

Among mild cognitive impairment subjects, alpha reactivity additively contributed to predict cerebral Aβ positivity beyond the clinical predictors, including vascular risks, impaired memory function, and apolipoprotein E ɛ4. These findings support that EC-to-EO alpha reactivity acts as an early biomarker of cerebral Aβ deposition and is a useful measurement for screening early-stage AD.

Keywords

Aged alpha rhythms electroencephalography eyes-closed resting state eyes-open resting state

INTRODUCTION

Cerebral amyloid-β (Aβ) deposition begins 15–20 years earlier than the initial onset of cognitive symptoms in Alzheimer’s disease (AD) [1]. Thus, predicting cerebral Aβ deposition is important for early diagnosis and intervention of AD, including therapeutic clinical trials [1, 2]. Although cerebral Aβ accumulation can be non-invasively detected in the living human brain by an amyloid positron emission tomography (PET) scan [3], it is not always performed due to the high cost and the risk of radiation exposure. Therefore, it is important to develop a more feasible test to distinguish Aβ pathology before AD progresses [4].

Electroencephalogram (EEG) is a non-invasive, widely available, and sensitive approach to detect synaptic changes [5]. Many previous reports have observed Aβ oligomer neurotoxicity in synaptic proteins and decreases in synaptic spines [6]. EEG has been proposed as a tool for examining larger populations with AD-related risks [7]. Many studies have compared EEG characteristics of AD dementia patients to those of healthy counterparts. Widespread slowing and decreased complexity of brain electrical activity have been repeatedly reported in patients with AD dementia [8]; a relative increase in theta and delta activities and a decrease in alpha and beta activities have been reported by quantitative EEG studies [9 –11].

However, only a few studies have investigated the EEG parameters related to Aβ pathology in the early stages of AD and showed mixed results. One study reported that a decrease in cerebrospinal fluid Aβ₄₂ level was significantly correlated with increased theta and delta global field power and decreased alpha and beta global field synchronization; however, significance in subgroup analyses remained only in patients with mild cognitive impairment (MCI) and dementia, and not in cognitively normal older adults [12]. Another study that examined EEG connectivity related to Aβ pathology based on florbetapir-PET did not reveal any significant correlations in older subjects with normal cognition [13].

On the other hand, increased alpha power in the medial frontal cortex of cognitively normal subjects with cerebral Aβ deposition was reported by a magnetoencephalography study [14], and a recent EEG study showed significant alpha rhythm features in preclinical AD subjects with high education attainment [15]. In general, resting-state alpha rhythm is known to probe general neurophysiological neural synchronization [16], and abnormally low amplitude posterior alpha activity has been consistently reported to be related to clinically defined AD dementia and MCI [15]. Taken together, alterations in alpha rhythms appear to be a possible marker and begin earlier than other EEG abnormalities from the preclinical stage of AD [17].

Alpha reactivity, which is the degree of decrease in alpha amplitude from the eyes-closed (EC) to the eyes-open (EO) condition, is a sensitive indicator of mental activity [18]. Suppressed EC-to-EO alpha reactivity has been hypothesized to reflect impaired cortical neural desynchronization [16], and has been suggested as a neural marker of impaired cognitive health [19, 20]. Reduced alpha reactivity in AD dementia patients has been reported consistently [21], and a topographical difference in alpha reactivity during the progression of clinical AD has also been revealed in a previous study; the authors suggested that an examination of EC-to-EO alpha reactivity could be a simple and sensitive method to probe the relationships between aging and mechanisms of cortical neural synchronization/desynchronization without confounding factors of task complexity or learning bias [16]. Therefore, the alpha reactivity can be an early possible biomarker for detecting cerebral Aβ accumulation; however, their association has not been investigated.

The aim of this study was to examine the quantitative EEG parameters related to cerebral Aβ accumulation in participants without dementia. In particular, we investigated the effect of Aβ pathology on the alpha reactivity parameter; we tested the hypothesis that subjects with cerebral Aβ deposits have decreased EC-to-EO alpha reactivity compared with those without Aβ.

MATERIALS AND METHODS

Participants

A total of 263 participants aged ≥55 years who visited the memory clinic of Seoul National University Hospital (SNUH) were enrolled between March 2018 and June 2019. Among them, 143 subjects were included in the study after 120 individuals were excluded for the following reasons: 1) diagnosed with dementia, 2) major psychiatric disorder (e.g., schizophrenia, bipolar disorder, or substance use disorder) other than depressive disorder, 3) any serious medical or neurological disease that could affect mental function (e.g., stroke diagnosed within 1 year, seizure disorder, parkinsonism, or delirium), or 4) insufficient data on any of three modalities, including EEG, brain magnetic resonance imaging (MRI), and [¹⁸F]florbetaben PET scan.

This study was approved by the Institutional Review Board of SNUH (IRB No. 1909-149-1067), and all subjects provided written informed consent before participation.

Clinical assessments

Participants were evaluated for cognitive disorders by a psychiatrist using semi-structured interviews and a clinical psychologist who conducted neuropsychological tests, both of which were based on the Korean version of the Consortium to Establish a Registry for Alzheimer’s disease (CERAD-K) [22]. During the evaluation, reliable informants were interviewed, and medical records were reviewed to acquire accurate data. Final diagnoses were confirmed in a consensus meeting that included four or more geriatric psychiatrists.

In the present study, we excluded participants diagnosed with dementia based on the criteria of the Diagnostic and Statistical Manual of Mental Disorders fourth edition (DSM-IV-TR) [23]. The remaining 27 individuals with subjective memory complaints without objectively defined impairment (SMI), and 116 MCI individuals who met the following core clinical criteria recommended by the NIA-AA guidelines [24] were included: 1) memory complaints corroborated by self-report, an informant, or a clinician; 2) objective memory impairment according to age, education level, and sex; 3) largely intact functional activities; and 4) no dementia. In terms of criterion 2), all MCI participants showed a performance score of at least 1.5 SD below the respective age-, education-level-, and sex-specific mean for at least one of the tests in the CERAD-K neuropsychological battery (i.e., Sematic Fluency, Boston Naming, Word List Memory, Visuospatial Construction, Word List Recall, Word List Recognition, Constructional Recall, and Stroop color-word tests).

