Sage Journals: Discover world-class research

Abstract

Background

Alzheimer's disease (AD) accounts for the majority of dementia-related disorders among aging populations. Given the irreversible nature of AD, early detection of cognitive impairment is critical for improving prognosis and facilitating timely interventions before individuals meet the clinical criteria for AD.

Objective

The objective of this review is to provide a comprehensive summary of recent advances in task-based electroencephalography (EEG), such as using attention and inhibitory control tasks, which has recently emerged as a promising non-invasive biomarker for assessing neurophysiological alterations associated with AD and mild cognitive impairment (MCI).

Methods

This systematic review evaluates the efficacy of task-based EEG biomarkers in distinguishing cognitively impaired individuals from those without impairment. A comprehensive literature search was conducted across PubMed, Semantic Scholar, and SpringerLink databases for studies published between 2017 and 2024.

Results

Findings indicate consistent neurophysiological alterations in MCI and AD, particularly reductions in event-related potential amplitudes and prolonged latencies, with P3 abnormalities observed in about half of the studies assessing selective attention and inhibitory control. Similarly, a comparable number of studies using working memory tasks report disrupted functional connectivity patterns, increased low-frequency oscillations (delta and theta activity), and reduced fast oscillations (alpha and beta activity).

Conclusions

These EEG-based indices demonstrate potential as objective biomarkers for detecting neural alterations associated with cognitive decline in both AD and preclinical dementia stages. Further research is needed to standardize EEG protocols and validate their clinical utility for early diagnosis and disease progression monitoring.

Keywords

Alzheimer's disease biomarkers cognitive dysfunction electroencephalography electroencephalography phase synchronization evoked potentials systematic review

Introduction

Dementia is a progressive neurodegenerative disease and a major cause of disability and dependency in aging societies. It is currently recognized as the seventh leading cause of death worldwide.¹ Dementia is marked by a progressive decline in cognitive functions, affecting memory, attention, reasoning, visual-spatial abilities, learning, and judgment. While age is a major risk factor, dementia is not an inevitable part of aging. As the global population grows older, dementia has become an urgent healthcare challenge, further intensified by the lack of a cure. Therefore, there is a pressing need to develop early diagnostic tools that can enable timely interventions before significant brain deterioration occurs. Cognition associated with memory, attention, executive function, language, learning, perceptual-motor function, and social cognition typically declines with advancing age.² Dementia represents a significant decline in one or more cognitive domains, with memory being among the first affected in many common types of disease. As such, memory impairment is often considered a potential early marker of mild cognitive impairment (MCI), a condition that carries a heightened risk of progressing to Alzheimer's disease (AD).³ In addition to memory loss, attention—the ability to filter relevant information, ignore distractions, and guide decision-making in response to environmental changes—is also among the earliest cognitive functions disrupted in the initial stages of AD.^4,5 Moreover, even in the very early phase of the disease, patients with AD exhibit impaired executive function, which is a comprehensive term encompassing the cognitive processes, including reasoning, problem-solving, and regulation of behavior.^6,7 Within the non-memory domain, attention and executive function impairment precedes observable deficits in language and visuospatial abilities.^8,9 Cognitive decline often precedes the clinical diagnosis of dementia by many years, highlighting the need for tools capable of distinguishing the early transitions along the AD continuum, from healthy aging to MCI and from MCI to dementia. In clinical practice, this need has traditionally been addressed through standardized screening instruments. The Mini-Mental State Examination (MMSE) and Montreal Cognitive Assessment (MoCA) are commonly used screening tools, with MoCA being superior for the identification of MCI.^10–14 Alternative screening tests include the Addenbrooke's Cognitive Examination Revised (ACE-R)¹⁵ and the Mini-Cog test.¹⁶ Validated neuropsychological assessments such as the Trail Making test,¹⁷ the Clock Drawing test,¹⁸ and the Rey-Osterrieth Complex Figure test¹⁹ are also used to evaluate specific cognitive domains. Nevertheless, these conventional pen-and-paper tests demand extensive training to ensure they are administered uniformly, and their findings are interpreted accurately.

MCI represents a transitional stage between the subtle cognitive changes of normal aging and the more pronounced decline characteristic of dementia, most often AD, which accounts for 60–70% of dementia cases worldwide.^1,20 With advances in neurotechnology for age-related cognitive intervention,²¹ the demand for sensitive and accessible tools to detect the earliest stage of AD is becoming increasingly urgent. Identifying biomarkers that can reliably predict the progression from preclinical MCI to overt cognitive decline and ultimately to AD is therefore critical.²² A deeper understanding of the how AD's pathophysiological processes give rise to clinical symptoms will be essential for the development of effective disease-modifying therapies and for improving early detection of both MCI and AD.

Detecting the progression of AD before the onset of brain damage is essential. Biomarkers offer non-invasive methods for diagnosing and evaluating disease progression. While several types of biomarkers, including biochemical, genetic, and neuroimaging biomarkers, are widely used in clinical research, neuroimaging techniques can be costly, invasive, and typically detect disease-related damage only in advanced stages. Additionally, biomarkers derived from laboratory tests, including blood tests, can be unstable due to variability in levels influenced by external factors such as age, health status, and differences in laboratory methods, leading to inconsistent results. Biochemical biomarkers like amyloid-β (Aβ) and tau protein found in the cerebrospinal fluid (CSF) are hallmarks of AD pathophysiology.²³ In contrast, genetic biomarkers like the ε4 allele of Apolipoprotein E (APOE) indicate an additional risk of developing AD.²⁴ AD is associated with increased Aβ and tau levels in the CSF, which makes both of them potential biomarkers for AD diagnosis.²⁵ Neuroimaging biomarkers such as positron emission tomography (PET) and magnetic resonance imaging (MRI) reveal specific changes that characterize AD.^26,27 Although amyloid PET has good sensitivity, it lacks specificity for diagnosing different dementia subtypes, and involves exposure to radiation.²⁸ Alternatively, MRI provides anatomical information, including the detection of structural changes associated with neurodegenerative disease such as brain atrophy, but it is not specific for AD.²⁹ While biochemical and genetic biomarkers reveal levels of specific compounds in the brain or body fluids, indicating biological processes or responses to interventions, neuroimaging biomarkers are non-invasive methods used to determine specific changes that characterize AD. In contrast to biochemical biomarkers, electroencephalography (EEG) offers neural activity with high temporal resolution, allowing for the identification of functional brain alterations and network dynamics in AD and MCI. Therefore, measures of neural oscillatory activity, event-related potentials (ERPs), as well as connectivity and complexity measures as revealed in task-based EEG are both promising candidate biomarkers for AD. Consequently, task-based EEG holds, for example, potential utility as a pre-screening tool prior to MRI implementation.

This systematic review aimed to identify studies on neural oscillatory activity, ERPs, and connectivity and complexity measures reflected in task-based EEG, investigating their potential as biomarkers for identifying MCI and AD. Through a systematic exploration of empirical studies, the review seeks to clarify the effectiveness and reliability of task-based EEG biomarkers in identifying distinct patterns that can aid in the early differentiation and potential prognostication of cognitive decline along the continuum from MCI to AD-related dementia. This review synthesizes recent findings on task-based EEG biomarkers and highlights their emerging role in the early detection and monitoring of MCI and AD, offering timely updates in a field where such insights are urgently needed Our systematic review will be critical for shaping future therapeutic strategies, as early detection could enable timely access to treatments and empower individuals to adopt practices that may slow disease progression.²¹ In line with this goal, this work presents recent advances in task-based EEG to guide biomarker development and deepen understanding of MCI and AD neurophysiology.