Depression was defined by the National Institute of Mental Health provisional criteria for depression in AD [25] and the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition criteria [26] for minor and major depressive disorders: individuals with any of three diagnoses were defined to have depression [25]. The Hamilton Depression Rating Scale (HAMD) was used to quantify the depressive symptoms of the participants [27, 28].

Vascular risk factors, including hypertension, diabetes mellitus, dyslipidemia, coronary heart disease, transient ischemic attack, and stroke were assessed using a systematic interview and chart review. We investigated the usage of sedative-hypnotics (e.g., benzodiazepines and Z-drugs), and the anticholinergic cognitive burden (ACB) score was assessed based on a review article: medications were identified to have either possible (score of 1) or definite anticholinergic effects (score of 2 or 3), and the sum of the scores was defined as the ACB score of each participant [29]. All medications used were confirmed by prescription reports at the time of EEG acquisition. Lastly, all subjects underwent apolipoprotein E (APOE) genotyping.

EEG acquisition and processing

Participants were seated in an air-conditioned room and fitted with the EEG equipment. The EEG recording began with a 7 min EC phase followed by a 3 min EO phase and 5 min EC phase; we designed the first EC phase to have a longer duration to obtain the obvious resting state EEG [30]. EEG data were recorded at a sampling frequency of 500 Hz from 19 scalp electrodes positioned according to the 10–20 system (i.e., Fp1, Fp2, F7, F3, Fz, F4, F8, T3, C3, Cz, C4, T4, T5, P3, Pz, P4, T6, O1, and O2). As a reference, a single channel with linked electrodes was attached to both ears (A1, A2). The vertical electrooculogram (EOG) was also recorded from electrodes placed above and below the left eye, and the horizontal EOG was recorded from lateral sites of both eyes.

MATLAB (version 2018b; MathWorks, Natick, MA, USA) and EEGLAB toolbox (Version 2019.0; Swartz Center for Computational Neuroscience, La Jolla, CA, USA) were used to process the data. EEG signals were processed with a bandpass filter from 0.1 to 50 Hz; Artifacts such as participant eye movements, temporal muscle activity, and line noise were removed from data epochs using the independent component analysis (ICA) algorithms implemented in EEGLAB. After removing the artifacts, the signals were re-referenced using the common average of the potentials at 21 electrodes (19 scalp and 2 ear electrodes). The removed artifact data were separated by two periods as the first EC and EO. A spectral analysis was performed using short-time Fourier transform in each period, which is commonly used for a frequency domain analysis [31]. A 30 s Hamming window was used to obtain the frequency power during each period. Five frequency bands were integrated: delta (1–4 Hz), theta (4–8 Hz), alpha (8–13 Hz), beta (13–30 Hz), and gamma (30–50 Hz) in each period.

Neuroimaging

All participants underwent a conventional axial fluid-attenuated inversion-recovery MRI scan on a SIEMENS Verio 3-Tesla scanner or Skyra 3-Tesla scanner (Siemens, Erlangen, Germany). The degree of white matter hyperintensities (WMH) was assessed using Fazekas’ scale (range 0–3) [32, 33] by trained psychiatrists and confirmed in a diagnostic consensus meeting; the deep WMH score was used as a covariate in analyses.

[¹⁸F]florbetaben PET scans were conducted for all subjects (Biograph True Point 40 and Biograph mCT, Siemens Healthcare). Following intravenous administration of 300 MBq of [¹⁸F]florbetaben, a 10 min emission scan was obtained 90 min after injection. PET images were reconstructed using an iterative algorithm (ordered-subset expectation maximization [OSEM]) with five iterations and 24 subsets. Attenuation correction was performed by the CT scanner after the emission scan. Images were reconstructed into a 256×256 image matrix and a 4 mm Gaussian post-reconstruction filter was applied.

Participants were classified as amyloid positive (Aβ+) or amyloid negative (Aβ-) according to the Brain Amyloid Plaque Load (BAPL) score [34]. The BAPL score was determined by visually inspecting the degree of amyloid accumulation. First, the regional cortical tracer uptake (RCTU) score of four areas of interest, i.e., lateral temporal cortex, frontal cortex, posterior cingulate cortex/precuneus, and parietal cortex, in a three-grade scoring system (1, no tracer uptake; 2, moderate tracer uptake; and 3, pronounced tracer uptake). A RCTU score of 1 in all four areas led to a BAPL score of 1, and an RCTU score of 2 in any region led to a BAPL score of 2. An RTCU score of 3 in any of the four areas led to a BAPL score of 3. If the BAPL score was >1, the participant was defined as Aβ+ [35]. Clinical ratings were provided by SNUH nuclear medicine physicians and confirmed in a diagnostic consensus meeting.

Statistical analyses

Differences in sociodemographic variables and clinical data between the Aβ+ and Aβ- groups were examined using the t-test for continuous variables and the chi-square test for categorical variables. Fisher’s exact test was used for the variables if a cell had an expected value <5.

For topographic analyses, EEG data were averaged across nine regions of interest (ROIs) based on a previous study [36], after which they were reorganized as polarity (frontal, central, and posterior) and laterality (left, midline, and right) factors in the analyses according to group (Aβ+ and Aβ–) and condition (EC and EO) [37].