Methods

Search strategies

We adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines.³⁰ The PubMed, Semantic Scholar, and SpringerLink online databases were searched to identify articles published in English between 2017 and 2024. This review focused on recent publications to capture current EEG methodologies and developments. Articles reporting task-based EEG biomarkers for cognitive impairment among patients with MCI or AD and in cognitively normal individuals, were included. The search strategy involved the following elements:

▪ Keywords: Four main keywords were established to retrieve articles from the databases. The first two keywords were “EEG” and “cognitive” which had to co-occur with a third keyword related to patient groups, such as “Alzheimer's disease,” “mild cognitive impairment,” or its abbreviation “MCI”. A fourth keyword defined specific cognitive tasks used to assess cognitive functions, including “oddball,” “attention inhibitory,” “working memory,” or “decision-making”.

▪ Used Combinations: Boolean operators were employed, using AND between all four main keywords and OR within the third and fourth keywords. For example: “EEG” AND “cognitive” AND (“Alzheimer's” OR “mild cognitive impairment” OR “MCI”) AND (“oddball” OR “attention inhibitory” OR “working memory” OR “decision-making”). All identified keywords and Boolean operators were adapted to the specific search features of each database.

▪ Filters: Filters were applied for publication date (2017–2024) and language (English). Additionally, filters for human subjects were applied in certain applicable databases.

Eligibility criteria

The inclusion criteria were limited to original research articles, excluding case reports, systematic reviews, and meta-analyses. The search was restricted to articles published in peer-reviewed journals, while other publication types, such as conference proceedings papers, preprints, letters, editorials, and books, were not considered. Studies met the following criteria were included in the review: 1) involved older adult populations, including individuals meeting standardized diagnostic criteria for AD or MCI, as well as those identified as at-risk MCI based on cognitive scores assessed with validated neuropsychological assessments; and 2) analyzed EEG under cognitive tasks. Meanwhile, the following studies were excluded: 1) populations with degenerative disorders in which dementia was not the primary manifestation of the disorder, such as dementia in Huntington's disease, Parkinson's disease with dementia, post-stroke dementia, and secondary dementias; 2) studies limited to healthy participants without reported memory or cognitive impairment; 3) studies on other forms of dementia or presymptomatic stage; 4) studies focusing on resting-state EEG or the effect of training or intervention on cognitive ability; and 5) studies on the development of mathematical foundations of machine learning, as they did not primarily underscore the aspects related to biomarkers for disorder detection.

Article selection

Our search terms initially identified one thousand nine hundred thirty-eight articles (n = 1938) from the PubMed, Semantic Scholar, and SpringerLink databases. Among the identified records, one thousand five hundred forty-three (n = 1543) were removed due to duplication. Furthermore, one hundred and eight studies (n = 108) were excluded for varying reasons, including fifty-four conference proceedings and/or abstract book titles, nine editorial and perspective articles, twenty-seven preprint records, and eighteen in-scope systematic reviews. Next, the titles and abstracts of the two hundred eighty-seven (n = 287) were screened according to the established inclusion and exclusion criteria in the first stage of the screening process. Out of these, fifty-six studies explored various methods for dementia screening without using EEG. Additionally, seventy studies were found to be irrelevant to dementia or focused on different clinical conditions, such as post-stroke dementia, Parkinson's disease, attention deficit hyperactivity disorder, schizophrenia, epilepsy, depression, and autism. Also, twenty-three studies were conducted on the young adult population, and three focused on tracing the developmental trajectories evident across the lifespan. Furthermore, thirty-one studies addressed resting-state EEG, while fourteen articles predominantly focused on developing machine learning models. Moreover, twenty-three records examined the influences of medication, interventions, and/or commercial product outcomes. As a result, a total of two hundred twenty articles (n = 220) were excluded from the analysis due to their failure to meet the predetermined inclusion criteria, and sixty-seven articles (n = 67) were selected for full-text review, independently screened by two reviewers (KT and AW). Out of these full-text articles, twenty-three studies (n = 23) were subsequently excluded due to the following reasons: thirteen studies on individuals with other forms of dementia or presymptomatic stage; one study involving animals; six studies on multimodal measures, which involved a combination of different approaches and could lead to deviations into non-related EEG research; and three studies with a small sample size (less than ten participants in the cognitively impaired group). Notably, to ensure a comprehensive identification of relevant studies, the reference lists of full-text articles were independently screened by KT and AW. This practice, as recommended in systematic review manuals and recent publications,^22,31 proved effective yielding 30 additional articles (n = 30) that were incorporated into the review. These studies were selected based on their relevance and contributions to understanding EEG biomarkers in MCI and AD, following manual searches of the references in the identified articles. All included studies met our predefined inclusion and exclusion criteria. Hence a final number of seventy-four studies (n = 74) was included in the review. The PRISMA flow diagram of search results has been depicted in Figure 1.

Figure 1.

PRISMA flow diagram describing the selection process of the included studies. This output plot was generated via the https://estech.shinyapps.io/prisma_flowdiagram/.³²

Results

Task-based EEG associated with cognitive function

A detailed summary of EEG biomarkers across cognitive tasks and patient populations is provided in Table 1, while overview of the changes in ERPs and oscillatory synchrony in AD and MCI presented in this review is shown in Tables 2 and 3.

Table 1.

Summary of EEG biomarkers across cognitive tasks and patient groups.

Cognitive function	Task	Group	EEG biomarkers
Selective attention	Oddball, Fastball	AD	Decreased P3 amplitude, increased P3 latency Decreased SNR during recognition condition
Selective attention	Oddball, Fastball	MCI	Increased P50 amplitude for auditory stimuli Decreased N1 amplitude for olfactory stimuli Decreased delta ERO Prefrontal ERP features
Attentional and inhibitory control	Go/NoGo, Flanker, CPT	AD	PSD, skewness, kurtosis, SS, SK, SCF, FD, SE features*
Attentional and inhibitory control	Go/NoGo, Flanker, CPT	MCI	Decreased N2 amplitude, increased N2 latency Alteration in P3 amplitude+, increased P3 latency Decreased theta power at NoGo condition, decreased alpha power at Go condition Increased theta ERS at Go condition and absence of alpha ERD difference between Go and NoGo conditions for late MCI group‡ Pre-stimulus power and difference in the pre-stimulus alpha power between fast and slow response trials features
Attention, Working memory	N-back, DMS, Sternberg, Recognition memory, Visuospatial memory/Navigation	AD	Decreased P450 amplitude Decreased alpha ERD Increased delta and theta power Alterations in alpha power+ Decreased connectivity at left hemisphere Increased repetition effect at posterior sites
Attention, Working memory		MCI	Decreased P2, P3, P450, PNwm, and N3 amplitude Increased P2 and N2 latency Decreased left frontal ERP +Alterations in theta, alpha, beta power Decreased alpha ERD Decreased beta coherence Decreased connectivity and complexity Remained repetition effect
Attention, Auditory processing, Working memory	Auditory detection task	MCI	Increased P50 amplitude A0 (extracted with LDA) and ST4 (extracted with PCA) features
Attention, Perceptual processing, Working memory	Strategic processing task/ object discrimination task	AD	Decreased N2 and P3 amplitude
Attention, Perceptual processing, Working memory	Strategic processing task/ object discrimination task	MCI	Decreased P3 amplitude, increased P3 latency +Decreased theta power ERP (N1, P1, P2) and spectral (relative power) features
Decision-making, Working memory	Gambling/reward processing	MCI	Increased FRN amplitude
Attention, Language processing, Working memory	Word repetition VR-based language	AD	Visibility graphs features*
Attention, Language processing, Working memory	Word repetition VR-based language	MCI	Decreased N400 amplitude Decreased P600 repetition effect and alpha repetition effect Combined features (frequency domain EEG features: mean power, signal energy, SE, asymmetric features; speech features: acoustic, rhythmic features; digital cognitive features)
Emotional regulation, Working memory	Cognitive reappraisal	MCI	Decreased LPP difference between more neutral (Rea condition) and more negative (Neg condition)

‡ within-group, between-conditions; + Conflicting results have been reported, with some studies indicating an increase while others demonstrate a decrease. * This biomarker found in both AD and MCI groups.

AD: Alzheimer's disease; CPT: continuous performance task; DMS: delayed match-to-sample; ERD/ERS: event-related desynchronization/synchronization; ERO: event-related oscillation; FRN: feedback-related negativity; MCI: mild cognitive impairment; LPP: late positive potential; PSD: power spectral density; SCF: spectral crest factor; SE: spectral entropy; SK: spectral kurtosis; SNR: signal-to-noise ratio; SS: spectral skewness.

Table 2.

Overview of the changes in amplitude and latency of ERP components in ad and MCI presented based on the articles included in this review.

ERP component	Amplitude in AD	Latency in AD	Amplitude in MCI	Latency in MCI
P50	*³⁶	–	↑^46,108	*⁴⁶
N1, N160	–	–	↓⁴⁵ *^75,91,109	*^45,74
P1	–	–	*⁷⁵	*⁷⁴
P2	–	↑⁷⁶	*⁹¹	–
N2, N2pc	↓⁵⁷	↑^36,76 *⁵⁷	↓^58,59 *^60,61,91,92	↑^59–61 *^58,83
P3	↑⁸⁴ ↓^{34,35,57,71,83,91}	↑^34–36 *^57,58,83	↑^62,84 ↓^{69,70,100,109} *¹¹⁰	↑^{34,59,61,71,100} *^62,69
N400, P600 or LPC, LPP	↓¹⁰⁵	–	↑^45,87 ↓⁸⁶^102–104^,107	↑¹⁰²
P450	↓^72,73	–	↓⁷³	–
PNwm	–	–	↓^74,75	–
CDA	↓⁹²	–	↑⁹²	–
Left frontal ERP	–	–	↑ in non-target condition and ↓ in target condition ^84,85	–
FRN	–	–	↑⁹⁹	*⁹⁹

↑: increase; ↓: decrease; *: no significant difference; –: data not available/reported; AD: Alzheimer's disease; CDA: contralateral delay activity; ERP: event-related potential; FRN: feedback-related negativity; MCI: mild cognitive impairment; LPC: late positive component; LPP: late positive potential; PNwm: positive-negative working memory.

Table 3.

Overview of the oscillatory synchrony in AD and MCI.

Frequency	AD			MCI
Frequency	Power spectrum (evoked power, ERSP)	ERS/ERD	ERO	Power spectrum (evoked power, ERSP)	ERS/ERD	ERO
Delta	↑⁹³	–	↓³⁹	↓¹¹¹ ↑⁹³	–	↓^40,41,42
Theta	↓⁹⁴ ↑⁹³	↓ERD⁷³	–	↓^63,95 ↑^93,98	↓ERD⁷³ ↑ERS⁶⁵
Alpha	↑^79,93↓⁹⁴	↓ERD⁷³	–	↓^63,64,82↑⁹³Absence of alpha ERD difference⁶⁵; alpha suppression^95,104	↓ERD^65,75,73↓ERS⁷⁷	↓⁴⁰*
Beta	–	↓ERS⁷⁶	–	↑⁹⁸	↓ERS⁷⁷	*⁴³
Gamma	–	–	↓amplitude⁴⁴↑latency⁴⁴	–	–	–

↑: increase; ↓: decrease; *: no significant difference; –: data not available/reported; AD: Alzheimer's disease; ERD/ERS: event-related desynchronization/synchronization; ERO: event-related oscillation; ERSP: event-related spectral perturbation; MCI: mild cognitive impairment.

In attention and target detection tasks, such as the oddball paradigm, the task design consists of a train of frequent irrelevant stimuli (standard/non-target) interspersed with infrequent task-relevant stimuli (deviant/target or “oddball”). See Supplemental Table 1 for a summary of oddball-paradigm studies. Generally, frequent non-target auditory stimuli elicit early sensory ERP components such as P50, N1, and P2, whereas infrequent target stimuli elicit later components, most prominently N2 and P3.³³ Within this paradigm, MCI and AD groups consistently demonstrated reduced P3 amplitude and prolonged latency compared to HCs,^34–36 with P3 latency showing the strongest correlation with AD severity (r = –0.512).³⁶ A preclinical MCI group, who had not yet been clinically diagnosed but whose neuropsychological screening indicated risk for developing MCI, also exhibited reduced P3 amplitude.³⁷ Notably, selective attention ERP features related to the P3 response demonstrated significant clinical utility, effectively classifying AD cases from healthy controls with effect sizes ranging from small to large (0.349–0.521 and 2.130–2.364), and achieving high predictive accuracy (AUC = 0.891) that outperformed traditional clinical measures such as the MMSE (AUC = 0.803).³⁸ Event-related oscillations (EROs), phase- or time-locked to a sensory or cognitive event obtained by filtering ERP signals, have also been proposed as biomarkers for detecting subtle cognitive process abnormalities in AD. Studies have reported that delta-range EROs derived from the P3 response were reduced in both AD³⁹ and MCI patients,^40–42 whereas alpha and beta EROs showed no group differences.^40,43 AD patients also showed reduced low-gamma (25–30 Hz) sensory-evoked oscillation amplitude over the left hemisphere and delayed gamma ERO latency compared to HCs.⁴⁴ Other ERP components have also been investigated. The N1 component, an indicator of cortical sensory processing, showed smaller amplitude in individuals with MCI compared to HCs.⁴⁵ In contrast, the P50 amplitude was elevated in MCI patients, and this increase was strongly associated with metabolic abnormalities in the left frontal cortex, suggesting that altered sensory gating responses may reflect underlying neuronal dysfunction.⁴⁶ Similarly, the late positive component (LPC) amplitude was enhanced in MCI, indicating alterations in memory encoding and retrieval processes.⁴⁵ Beyond ERP findings, connectivity analyses revealed decreased low-frequency coherence in AD,⁴⁷ increased theta coherence in MCI,⁴⁸ and disrupted alpha, beta, and gamma synchronization in both AD and amnestic mild cognitive impairment (aMCI).⁴⁹ Additionally, network-level synchronization parameters, including phase-locking value and amplitude coherence, demonstrated strong discriminative capacity, achieving classification accuracies exceeding 90% in distinguishing healthy controls from MCI and in differentiating AD from MCI.⁴⁹ Moreover, steady-state responses evoked during Fastball task effectively distinguished recognition memory performance between AD and HC groups, with a high AUC of 0.86.⁵⁰