As the EC-to-EO alpha band power would be different based on cerebral Aβ deposition, we created an alpha reactivity index (ARI) based on a previous study as follows [20]: $\begin{matrix} Alpha Reactivity Index (ARI) \\ = \frac{EC alpha power - EO alpha power}{EC alpha power} \end{matrix}$

The following four-way multivariate analysis of covariance (MANCOVA) was performed: Group (Aβ+ and Aβ–)×Condition (EC and EO)×Polarity (frontal, central, and posterior)×Laterality (left, midline, and right). Group was an independent-group factor, in that each participant belonged to either the Aβ+ or Aβ–group. The other three factors, group, polarity, and laterality, were repeated-measures factors. Each participant experienced both conditions (EC and EO) with measuring their EEG signals on three horizontally different locations (frontal, central, and posterior) as well as three vertically different locations (left, midline, and right). Analyses were performed with age, sex, education, and MMSE score as covariates (model 1), and with additional covariates of depression, sedative-hypnotic usage, ACB score, and WMH on MRI (model 2). The ARI between the ROIs of the Aβ+ and Aβ–groups was compared for post-hoc analyses.

We developed the Aβ prediction model using a stepwise multivariate logistic regression. Of the various factors possibly associated with Aβ, only the variables which are commonly obtained in clinic practice were selected. Demographic (age, sex, education), clinical (hypertension, diabetes, dyslipidemia, cardiovascular disease, stroke, and transient ischemic attack), neuropsychological (MMSE for global cognitive function, Word List Recall [WLR] for episodic memory, Boston Naming Test for language, Semantic fluency for executive function, and Constructional Praxis for visuospatial function; all of them were involved in CERAD-K neuropsychological battery), and APOE ɛ4 positivity were selected. The final model for Aβ prediction was made using forward likelihood ratio (LR) method. Then, a hierarchical logistic regression model was used to infer the contribution of the ARI to predict the likelihood of cerebral Aβ pathology (block 2) beyond the factors from the final model (block 1).

All analyses were conducted using SPSS version 23 (IBM Corp., Armonk, NY, USA). All statistical tests were two-tailed, and p < 0.05 was considered significant. As all contrasts were planned independently of the data, Bonferroni-type adjustments were not required [38].

RESULTS

No significant differences in demographic variables or clinical data were observed, including cognitive function, vascular health status, depression, or medication use between the Aβ+ and Aβ– groups, except for the higher prevalence of the APOE ɛ4 allele in Aβ+ subjects (Table 1).

Table 1

Baseline characteristics of the subjects

	Amyloid-negative (n = 70)	Amyloid-positive (n = 73)	P
Demographic variables
Age, y (SD)	74.07 (7.50)	73.95 (7.79)	0.921
Sex, n (female/male)	47/23	50/23	0.863
Education, y (SD)	10.61 (5.96)	10.38 (5.05)	0.803
APOE4 allele, n (%)	11 (15.94)	47 (66.20)	<0.001
Cognitive status
MMSE, score (SD)	22.31 (4.72)	22.66 (3.62)	0.627
MCI, n (%)	53 (75.71)	63 (86.30)	0.106
Vascular health status
Vascular risk score (SD)	1.44 (1.20)	1.36 (1.19)	0.666
DWMHI, score (SD)	1.50 (0.70)	1.32 (0.60)	0.105
PVHI, score (SD)	1.77 (0.78)	1.66 (0.67)	0.351
Mood status
^aDepression, n (%)	31 (44.3)	32 (43.8)	0.957
HAMD, score (SD)	5.91 (5.04)	5.75 (6.13)	0.864
Drug use
ACB, score (SD)	0.59 (1.42)	0.41 (1.19)	0.425
Sedative-hypnotics user, n (%)	15 (22.86)	15 (20.55)	0.738

APOE4, apolipoprotein E ɛ4; MMSE, Mini-Mental State Examination; MCI, mild cognitive impairment; DWMHI, deep white matter hyperintensities; PVHI, periventricular hyperintensities; HAMD, Hamilton Depression Rating Scale; ACB, anticholinergic cognitive burden score. Student’s t-test was performed for continuous variables and the chi-square test was used for categorical variables. ^aDepression included patients who were diagnosed with a major depressive episode, minor depressive episode, or satisfied the diagnostic criteria for depression in AD proposed by the National Institute of Mental Health.

The mean power spectral density of each frequency band was not different between the Aβ+ and Aβ–groups, except for the alpha band in the EO phase (F = 4.123, p = 0.044). Table 2 shows a significant three-way interaction among Aβ, eye conditions, and the laterality factor on alpha-band power after adjusting for age, sex, education, and MMSE score (model 1); this significant interaction was maintained after additional adjustment for depression, sedative-hypnotic usage, ACB score, and deep WMH on MRI (model 2). Consequently, EC-to-EO alpha reactivity was compared by laterality in a post-hoc analysis; ARI decreased significantly in the left hemisphere of Aβ+ subjects, and a more prominent decrease was observed in the left posterior area (F = 4.966, p = 0.027; Fig. 1 and Table 3).

Table 2

Four-way MANCOVA analysis including the interaction terms among amyloid-beta positivity, eye condition, laterality, and polarity on alpha-band power

	^aModel 1		^bModel 2
	F-value	P	F-value	P
Amyloid	1.433	0.233	0.808	0.370
Condition	2.066	0.153	2.849	0.094
Laterality	1.552	0.215	1.415	0.236
Polarity	0.847	0.359	0.192	0.662
Amyloid×condition	0.465	0.496	0.37	0.544
Amyloid×condition×laterality	5.987	0.016	6.618	0.011
Amyloid×condition×polarity	0.139	0.710	0.133	0.716
Amyloid×condition×laterality×polarity	1.230	0.269	2.315	0.131

MANCOVA, multivariate analysis of covariance. All df = 1. ^aModel 1 was adjusted age, sex, education years, and the Mini-Mental State Examination. ^bModel 2 was additionally adjusted for depression, vascular risk score, deep white matter hyperintensities, anticholinergic cognitive burden score, and sedative-hypnotics usage.

Fig. 1

Regional relative band power differences according to cerebral Aβ deposition from the eyes-closed to eyes-open condition.