In response inhibition tasks—such as Go/NoGo, Stroop, flanker, stop-signal, and continuous performance task (CPT)—executive dysfunction can be observed through impairments in conflict monitoring and conflict resolution,⁵¹ two core processes of cognitive control. Findings from these tasks are summarized in Supplemental Table 2. Among these paradigms, the Go/NoGo task is one of the most widely used methods for assessing response inhibition, requiring participants to quickly respond to certain stimuli (Go stimuli) and withhold responses to others (NoGo stimuli). N2 and P3 are two key ERP components linked to response inhibition: N2 is typically elicited during the NoGo condition and is associated with inhibitory processes,^52,53 whereas P3 is maximally generated in centroparietal areas during the Go condition and in frontocentral areas during the NoGo condition.^54,55 These neural alterations are also observed in older adults, as response inhibition becomes less efficient with advancing age.⁵⁶ Additionally, paradigms that engage similar inhibitory control mechanisms—requiring selective attention and the ability to suppress responses to irrelevant stimuli—include tasks such as value-directed list learning and attentionally cued orientation discrimination. EEG studies have demonstrated that in such tasks, executive dysfunction in AD and aMCI, irrespective of subtype distinctions such as multi-domain (mdaMCI) and single-domain (sdaMCI), when compared with HCs, is indexed by reduced N2 amplitude,^57–59 and prolonged N2 latency in aMCI.^59–61 Aging appeared to magnify this effect, showing a stronger correlation between age and N2 latency in individuals with MCI than in healthy controls (r = 0.709 and 0.234, respectively).⁶⁰ Significant group differences in N2 latency (Cohen's d = 1.14) and amplitude (Cohen's d = 0.76) were reported between MCI and controls, whereas differences between MCI subtypes were comparatively smaller (Cohen's d = 0.33).⁵⁹ This pattern was also evident in N2 amplitude, showing a large effect size for differences between MCI and control groups (Cohen's d = 0.76) and a smaller effect between MCI subtypes (Cohen's d = 0.39).⁵⁹ Moreover, P3 measures demonstrated strong discriminative value, effectively distinguishing cognitively impaired individuals from healthy controls (Cohen's d ranging from 0.98 to 1.08; $η_{p}^{2}$ = 0.156, AUC = 0.84),^34,37,62 and also differentiating between MCI subtypes.⁶² Consistent with this, reductions in P3 amplitude were commonly observed in AD,⁵⁷ whereas elevated P3a responses were occasionally been reported in MCI, particularly in multi-domain subtypes.⁶² Oscillatory analyses revealed reduced theta synchronization^63,64 and diminished alpha ERD⁶⁵ in MCI. Furthermore, alterations in theta power demonstrated a moderate discriminatory effect between cognitively healthy and impaired groups ( $η_{p}^{2}$ = 0.14).⁶³ Notably, late aMCI cases exhibited greater theta ERS compared to early cases (Cohen's d = 1.155), while alpha ERD differences between Go and NoGo trials were minimal (Cohen's d = 0.070 and 0.009 for central and centroparietal regions), reflecting subtle impairments in attentional control and working memory.⁶⁵ Furthermore, ERP and machine learning-based features derived from the auditory detection task successfully differentiate cognitively impaired individuals (those with MCI and at greater risk of being diagnosed with dementia) from HCs.⁶⁶ Alpha power features during the Eriksen flanker task, which probes conflict processing by requiring rapid identification of a central target amid flanking non-target stimuli, effectively distinguished individuals with MCI from HCs (AUC of 0.77).⁶⁷ Moreover, classification models combining ERP and oscillatory features have achieved accuracies of up to 88% across binary group comparisons.⁶⁸

Working memory tasks evaluate the ability to temporarily store and update information. To evaluate working memory, researchers commonly employ memory span tasks, including N-back, delayed match-to-sample (DMS), recognition memory, and Sternberg tasks (see Supplemental Table 3 for a summary of working-memory studies). In these paradigms, patients with MCI and AD consistently exhibited reduced P3 amplitude and prolonged latency,^69–71 as well as smaller P450 amplitudes.^72,73 Longitudinal data further showed that reduced the positive-negative working memory component (PNwm) amplitudes predicted progression to dementia.⁷⁴ Meanwhile, early visuo-cortical P1 and N1 components were comparable between individuals with and without cognitive impairment, indicating the preservation of visual processing.⁷⁵ Additional ERP and oscillatory features demonstrated predictive value, with P2 and N2 latencies and beta ERS reaching AUC values of 0.93, 0.78, and 0.72, respectively; combined, these features boosted classification accuracy to 90%.⁷⁶ Oscillatory analyses revealed reduced alpha ERD⁷³ and reduced alpha/beta ERS in MCI and AD cases, with individuals with MCI exhibiting more theta desynchronization than those with AD. ^73,77,78 Functional connectivity was broadly reduced, particularly in delta and alpha bands.^71,72,79 EEG complexity measures, characterized by higher Hurst exponent and lower entropy measures, diminished fractal dimension, and reduced mutual information, have been shown to distinguish MCI and AD from controls.^80–82 Importantly, these altered connectivity and oscillatory markers not only discriminated between groups but also indicated risk of progression, with smaller alpha- and beta-band ERD in stable controls (sCON) compared to deteriorating controls (dCON) and reduced beta-band inter-trial phase synchrony in dCON relative to sCON.⁷⁵

Memory paradigms such as the DMS task, which comprises encoding, maintenance, and retrieval phases, are frequently used to assess recognition memory processes. In MCI patients, DMS tasks revealed reduced P3 amplitudes, particularly at left frontal sites,^83–85 while repetition effects at P600 remained preserved⁸⁶ and even enhanced at posterior sites,⁸⁷ suggesting partial compensation in memory retrieval processes. For visuospatial memory tasks, which rely on the integrity of attention and working memory circuits, showed reduced alpha-band connectivity in MCI under high memory load conditions, with increased network centralization in both alpha and beta bands.⁸⁸ In contrast, AD patients exhibited left-hemisphere network dropout, characterized by reduced temporo-frontal connectivity during recognition and delayed signal transmission during priming, indicating lateralized memory circuit dysfunction.⁸⁹ Reduced connectivity in the upper alpha and beta bands, as measured by synchronization likelihood, was also observed in AD patients.⁹⁰

Other ERP markers of working memory and visual short-term memory showed variable results across studies. For instance, N2 and contralateral delay activity (CDA), reflecting encoding and maintenance of multiple objects, were altered in patient groups.^37,91,92 Power spectral analyses provided further evidence of abnormalities: greater delta, theta, and alpha power were observed in both AD and MCI,⁹³ although reductions in theta power were specifically noted in AD patients⁹⁴ and in MCI individuals who later progressed to AD.⁹⁵ A consistent trend of diminishing alpha activity has been reported, with reduced alpha synchronization in MCI and diminished alpha power, particularly in the right hemisphere, in AD.^82,94 Interestingly, no alpha suppression was observed in either MCI converters or non-converters,⁹⁵ suggesting that alpha modulation may differentiate between stable and progressive impairment. Beyond memory tasks, non-memory paradigms such as backward counting revealed increased total power in frontal and central regions for AD and MCI, respectively, achieving a classification accuracy of 85.4% when spectral and complexity features were combined.⁹⁶ Furthermore, substantial beta-band coherence differences, particularly in frontal and frontoparietal regions, were observed in MCI patients, with large effect sizes (Cohen's d > 0.5), suggesting that disrupted prefrontal connectivity underlies spatial navigation difficulties.⁹⁷ In addition, the MCI group exhibited increased beta- and theta-band power compared to the SCD group,⁹⁸ further implicating frequency-domain alterations in early disease stages.