Table 3

Regional ARI difference by cerebral Aβ deposition

		Mean	SD	F-value	P
Left	Aβ-	0.685	0.249	3.984	0.048
	Aβ+	0.614	0.224
Midline	Aβ-	0.639	0.292	2.867	0.093
	Aβ+	0.570	0.235
Right	Aβ-	0.678	0.262	0.627	0.430
	Aβ+	0.654	0.191

SD, standard deviation; Aβ, amyloid-β; ARI, alpha reactivity index. All df = 1. Adjusted for age, sex, education years, and the Mini-Mental State Examination.

Based on the results of stepwise multivariate logistic regression analysis using the forward LR method, diabetes, dyslipidemia, the WLR score, and APOE ɛ4 positivity in combination significantly predicted Aβ deposition. Therefore, these four variables were included simultaneously in the final Aβ prediction model as covariates. In MCI subjects, left ARI contributes significantly to predicting cerebral Aβ deposition in addition to those factors included in the final model. A significant increase in model fit was detected after adding ARI as a predictor in the MCI group (χ² = 4.346, p = 0.037). However, the contribution of the ARI in predicting Aβ pathology was not significant in total AD sample (χ² = 2.905, p = 0.088, Table 4).

Table 4

Additive contribution of the left ARI to predict cerebral Aβ deposition from a hierarchical multivariate logistic regression model

	Adjusted odds ratio (95% confidence interval)
	Total sample (n = 140)^a				MCI (n = 116)^a
	Final prediction model for Aβ^b				Final prediction model for Aβ^b
	without ARI		with ARI		without ARI		with ARI
Diabetes	0.20	(0.062–0.631)^*	0.19	(0.058–0.610)^*	0.17	(0.047–0.644)^*	0.18	(0.047–0.702)^*
Dyslipidemia	3.00	(1.162–7.739)^*	2.91	(1.102–7.672)^*	5.84	(1.835–18.616)^*	5.45	(1.636–18.140)^*
WLR score	0.54	(0.359–0.811)^*	0.54	(0.361–0.816)^*	0.43	(0.258–0.712)^*	0.39	(0.233–0.669)^*
APOE ɛ4 carrier	11.02	(4.494–27.006)^*	11.56	(4.622–28.933)^*	7.87	(2.941–21.066)^*	8.33	(3.015–23.009)^*
Left ARI	0.22	(0.037–1.267)	0.12	(0.015–0.917)^*
Chi-square		54.618^*		57.523^*		47.554^*		51.90^*
Cox &Snell R ²		0.323		0.337		0.344		0.368

Aβ, amyloid-β; ARI, alpha reactivity index; WLR, word list recall. ^aThree missing values on APOE ɛ4 carrier. ^bStepwise multivariate logistic regression analysis was conducted with demographic (age, sex, education), clinical (hypertension, diabetes, dyslipidemia, cardiovascular disease, stroke, and transient ischemic attack), neuropsychological (Mini-Mental State Examination for global cognitive function, Word List Recall for episodic memory, Boston Naming Test for language, Semantic fluency for executive function, and Constructional Praxis for visuospatial function), and APOE ɛ4 positivity using the forward likelihood ratio method. The final model includes diabetes, dyslipidemia, WLR, and APOE ɛ4. ^*Statistically significant p < 0.05.

DISCUSSION

EC-to-EO alpha reactivity in the left hemisphere was associated with Aβ positivity in non-demented older adults. The ARI, which measures the degree of suppression of regional alpha band power from EC to EO, decreased in the Aβ+ group after adjusting for covariates possibly affecting brain electrical activities, such as the degree of cerebrovascular burden and medication use. Left ARI contributed to prediction of cerebral Aβ deposition beyond age, sex, education, and global cognition among MCI subjects.

Our finding of decreased alpha reactivity in AD is in line with previous EEG studies, even though most studies have been conducted in clinically defined AD dementia patients: reduced alpha suppression under the EO condition is one of the EEG parameters used to discriminate AD dementia patients from controls [21 , 39–41]. A previous study also showed a progressive decrease in EC-to-EO alpha reactivity across AD progression, and this reactivity was correlated with the subjects’ global cognitive function [16]. Meanwhile, cognitively normal older adults presented similar levels of EC-to-EO alpha reactivity compared with healthy young adults [37].

However, all these previous studies were performed without confirming AD-related pathology. Clinically-defined MCI and AD dementia cases usually include many AD phenocopies: 47% of MCI and 12% of probable AD dementia patients do not have Aβ deposits in their brain [42]. Thus, it was not conclusive whether the previous findings of reduced alpha blockage in AD dementia and MCI subjects were due to AD-specific pathology or another neurodegenerative mechanism. Meanwhile, the prevalence of cerebral Aβ positivity without apparent cognitive symptoms increases from 10% in cognitively normal adults aged 50 years to 44% in those aged 90 years [42]. Therefore, cognitively healthy subjects in previous studies might have partly included preclinical AD cases; that is, previous findings regarding EC-to-EO alpha reactivity are not conclusive. A recent study that included cognitively normal older adults revealed that lesions in white matter tracts linking the basal nucleus of Meynert and the visual cortex are associated with reduced alpha reactivity [20]; however, that study did not consider the influence of AD-related pathology.

In the present study, decreased EC-to-EO alpha reactivity was related to pathologically confirmed preclinical and prodromal AD defined by the revised NIA-AA guidelines [24]. Impaired alpha attenuation is often described as loss of the dynamics of responsiveness to the environment [40], and functional disconnections in cortico-cortical and thalamo-cortical networks were suggested as mechanisms [43]. The neuropathological changes underlying AD may rely on abnormal postsynaptic potentials generated in pyramidal neurons of the occipitoparietal region and in neurons conveying signals from the parietal nodes of attention networks to the visual cortex [16]. Resting alpha rhythms reflect cortical neural synchronization, and a previous study proposed that EC-to-EO alpha reactivity may mirror another aspect of synchronization called desynchronization [16]. A decreased ARI reflects functional impairment of cortical neural desynchronization. Our results did not show a decrease in resting-state alpha amplitude, which has been frequently reported in previous studies that compared AD dementia and controls. Thus, we cautiously suggest that impaired desynchronization reflected by the EC-to-EO alpha reactivity is more easily detectable during the early stage of AD compared with disrupted synchronization reflected by resting-state alpha activity.