In decision-making tasks, a component associated with processing negative feedback (feedback-related negativity: FRN) was reduced in aMCI during gambling paradigms,⁹⁹ while categorization tasks elicited reduced and delayed P3 responses.¹⁰⁰ and ERP as well as time–frequency features derived from these tasks discriminated MCI from controls with an accuracy of 87.9%, even when recorded from a single EEG channel.¹⁰¹

In word repetition tasks, repeated congruent words elicited reduced P600, and incongruous words produced attenuated N400. Reduced P600 was observed in MCI and individuals progressing to MCI.^102–104 N400 attenuation occurred in non-converters across right-hemisphere sites, while converters showed no effect¹⁰³; only one study reported prolonged N400 latency in MCI.¹⁰² Visibility graph features from EEG differentiated AD and MCI converters from controls with 100% and 95.83% accuracy, respectively.¹⁰⁵ Meanwhile, EEG features derived from multimodal paradigms combining EEG, speech, and virtual reality achieved accuracy of 87% for MCI detection.¹⁰⁶ In emotional processing tasks, reduced late positive potential (LPP) in MCI, particularly in those at risk of dementia, reflected impaired reappraisal ability, with smaller amplitude differences between neutral and negative stimuli indicating difficulty attenuating negative responses.¹⁰⁷ Studies involving decision-making, language processing, and other cognitive domains are listed in Supplemental Table 4.

Discussion

This systematic review summarizes the use of EEG as a predictive biomarker for identifying different stages of the dementia spectrum, i.e., MCI and AD, and highlights the importance of cognitive paradigms over resting-state EEG. Cognitive paradigms offer more detailed insights into attention, memory, and cognitive processing. Since EEG can reveal early functional changes in the brain caused by AD, even before tissue damage or symptoms appear, this makes EEG a useful biomarker for predicting neuronal degeneration and a promising alternative or complement to current tools like MRI or PET scans. Together, these findings indicate that EEG captures disease stage–specific alterations, with MCI often characterized by early compensatory changes and AD by more extensive slowing, connectivity loss, and network breakdown. Such stage-specific differentiation between healthy aging, MCI, and early AD is especially valuable in the context of emerging disease-modifying treatments, such as monoclonal antibody therapies targeting beta-amyloid, which are most effective at the prodromal or early AD stage. By capturing functional changes before overt clinical decline, task-based EEG biomarkers provide a cost-effective and non-invasive means to support early detection, refine patient stratification, and monitor disease trajectory in clinical practice.

Prediction of MCI and AD via task-based EEG biomarkers

Several task-based EEG measures have demonstrated predictive utility in identifying individuals at risk of cognitive decline, offering insights into early neural changes that precede measurable behavioral impairments. In oddball tasks, reductions in P3 amplitude were repeatedly observed not only in clinically diagnosed MCI and AD patients but also in preclinical individuals identified as at risk on neuropsychological screening.^34–37 This reduction suggests impaired allocation of attentional resources and diminished cognitive updating, processes critical to detecting and responding to salient stimuli.¹¹² Importantly, the strong correlation between P3 latency and disease severity underscores its value as an early marker and a progression indicator.³⁶ Moreover, P3-related responses have demonstrated superior discriminative accuracy compared to conventional cognitive screening tools such as the MMSE in distinguishing AD from healthy controls,³⁸ further reinforcing the role of P3 as a sensitive early marker of cognitive decline. This suggests that attentional ERPs, particularly those related to the P3 response, can provide both predictive and diagnostic value, reflecting heightened dementia risk more sensitively than conventional cognitive screening tools. Moreover, task manipulations such as varying stimulus presentation rates further reveal how neural responses are modulated; for instance, elevated auditory N1 amplitudes in MCI patients¹⁰⁹ suggest that rapid stimuli predominantly engage the auditory cortex, whereas slower rates recruit additional frontal regions,^113,114 highlighting the sensitivity of ERP markers to both cognitive state and task context.

In language-based paradigms, particularly word repetition tasks, reductions in P600 and attenuated N400 responses have been associated with later conversion to MCI or AD, whereas individuals who remained stable retained N400 repetition effects.^102–104 Similarly, decreases in alpha and beta synchrony differentiated deteriorating from stable control groups,⁷⁵ and reductions in theta power were specifically observed in MCI patients who subsequently progressed to AD.⁹⁵ Together, these findings suggest a consistent electrophysiological signature of disrupted semantic integration and memory retrieval processes across studies. Consequently, these measures may serve as sensitive neurophysiological indicators for identifying individuals at elevated risk of cognitive decline before the onset of overt clinical symptoms.

A consistent finding across working-memory paradigms is the emergence of compensatory frontal activity. Studies have reported enhanced left frontal ERP responses during non-target stimuli in MCI, alongside diminished responses to target stimuli, reflecting a compensatory mechanism to counter their impaired ability in inhibitory control.^84,85 Importantly, these frontal ERP alterations have been detected up to five years before clinical diagnosis,⁸⁵ supporting their value of frontal memory-related neuromarkers for individualized monitoring of disease progression. Furthermore, as individuals aged, particularly with the onset of MCI, the ability to suppress distractors became more challenging. Correspondingly, reductions in P3 (memory recall), N400 and P600 (semantic memory) and alpha/beta activity (attention and inhibitory control) are frequently observed among those at increased likelihood of cognitive impairment. Additionally, variations in theta oscillations observed across studies may reflect differences in cognitive control strategies rather than uniform patterns of neural dysfunction. Therefore, despite differences in task design and analytical methods, the consistent decreases observed in these task-based EEG measures reinforce their reliability as biomarkers for detecting early cognitive impairment.

Classification of MCI and AD via task-based EEG biomarkers

Several task-based EEG features have demonstrated strong potential for classifying between groups. Network-level measures of neural synchronization and cross-frequency coupling metrics show improved classification accuracies by capturing disrupted hierarchical communication across frequency bands, even for more complex three-class classifications.^{49,52,115,116} This evidence reinforces the clinical value of network-based synchronization measures as robust discriminators along the AD continuum. As individuals age, EEG patterns typically show a decrease in alpha power and an increase in slower brain wave frequencies like delta and theta. These slow waves differ significantly from those typically associated with AD. Unlike AD-related slow activity, these nonpathological slow waves are not linked to a substantial asymmetry of the alpha rhythm, tend to diminish with mental activity or eye opening, and occur only sporadically as single waves or pairs rather than in continuous rhythmic trains.¹¹⁷ Reductions in slow-wave activity may indicate cognitive improvement in patients responding to cholinesterase inhibitor therapy further supports their clinical relevance.¹¹⁸ Desynchronization in the alpha band remarkably increased (less negative) in centroparietal regions and has been associated with attention, indicating reduced attentiveness and correlating with poor performance.^119,120 Moreover, variations in theta power have also demonstrated discriminative potential in distinguishing cognitively healthy individuals from those with impairment,⁶³ while abnormalities in theta ERS further suggest disruptions in attentional control networks that contribute to cognitive dysfunction.¹²¹ However, results contrary to the findings in MCI patients have emerged, demonstrating an absence of beta ERD in individuals with AD during the retrieval phase of short-term memory processing.⁷⁸ This may suggest that the AD pathology disrupts the normal desynchronization pattern of neural activity in this frequency band. These effects, known as a general “slowing” of EEG activity.¹²² Given that dementia is typified by a “disconnection syndrome,” marked by diminished brain networks,^123,124 individuals consistently manifest impaired functional connectivity, reduced complexity and perturbation of synchrony as AD progresses. Similarly, selective-attention ERP and spectral features have demonstrated high classification performances between cognitively impaired and healthy individuals.^67,101 These findings underscore the diagnostic utility of EEG markers that capture deficits in attention, memory, and inhibitory control.