Some previous studies proposed that structural changes in AD are related to alpha rhythm abnormalities. Decreased grey matter volume including occipital lobe atrophy has been associated with reduced resting-state alpha amplitude [44]. More importantly, cholinergic tract atrophy and disrupted white matter connectivity have been suggested to be related to abnormalities in alpha activity [13, 45], and synaptic dysfunction in the cholinergic tract is positively related with Aβ pathology [46, 47]. However, the deficiency of cholinergic activity is not evident during the initial stage of AD [48]. Although our results support that cholinergic disruption reflected by EC-to-EO alpha reactivity may start during the early phase of AD, its underlying mechanism of dysfunctional neurotransmission caused by AD-related pathology needs to be clarified in future studies. A previous study proposed that tauopathy rather than amyloidosis in the nucleus basalis is more associated with cholinergic dysfunction [49].

Monoaminergic abnormalities in AD are another possible mechanism for changes in EEG activities [50]. One of the most important AD pathologic findings, tau aggregates, was observed first in the locus coeruleus (LC) [51]. Decreased noradrenergic projections from the LC are related to amyloid-induced toxicity, and its dysfunction was related with arousal, attention, and memory function [52]. Furthermore, the noradrenergic output from LC may be related to arousal modulation in the basal nucleus [49]. Therefore, the effects of tauopathy on LC during the early stage of AD and its dysfunctional noradrenergic transmission may also contribute to changes in alpha reactivity.

In the present study, the ARI differed between the Aβ+ and Aβ–groups in the left hemisphere. Topographic differences in alpha blocking have been suggested as a marker of laterality of brain lesion; that is, impaired alpha attenuation was lateralized to the side of the lesion, such as a cerebral infarction. The authors also proposed that may be an early manifestation or the only EEG presentation of a disturbance of cerebral function [53]. All participants in our study were right-handed, and important cognitive functions, such as memory and language, which are frequently impaired in patients with AD, mainly involve the dominant hemisphere. Defective cortical networks have been frequently reported in the left-brain region of AD dementia patients [54 –58]. Although EEG abnormalities in AD, including decreased alpha reactivity, do not consistently show left-sided dominance, our Aβ+ sample from an earlier stage of AD presented a lateralizing pattern, which was consistent with a previous study [21]. Whether this laterality is transient or consistent across disease progression should be more investigated with the exploration of underlying ipsilateral structural and functional changes in AD.

Previous attempts have been made to predict Aβ pathology in individuals without prominent cognitive decline using clinical variables. Most studies included neuropsychological, laboratory, and neuroimaging measures [59, 60]. We developed a model to predict Aβ pathology based on commonly available clinical information and APOE ɛ4 positivity. Reduced ARI acts as a significant predictor of Aβ pathology in addition to the factors included in the final model, and improved the prediction model fit in the MCI group. As the EC-to-EO ARI is also an easily available measurement in a clinical setting, it may be useful to distinguish Aβ positivity among patients with MCI. No additional contribution of ARI to the prediction of Aβ pathology was found in total sample, suggesting that the EC-to-EO ARI works as an Aβ predictor after the onset of cognitive symptoms even though it is not conclusive due to the small number of preclinical AD subjects in this study.

Several strengths of this study should be discussed. The study participants received comprehensive clinical assessments and imaging evaluation, including structural MRI and amyloid PET. We classified the participants based on pathological confirmation; therefore, the confounding effect of non-AD cases was minimized. Also, we evaluated and adjusted for various factors that commonly affect brain electrical activities, such as the degree of cerebrovascular burden and medication use. This study also had some potential limitations that should be discussed. First, as we included a relatively large number of patients with MCI, decreased alpha reactivity may have been affected by further neurodegeneration in addition to amyloid deposition during the course of AD. Although we controlled for the degree of WMH, other neurodegeneration markers, such as the structural and functional changes observed in AD, should be included in future studies to exclude the possibility of false positive findings. Second, only a crude topographic analysis based on nine ROIs was conducted in this study, as we used a 19-channel EEG recording; however, we examined the predictability of the EC-to-EO ARI as a marker for early AD in a common clinic setting. Third, the simply calculated alpha reactivity can be easily influenced by individual-level EEG variability, such as impaired resting state alpha activity. Therefore, it will be necessary to elaborate on EC-to-EO alpha reactivity considering the size of alpha suppression and the characteristics of alpha-blocking such as appearance time, duration, frequency, and variability. Fourth, multiple comparisons may increase the possibility of false positive discovery even though we combined individual ROIs into polarity and laterality factors for repeated-measures 4-way MANCOVA and its positive finding remained significant after Bonferroni correction for multiple comparisons (p < 0.017). Lastly, both WMH and BAPL scores were confirmed in the diagnostic consensus meeting to increase the concordance rate of evaluation; however, the evaluators were not blinded to the clinical condition of the participants.

Conclusion

In the present study, we examined whether older adults in the early stage of AD present with abnormal quantitative EEG parameters, particularly focusing on changes in the alpha rhythm from EC to EO. Pathologically confirmed early AD subjects showed decreased EC-to-EO alpha reactivity in the left hemisphere. In MCI group, this index contributed to prediction of cerebral Aβ deposition beyond the clinical factors, including vascular risks, impaired memory function, and APOE ɛ4. The EC-to-EO alpha reactivity may work as an early biomarker of AD and would be a useful measurement for screening a large elderly population with risk of AD. Further studies with a larger preclinical AD sample size, a more elaborate ARI, and concurrent analyses of functional brain changes are needed to confirm the specific effects of cerebral Aβ pathology on impaired alpha reactivity.

Footnotes

ACKNOWLEDGMENTS

This study was supported by grant No. 0320180150 from the SNUH Research Fund, the National Research Foundation of Korea (NRF) grant funded by the Korea government (Ministry of Science and ICT; No. 800-20200212), and a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI) funded by the Ministry of Health & Welfare, Republic of Korea (Grant No: HI19C0149).

Authors’ disclosures available online ().

References

Dubois

, Hampel

, Feldman

, Scheltens

, Aisen

, Andrieu

, Bakardjian

, Benali

, Bertram

, Blennow

, Broich

, Cavedo

, Crutch

, Dartigues

, Duyckaerts

, Epelbaum

, Frisoni

, Gauthier

, Genthon

, Gouw

, Habert

, Holtzman

, Kivipelto

, Lista

, Molinuevo

, O’Bryant

, Rabinovici

, Rowe

, Salloway

, Schneider

, Sperling

, Teichmann

, Carrillo

, Cummings

, Jack

Jr. (2016) Preclinical Alzheimer’s disease: Definition, natural history, and diagnostic criteria. Alzheimers Dement 12, 292–323.

Small

, Gagnon

, Robinson

(2007) Early identification of cognitive deficits: Preclinical Alzheimer’s disease and mild cognitive impairment. Geriatrics 62, 19–23.

Klunk

, Engler

, Nordberg

, Wang

, Blomqvist

, Holt

, Bergstrom

, Savitcheva

, Huang

, Estrada

, Ausen

, Debnath

, Barletta

, Price

, Sandell

, Lopresti

, Wall

, Koivisto

, Antoni

, Mathis

, Langstrom

(2004) Imaging brain amyloid in Alzheimer’s disease with Pittsburgh Compound-B. Ann Neurol 55, 306–319.

Maestu

, Cuesta

, Hasan

, Fernandez

, Funke

, Schulz

(2019) The importance of the validation of M/EEG with current biomarkers in Alzheimer’s disease. Front Hum Neurosci 13, 17.

Kramberger

, Kareholt

, Andersson

, Winblad

, Eriksdotter

, Jelic

(2013) Association between EEG abnormalities and CSF biomarkers in a memory clinic cohort. Dement Geriatr Cogn Disord 36, 319–328.

Colom-Cadena

, Spires-Jones

, Zetterberg

, Blennow

, Caggiano

, DeKosky

, Fillit

, Harrison

, Schneider

, Scheltens

, de Haan

, Grundman

, van Dyck

, Izzo

, Catalano

, Synaptic Health Endpoints Working Group (2020) The clinical promise of biomarkers of synapse damage or loss in Alzheimer’s disease. Alzheimers Res Ther 12, 21.

Adeli

, Ghosh-Dastidar

, Dadmehr

(2005) Alzheimer’s disease and models of computation: Imaging, classification, and neural models. J Alzheimers Dis 7, 187–199.

Jeong

(2004) EEG dynamics in patients with Alzheimer’s disease. Clin Neurophysiol 115, 1490–1505.

Kim

, Lee

, Park

, Kim

, Bae

, Kim

, Im

(2012) Clinical implications of quantitative electroencephalography and current source density in patients with Alzheimer’s disease. Brain Topogr 25, 461–474.

10.

Jung

, Lee

, Kim

, Lee

, Chung

(2007) Quantitative electroencephalography and low resolution electromagnetic tomography imaging of Alzheimer’s disease. Psychiatry Investig 4, 31–37.

11.

Roh

, Park

, Ko

, Park

, Lee

, Han

, Jo

, Yang

, Jung

(2011) Region and frequency specific changes of spectral power in Alzheimer’s disease and mild cognitive impairment. Clin Neurophysiol 122, 2169–2176.

12.

Smailovic

, Koenig

, Kareholt

, Andersson

, Kramberger

, Winblad

, Jelic

(2018) Quantitative EEG power and synchronization correlate with Alzheimer’s disease CSF biomarkers. Neurobiol Aging 63, 88–95.

13.

Teipel

, Bakardjian

, Gonzalez-Escamilla

, Cavedo

, Weschke

, Dyrba

, Grothe

, Potier

, Habert

, Dubois

, Hampel

, INSIGHT-preAD study group (2018) No association of cortical amyloid load and EEG connectivity in older people with subjective memory complaints. Neuroimage Clin 17, 435–443.

14.

Nakamura

, Cuesta

, Fernandez

, Arahata

, Iwata

, Kuratsubo

, Bundo

, Hattori

, Sakurai

, Fukuda

, Washimi

, Endo

, Takeda

, Diers

, Bajo

, Maestu

, Ito

, Kato

(2018) Electromagnetic signatures of the preclinical and prodromal stages of Alzheimer’s disease. Brain 141, 1470–1485.

15.

Babiloni

, Lopez

, Del Percio

, Noce

, Pascarelli

, Lizio

, Teipel

, Gonzalez-Escamilla

, Bakardjian

, George

, Cavedo

, Lista

, Chiesa

, Vergallo

, Lemercier

, Spinelli

, Grothe

, Potier

, Stocchi

, Ferri

, Habert

, Fraga

, Dubois

, Hampel

(2020) Resting-state posterior alpha rhythms are abnormal in subjective memory complaint seniors with preclinical Alzheimer’s neuropathology and high education level: The INSIGHT-preAD study. Neurobiol Aging 90, 43–59.

16.

Babiloni

, Lizio

, Vecchio

, Frisoni

, Pievani

, Geroldi

, Claudia

, Ferri

, Lanuzza

, Rossini

(2010) Reactivity of cortical alpha rhythms to eye opening in mild cognitive impairment and Alzheimer’s disease: An EEG study. J Alzheimers Dis 22, 1047–1064.

17.

López-Sanz

, Bruña

, Garcés

, Camara

, Serrano

, Rodríguez-Rojo

, Delgado

, Montenegro

, López-Higes

, Yus

, Maestú

(2016) Alpha band disruption in the AD-continuum starts in the subjective cognitive decline stage: A MEG study. Sci Rep 6, 37685.

18.

Bazanova

, Vernon

(2014) Interpreting EEG alpha activity. Neurosci Biobehav Rev 44, 94–110.