Across studies, prolonged N2 latency and reduced N2 amplitude indicate slowed cognitive processing and deficits in early attentional control among individuals with MCI,^59,60 while reduction in P3 amplitude and delayed latency further reflect impairments in attentional allocation and working memory updating.^34,37,62 Beyond amplitude and latency, trial-to-trial variability in P3 responses has been reported as an additional discriminative feature, reflecting the instability of attentional resource allocation in MCI and AD.^57,91 The robustness of these findings supports the relevance of P3 as a potential clinical biomarker. Meanwhile, elevated P50 amplitude in MCI, which correlated negatively with left frontal metabolic abnormalities but not with MMSE performance, suggests that abnormal sensory gating may capture localized neuronal dysfunction rather than global cognitive status, limiting its reliability as a general clinical marker.^36,46 In contrast, the increased LPC amplitude observed in MCI likely reflects compensatory perceptual processing in response to early sensory deficits,⁴⁵ underscoring the potential of ERP components to differentiate localized dysfunction from adaptive mechanisms.

Finally, differentiation between stable and progressive MCI remains underexplored, though it is critical for clinical decision-making. Evidence suggests that altered oscillatory activity (e.g., reduced theta power,⁹⁵ smaller alpha/beta ERD⁷⁵) and P3 amplitude differences^62,85 may serve as indicators of progression risk, yet few studies stratify MCI patients by conversion status. Clarifying these criteria is necessary for EEG biomarkers to provide meaningful clinical guidance, particularly in identifying which patients are most likely to benefit from early interventions or monitoring. Nevertheless, classification outcomes vary widely across studies due to differences in task design, including stimulus modality, stimulus duration, and number of trials. The number of extracted features also varies widely, from large sets to only a few selected features, making comparisons between studies difficult. Preprocessing strategies also vary considerably, encompassing artifact rejection criteria, filtering approaches, and baseline correction procedures, each of which can affect the resulting EEG signals. These methodological differences highlight the adaptability of EEG but also underscore the pressing need for standardized protocols and consensus on core feature sets to ensure reproducibility and facilitate translation into clinical practice.

Monitoring disease progression via task-based EEG biomarkers

In oddball paradigms, P3 latency correlated strongly with AD severity and increased longitudinally over 18 months in probable AD patients, whereas N2 latency exhibited a weaker but still significant correlation.^36,100 Findings from working memory paradigms further support this progression pattern, as reductions in PNwm amplitude have been linked to the transition from MCI to dementia.⁷⁴ Connectivity-based studies complement these results, demonstrating greater network centralization in late compared to early aMCI.⁶⁵ Moreover, oscillatory analyses revealed progressive reductions in theta and alpha power in both MCI and AD,^94,95 with diminished alpha synchronization particularly evident in the right hemisphere of AD patients.^82,94 Additionally, differences in theta ERS have shown promise in distinguishing between early and late aMCI groups.⁶⁵ These findings indicate that differences in theta and alpha oscillations between early and late MCI may serve as neural progression markers capable of tracking disease trajectory even when behavioral performance remains similar.

The N400 and P600 (or LPC) components reflect semantic processing and episodic memory, and their modulation by repeated (old) versus novel words, known as the repetition effect, provides insight into neural memory function .¹⁰³ In individuals with MCI, both N400 and P600 repetition effects were reduced, with the most pronounced deficits observed in those who later progressed to AD.^102–104 Notably, the P600 effect was absent in MCI converters within three years, and by the second year, the N400 repetition effect had disappeared entirely in converters, while non-converters maintained a pattern similar to that observed at baseline .¹⁰³ These findings indicate that diminished or absent N400 and P600 repetition effects may serve as sensitive early markers of disease progression, offering a potential tool for identifying individuals at high risk for dementia.

Connectivity disruptions provided additional markers of progression. Left-hemisphere network dropout during recognition and delayed signal transmission during priming tasks were reported in AD, reflecting lateralized memory circuit dysfunction,⁸⁹ while reduced synchronization in upper alpha and beta bands characterized advanced disease stages.⁹⁰ Complexity measures, including higher Hurst exponent, lower entropy, and diminished fractal dimension, further distinguished controls from cognitively impaired groups and tracked worsening impairment.^80–82 The observed deterioration in AD may be attributed to the presence of plaques and cell death, leading to a reduction in connectivity between cortical neurons. This reduction could result in increased regularity of brain signals, ultimately impacting communication efficiency throughout the brain. These findings support the concept of a disconnection syndrome emerging in MCI, potentially progressing to AD or other forms of dementia. Key findings from studies on task-based EEG are summarized in Table 4.

Table 4.

Main findings from studies on task-based EEG (own elaboration based on reports from^36,^40–43^{,45,50,57,58,61–65,69–74,100–105,107,111,115,126}).

Cognitive domain	Task	Experimental conditions	EEG alterations in MCI/AD
Attention and target detection	Oddball	Utilizes low-probability targets (10–30%) embedded among frequent standards (70–90%). Task demands sustained attention that can be challenging for cognitively impaired individuals High signal-to-noise variant (FPVS) using 6 Hz standard and 1 Hz oddball stimuli, enables passive assessment of recognition memory, though rapid flickering may affect photosensitive participants.	Significant reduction in P3 amplitude and prolonged P3 latency, reflecting impaired attentional allocation and target detection. Frontal delta oscillations and parietal P3 responses (maximal at Pz) are attenuated in MCI/AD.
Inhibitory control and executive function	Go/NoGo, Stroop, Flanker	Electrode placement influences observed oscillatory patterns.	Reduced N2 and P3 amplitudes and delayed latencies, indicating deficits in response inhibition. mdaMCI subtype exhibits increased P3a responses, suggesting compensatory activation Decreased theta synchronization and alpha desynchronization, with attenuated alpha power during response execution in parietal and central regions
Working Memory	N-back, DMS, Sternberg	Variability in task load influences working memory demands and the amplitude–latency characteristics of task-evoked ERP components, particularly the P3	P3 and P450 amplitudes are consistently reduced in MCI and AD, reflecting deficits in attention and memory updating Reduced PNwm amplitude serves as a prognostic marker of conversion from MCI to dementia Oscillatory analyses reveal weaker delta connectivity in AD and increased theta desynchronization with reduced alpha ERD in MCI
Decision-making, language, and emotional processing	Semantic categorization, repetition priming, emotional reappraisal	-	Decreased P3 amplitude and delayed latency during categorization reflect attentional and processing inefficiencies Diminished LPP responses during emotional reappraisal indicate impairments in emotional regulation and cognitive flexibility among individuals with MCI Repetition-related P600 and N400 effects are attenuated in MCI, particularly in converters Visibility graph features achieve high classification accuracy for AD, while multimodal ERP and graph-theoretic EEG features show strong potential for distinguishing MCI and AD from healthy controls

AD: Alzheimer's disease; DMS: delayed match-to-sample; EEG: electroencephalography; ERD: event-related desynchronization; FPVS: fast periodic visual stimulation; LPP: late positive potential; MCI: mild cognitive impairment; mdaMCI: multi-domain amnestic MCI.

The scope of this study pertains to manuscripts derived from the Open Access Policy, ensuring accessibility and transparency in disseminating findings. The insights summarized here, drawn from scientifically acknowledged and accessible protocols, support the future development of innovative approaches to MCI screening, consistent with other publications.¹²⁵ Methodological and practical limitations should be acknowledged. Potential omissions of relevant studies due to database or search-term constraints, variability in diagnostic criteria, and limited reporting of effect sizes reduce cross-study comparability and the robustness of clinical interpretations. Moreover, translation of EEG findings into clinical practice remains constrained by heterogeneity in task design, stimulus parameters, preprocessing pipelines, and analytical methods. The absence of standardized protocols and normative datasets makes it difficult to compare results across laboratories and to validate cut-off values that could guide clinical decision-making. Moreover, there is a notable gap between the controlled conditions of laboratory-based cognitive paradigms and the realities of clinical or hospital environments, where factors such as time constraints, patient fatigue, and comorbidities may influence feasibility and data quality. To advance clinical translation, future studies should prioritize protocol harmonization, replication in large multi-center cohorts, and integration with existing diagnostic frameworks. Such steps are critical for establishing reliable, reproducible EEG-based predictors that can complement current neuropsychological and imaging assessments in routine clinical settings.

Conclusions

This review highlights the evidence supporting task-based EEG as biomarkers for early detection and monitoring of cognitive decline across the MCI-AD continuum. Characteristic EEG alterations, most notably decreased P3 and N2 amplitudes, prolonged latencies, reduced alpha and beta activity, and disrupted connectivity, reflect impairments in attention, memory, and cortical inhibition that arise prior to overt clinical symptoms. These converging findings indicate that task-evoked EEG measures capture both early neurophysiological dysfunction and compensatory processes, offering potential utility in disease staging and prognosis. Clinically, task-based EEG provides a low-cost, noninvasive complement to traditional cognitive and neuroimaging assessments, supporting its integration into multimodal diagnostic frameworks.

Supplemental Material

sj-docx-1-alz-10.1177_13872877251413700 - Supplemental material for Biomarkers for Alzheimer's disease and mild cognitive impairment: Recent advances in task-based EEG

Supplemental material, sj-docx-1-alz-10.1177_13872877251413700 for Biomarkers for Alzheimer's disease and mild cognitive impairment: Recent advances in task-based EEG by Kornkanok Tripanpitak, Alexandra Wolf, Maryna Kapitonova, Satoshi Umeda, Tonio Ball and Mihoko Otake-Matsuura in Journal of Alzheimer's Disease

Footnotes

Acknowledgements

The authors express gratitude to the RIKEN Center for Advanced Intelligence Project (AIP) for supporting this research.

ORCID iDs

Kornkanok Tripanpitak

Alexandra Wolf

Maryna Kapitonova

Satoshi Umeda

Mihoko Otake-Matsuura

Ethical considerations

Not applicable

Consent to participate

Not applicable

Consent for publication

Not applicable

Author contribution(s)

Kornkanok Tripanpitak: Conceptualization; Methodology; Writing – original draft; Writing – review & editing.

Alexandra Wolf: Conceptualization; Writing – review & editing.

Maryna Kapitonova: Writing – review & editing.

Satoshi Umeda: Writing – review & editing.

Tonio Ball: Writing – review & editing.

Mihoko Otake-Matsuura: Conceptualization; Funding acquisition; Writing – review & editing.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was partly supported by the Japan Science and Technology Agency AIP Trilateral AI Research (Grant Numbers: JPMJCR20G1) to MO-M and KT. MO-M was supported by JSPS KAKENHI (Grant Numbers: JP22H04872 and JP22H00544) and the Japan Science and Technology Agency (Grant Numbers: JPMJPF2101). AW was supported by the Special Postdoctoral Research (SPDR) grant under the RIKEN SPDR Fellowship. MO-M and SU was supported by the Japan Science and Technology Agency (Grant Numbers: JPMJMS2237). TB and MK were supported by the AI-Cog (DFG, Germany project BA 4695/4-1).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability statement

The data supporting the findings of this study are available within the article and/or its supplemental material.

Supplemental material

Supplemental material for this article is available online.

References

World Health Organization. Global action plan on the public health response to dementia 2017–2025. Geneva: World Health Organization, 2017, https://apps.who.int/iris/handle/10665/259615 (accessed 7 September 2022).

Woodford

George

. Cognitive assessment in the elderly: a review of clinical methods. QJM 2007; 100: 469–484.

Petersen

Smith

Waring

, et al. Mild cognitive impairment: clinical characterization and outcome. Arch Neurol 1999; 56: 303.

Perry

. Attention and executive deficits in Alzheimer’s disease: a critical review. Brain 1999; 122: 383–404.

Levinoff

Saumier

Chertkow

. Focused attention deficits in patients with Alzheimer’s disease and mild cognitive impairment. Brain Cogn 2005; 57: 127–130.

Baudic

Barba

Thibaudet

, et al. Executive function deficits in early Alzheimer’s disease and their relations with episodic memory. Arch Clin Neuropsychol 2006; 21: 15–21.

Guarino

Favieri

Boncompagni

, et al. Executive functions in Alzheimer disease: a systematic review. Front Aging Neurosci 2019; 10: 437.

Binetti

Magni

Padovani

, et al. Executive dysfunction in early Alzheimer’s disease. J Neurol Neurosurg Psychiatry 1996; 60: 91–93.

Reid

Broe

Creasey

, et al. Age at onset and pattern of neuropsychological impairment in mild early-stage Alzheimer disease: a study of a community-based population. Arch Neurol 1996; 53: 1056–1061.

10.

Folstein

McHugh

. Mini-mental state. J Psychiatr Res 1975; 12: 189–198.

11.

Shiroky

Schipper

Bergman

, et al.

Can you have dementia with an MMSE score of 30?

Am J Alzheimers Dis Other Demen 2007; 22: 406–415.

12.

Carnero-Pardo

. Should the mini-mental state examination be retired? Neurología 2014; 29: 473–481.

13.

Ciesielska

Sokołowski

Mazur

, et al. Is the Montreal Cognitive Assessment (MoCA) test better suited than the Mini-Mental State Examination (MMSE) in mild cognitive impairment (MCI) detection among people aged over 60? Meta-analysis. Psychiatr Pol 2016; 50: 1039–1052.

14.

Pinto

TCC

Machado

Bulgacov

, et al.

Is the Montreal Cognitive Assessment (MoCA) screening superior to the Mini-Mental State Examination (MMSE) in the detection of mild cognitive impairment (MCI) and Alzheimer’s disease (AD) in the elderly?

Int Psychogeriatr 2019; 31: 491–504.