19.

Miraglia

, Vecchio

, Bramanti

, Rossini

(2016) EEG characteristics in “eyes-open” versus “eyes-closed” conditions: Small-world network architecture in healthy aging and age-related brain degeneration. Clin Neurophysiol 127, 1261–1268.

20.

Wan

, Huang

, Schwab

, Tanner

, Rajan

, Lam

, Zaborszky

, Li

, Price

, Ding

(2019) From eyes-closed to eyes-open: Role of cholinergic projections in EC-to-EO alpha reactivity revealed by combining EEG and MRI. Hum Brain Mapp 40, 566–577.

21.

Fonseca

, Tedrus

GMAS

, Fondello

, Reis

, Fontoura

(2011) EEG theta and alpha reactivity on opening the eyes in the diagnosis of Alzheimer’s disease. Clin EEG Neurosci 42, 185–189.

22.

Lee

, Lee

, Kim

, Jhoo

, Kim

, Lee

, Kim

, Han

, Woo

(2002) Development of the Korean version of the Consortium to Establish a Registry for Alzheimer’s Disease Assessment Packet (CERAD-K): Clinical and neuropsychological assessment batteries.P. J Gerontol B Psychol Sci Soc Sci 57, 47–53.

23.

Frances

, Galanter

, Kleber

(2000) Diagnostic and Statistical Manual of Mental Disorders: DSM-IV-TR®, American Psychiatric Association Publishing.

24.

Jack

Jr. , Bennett

, Blennow

, Carrillo

, Dunn

, Haeberlein

, Holtzman

, Jagust

, Jessen

, Karlawish

, Liu

, Molinuevo

, Montine

, Phelps

, Rankin

, Rowe

, Scheltens

, Siemers

, Snyder

, Sperling

, Contributors (2018) NIA-AA Research Framework: Toward a biological definition of Alzheimer’s disease. Alzheimers Dement 14, 535–562.

25.

Teng

, Ringman

, Ross

, Mulnard

, Dick

, Bartzokis

, Davies

, Galasko

, Hewett

, Mungas

, Reed

, Schneider

, Segal-Gidan

, Yaffe

, Cummings

, Alzheimer’s Disease Research Centers of California-Depression in Alzheimer’s Disease Initiative (2008) Diagnosing depression in Alzheimer disease with the national institute of mental health provisional criteria. Am J Geriatr Psychiatry 16, 469–477.

26.

American Psychiatric Association (1994) Diagnostic and statistical manual of mental disorders, 4th ed. American Pyshicatric Association, Washington DC.

27.

Hamilton

(1960) A rating scale for depression. J Neurol Neurosurg Psychiatry 23, 56–62.

28.

Park

, Jang

, Kim

, Jun

, Lee

, Kim

, Yim

, Park

(2017) Clinical validation of the Psychotic Depression Assessment Scale, Hamilton Depression Rating Scale-6, and Brief Psychiatric Rating Scale-5: Results from the Clinical Research Center for Depression Study. Psychiatry Investig 14, 568–576.

29.

Boustani

, Campbell

, Munger

, Maidment

, Fox

(2008) Impact of anticholinergics on the aging brain: A review and practical application. Aging Health 4, 311–320.

30.

Babiloni

, Barry

, Basar

, Blinowska

, Cichocki

, Drinkenburg

, Klimesch

, Knight

, Lopes da Silva

, Nunez

, Oostenveld

, Jeong

, Pascual-Marqui

, Valdes-Sosa

, Hallett

(2020) International Federation of Clinical Neurophysiology (IFCN) - EEG research workgroup: Recommendations on frequency and topographic analysis of resting state EEG rhythms. Part 1: Applications in clinical research studies. Neurophysiol 131, 285–307.

31.

Zoubir

, Koivunen

, Chakhchoukh

, Muma

(2012) Robust estimation in signal processing: A tutorial-style treatment of fundamental concepts. IEEE Signal Process Mag 29, 61–80.

32.

Fazekas

, Chawluk

, Alavi

, Hurtig

, Zimmerman

(1987) MR signal abnormalities at 1.5 T in Alzheimer’s dementia and normal aging. AJR Am J Roentgenol 149, 351–356.

33.

Van Straaten

, Fazekas

, Rostrup

, Scheltens

, Schmidt

, Pantoni

, Inzitari

, Waldemar

, Erkinjuntti

, Maäntylaä

(2006) Impact of white matter hyperintensities scoring method on correlations with clinical data: The LADIS study. Stroke 37, 836–840.

34.

Seibyl

, Catafau

, Barthel

, Ishii

, Rowe

, Leverenz

, Ghetti

, Ironside

, Takao

, Akatsu

, Murayama

, Bullich

, Mueller

, Koglin

, Schulz-Schaeffer

, Hoffmann

, Sabbagh

, Stephens

, Sabri

(2016) Impact of training method on the robustness of the visual assessment of 18F-Florbetaben PET scans: Results from a phase-3 study. J Nucl Med 57, 900–906.

35.

Barthel

, Gertz

, Dresel

, Peters

, Bartenstein

, Buerger

, Hiemeyer

, Wittemer-Rump

, Seibyl

, Reininger

, Sabri

(2011) Cerebral amyloid-β PET with florbetaben (18F) in patients with Alzheimer’s disease and healthy controls: A multicentre phase 2 diagnostic study. Lancet Neurol 10, 424–435.

36.

Barry

, Clarke

, Johnstone

, Magee

, Rushby

(2007) EEG differences between eyes-closed and eyes-open resting conditions. Clin Neurophysiol 118, 2765–2773.

37.

Barry

, De Blasio

(2017) EEG differences between eyes-closed and eyes-open resting remain in healthy ageing. Biol Psychol 129, 293–304.

38.

Tabachnick

, Fidell

(2013) Using Multivariate Statistics, 6th Edn, New International Edition. Pearson Education Limited, Harlow.

39.

Dierks

, Perisic

, Frölich

, Ihl

, Maurer

(1991) Topography of the quantitative electroencephalogram in dementia of the Alzheimer type: Relation to severity of dementia. Psychiatry Res 40, 181–194.

40.

Pritchard

, Duke

, Coburn

, Moore

, Tucker

, Jann

, Hostetler

(1994) EEG-based, neural-net predictive classification of Alzheimer’s disease versus control subjects is augmented by non-linear EEG measures. Electroencephalogr Clin Neurophysiol 91, 118–130.

41.

Besthorn

, Zerfass

, Geiger-Kabisch

, Sattel

, Daniel

, Schreiter-Gasser

, Förstl

(1997) Discrimination of Alzheimer’s disease and normal aging by EEG data. Electroencephalogr Clin Neurophysiol 103, 241–248.

42.

Jansen

, Ossenkoppele

, Knol

, Tijms

, Scheltens

, Verhey

, Visser

, Aalten

, Aarsland

, Alcolea

(2015) Prevalence of cerebral amyloid pathology in persons without dementia: A meta-analysis. JAMA 313, 1924–1938.

43.

Pfurtscheller

, Da Silva

(1999) Event-related EEG/MEG synchronization and desynchronization: Basic principles. Clin Neurophysiol 110, 1842–1857.

44.

Babiloni

, Del Percio

, Boccardi

, Lizio

, Lopez

, Carducci

, Marzano

, Soricelli

, Ferri

, Triggiani

(2015) Occipital sources of resting-state alpha rhythms are related to local gray matter density in subjects with amnesic mild cognitive impairment and Alzheimer’s disease. Neurobiol Aging 36, 556–570.

45.

Bartus

, Dean

, Beer

, Lippa

(1982) The cholinergic hypothesis of geriatric memory dysfunction. Science 217, 408–414.

46.

Arendt

, Bigl

, Tennstedt

, Arendt

(1985) Neuronal loss in different parts of the nucleus basalis is related to neuritic plaque formation in cortical target areas in Alzheimer’s disease. Neuroscience 14, 1–14.

47.

Beach

, McGeer

(1992) Senile plaques, amyloid β-protein, and acetylcholinesterase fibres: Laminar distributions in Alzheimer’s disease striate cortex. Acta Neuropathol 83, 292–299.

48.

Davis

, Mohs

, Marin

, Purohit

, Perl

, Lantz

, Austin

, Haroutunian

(1999) Cholinergic markers in elderly patients with early signs of Alzheimer disease. JAMA 281, 1401–1406.

49.

Mesulam

(2013) Cholinergic circuitry of the human nucleus basalis and its fate in Alzheimer’s disease. J Comp Neurol 521, 4124–4144.

50.

Dringenberg

(2000) Alzheimer’s disease: More than a ‘cholinergic disorder’—evidence that cholinergic–monoaminergic interactions contribute to EEG slowing and dementia. Brain Res 115, 235–249.

51.

Mather

, Harley

(2016) The locus coeruleus: Essential for maintaining cognitive function and the aging brain. Trends Cogn Sci 20, 214–226.

52.

Kelly

, He

, Perez

, Ginsberg

, Mufson

, Counts

(2017) Locus coeruleus cellular and molecular pathology during the progression of Alzheimer’s disease. Acta Neuropathol Commun 5, 8.

53.

Westmoreland

, Klass

(1998) Defective alpha reactivity with mental concentration. J Clin Neurophysiol 15, 424–428.

54.

Jelic

, Julin

, Shigeta

, Nordberg

, Lannfelt

, Winblad

, Wahlund

(1997) Apolipoprotein E epsilon4 allele decreases functional connectivity in Alzheimer’s disease as measured by EEG coherence. J Neurol Neurosurg Psychiatry 63, 59–65.

55.

Adler

, Brassen

, Jajcevic

(2003) EEG coherence in Alzheimer’s dementia. J Neural Transm (Vienna) 110, 1051–1058.

56.

Babiloni

, Frisoni

, Pievani

, Vecchio

, Lizio

, Buttiglione

, Geroldi

, Fracassi

, Eusebi

, Ferri

, Rossini

(2009) Hippocampal volume and cortical sources of EEG alpha rhythms in mild cognitive impairment and Alzheimer disease. Neuroimage 44, 123–135.

57.

Canuet

, Tellado

, Couceiro

, Fraile

, Fernandez-Novoa

, Ishii

, Takeda

, Cacabelos

(2012) Resting-state network disruption and APOE genotype in Alzheimer’s disease: A lagged functional connectivity study. PLoS One 7, e46289.

58.

Babiloni

, Del Percio

, Lizio

, Noce

, Lopez

, Soricelli

, Ferri

, Pascarelli

, Catania

, Nobili

, Arnaldi

, Fama

, Orzi

, Buttinelli

, Giubilei

, Bonanni

, Franciotti

, Onofrj

, Stirpe

, Fuhr

, Gschwandtner

, Ransmayr

, Garn

, Fraioli

, Pievani

, D’Antonio

, De Lena

, Guntekin

, Hanoglu

, Basar

, Yener

, Emek-Savas

, Triggiani

, Taylor

, De Pandis

, Vacca

, Frisoni

, Stocchi

(2018) Functional cortical source connectivity of resting state electroencephalographic alpha rhythms shows similar abnormalities in patients with mild cognitive impairment due to Alzheimer’s and Parkinson’s diseases. Neurophysiol 129, 766–782.

59.

Bahar-Fuchs

, Villemagne

, Ong

, Chetélat

, Lamb

, Reininger

, Woodward

, Rowe

(2013) Prediction of amyloid-β pathology in amnestic mild cognitive impairment with neuropsychological tests. J Alzheimers Dis 33, 451–462.

60.

Haghighi

, Smith

, Morgan

, Small

, Huang

, for the Alzheimer’s Disease Neuroimaging Initiative (2015) Identifying cost-effective predictive rules of amyloid-β level by integrating neuropsychological tests and plasma-based markers. J Alzheimers Dis 43, 1261–1270.