15.

Mathuranath

Nestor

Berrios

, et al. A brief cognitive test battery to differentiate Alzheimer’s disease and frontotemporal dementia. Neurology 2000; 55: 1613–1620.

16.

Borson

Scanlan

Chen

, et al. The Mini-Cog as a screen for dementia: validation in a population-based sample: MINI-COG IN MOVIES. J Am Geriatr Soc 2003; 51: 1451–1454.

17.

Amodio

Wenin

Del Piccolo

, et al. Variability of Trail Making Test, Symbol Digit Test and Line Trait Test in normal people. A normative study taking into account age-dependent decline and sociobiological variables. Aging Clin Exp Res 2002; 14: 117–131.

18.

Shulman

. Clock-drawing: is it the ideal cognitive screening test? Int J Geriatr Psychiatry 2000; 15: 548–561.

19.

Berry

DTR

Allen

Schmitt

. Rey-Osterrieth complex figure: psychometric characteristics in a geriatric sample. Clin Neuropsychol 1991; 5: 143–153.

20.

World Health Organization. Dementia, https://www.who.int/news-room/fact-sheets/detail/dementia (2023, accessed 7 September 2022).

21.

Otake-Matsuura

Tokunaga

Watanabe

, et al. Cognitive intervention through Photo-Integrated Conversation Moderated by Robots (PICMOR) Program: a randomized controlled trial. Front Robot AI 2021; 8: 633076.

22.

Wolf

Tripanpitak

Umeda

, et al. Eye-tracking paradigms for the assessment of mild cognitive impairment: a systematic review. Front Psychol 2023; 14: 1197567.

23.

Leuzy

Ashton

Mattsson-Carlgren

, et al. 2020 Update on the clinical validity of cerebrospinal fluid amyloid, tau, and phospho-tau as biomarkers for Alzheimer’s disease in the context of a structured 5-phase development framework. Eur J Nucl Med Mol Imaging 2021; 48: 2121–2139.

24.

Sonnen

Montine

Quinn

, et al. Biomarkers for cognitive impairment and dementia in elderly people. Lancet Neurol 2008; 7: 704–714.

25.

Fagan

Perrin

. Upcoming candidate cerebrospinal fluid biomarkers of Alzheimer’s disease. Biomark Med 2012; 6: 455–476.

26.

Albert

DeKosky

Dickson

, et al. The diagnosis of mild cognitive impairment due to Alzheimer’s disease: recommendations from the National Institute on Aging- Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 2012; 7: 270–279.

27.

Jack

Jr. Bennett

Blennow

, et al. NIA-AA Research Framework: toward a biological definition of Alzheimer’s disease. Alzheimers Dement 2018; 14: 535–562.

28.

Selkoe

Hardy

. The amyloid hypothesis of Alzheimer’s disease at 25 years. EMBO Mol Med 2016; 8: 595–608.

29.

Wirth

Madison

Rabinovici

, et al. Alzheimer’s disease neurodegenerative biomarkers are associated with decreased cognitive function but not -amyloid in cognitively normal older individuals. J Neurosci 2013; 33: 5553–5563.

30.

Page

McKenzie

Bossuyt

, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. J Clin Epidemiol 2021; 134: 178–189.

31.

Horsley

Dingwall

Sampson

. Checking reference lists to find additional studies for systematic reviews. Cochrane Database Syst Rev 2011; 2011: MR000026.

32.

Haddaway

Page

Pritchard

, et al. PRISMA2020: an R package and Shiny app for producing PRISMA 2020-compliant flow diagrams, with interactivity for optimised digital transparency and Open Synthesis. Campbell Syst Rev 2022; 18: e1230.

33.

Donchin

Coles

MGH

. Is the P300 component a manifestation of context updating? Behav Brain Sci 1988; 11: 357–374.

34.

Parra

Ascencio

Urquina

, et al. P300 and neuropsychological assessment in mild cognitive impairment and Alzheimer dementia. Front Neur 2012; 3: 172.

35.

Pedroso

Fraga

Nascimento

CMC

, et al. Apolipoprotein E ɛ4 allele impairs cortical activity in healthy aging and Alzheimer’s disease. Behav Brain Res 2022; 420: 113700.

36.

Fruehwirt

Dorffner

Roberts

, et al. Associations of event-related brain potentials and Alzheimer’s disease severity: a longitudinal study. Prog Neuropsychopharmacol Biol Psychiatry 2019; 92: 31–38.

37.

Newsome

Pun

Smith

, et al. Neural correlates of cognitive decline in older adults at-risk for developing MCI: evidence from the CDA and P300. Cogn Neurosci 2013; 4: 152–162.

38.

Doan

DNT

Choi

, et al. Predicting dementia with prefrontal electroencephalography and event-related potential. Front Aging Neurosci 2021; 13: 659817.

39.

Yener

Güntekin

Örken

, et al. Auditory delta event-related oscillatory responses are decreased in Alzheimer’s disease. Behav Neurol 2012; 25: 3–11.

40.

Yener

Kurt

Emek-Savaş

, et al. Reduced visual event-related delta oscillatory responses in amnestic mild cognitive impairment. J Alzheimers Dis 2013; 37: 759–767.

41.

Yener

Emek-Savaş

Güntekin

, et al. The visual cognitive network, but not the visual sensory network, is affected in amnestic mild cognitive impairment: a study of brain oscillatory responses. Brain Res 2014; 1585: 141–149.

42.

Yener

Emek-Sava

Ada

, et al. Frontal delta event-related oscillations relate to frontal volume in mild cognitive impairment and healthy controls. Int J Psychophysiol 2016; 103: 110–117.

43.

Güntekin

Emek-Savaş

Kurt

, et al. Beta oscillatory responses in healthy subjects and subjects with mild cognitive impairment. Neuroimage Clin 2013; 3: 39–46.

44.

Başar

Emek-Savaş

Güntekin

, et al. Delay of cognitive gamma responses in Alzheimer’s disease. Neuroimage Clin 2016; 11: 106–115.

45.

Invitto

Piraino

Ciccarese

, et al. Potential role of OERP as early marker of mild cognitive impairment. Front Aging Neurosci 2018; 10: 272.

46.

Shao

Wang

, et al. Correlation of auditory event-related potentials and magnetic resonance spectroscopy measures in mild cognitive impairment. Brain Res 2010; 1346: 204–212.

47.

Güntekin

Saatçi

Yener

. Decrease of evoked delta, theta and alpha coherences in Alzheimer patients during a visual oddball paradigm. Brain Res 2008; 1235: 109–116.

48.

Fide

Yerlikaya

Güntekin

, et al. Coherence in event-related EEG oscillations in patients with Alzheimer’s disease dementia and amnestic mild cognitive impairment. Cogn Neurodyn 2022; 17: 1621–1635.

49.

Sedghizadeh

Aghajan

Vahabi

, et al. Network synchronization deficits caused by dementia and Alzheimer’s disease serve as topographical biomarkers: a pilot study. Brain Struct Funct 2022; 227: 2957–2969.

50.

Stothart

Smith

Milton

, et al. A passive and objective measure of recognition memory in Alzheimer’s disease using Fastball memory assessment. Brain 2021; 144: 2812–2825.

51.

Botvinick

Carter

Braver

, et al. Conflict monitoring and cognitive control. Psychol Rev 2001; 108: 624–652.

52.

Nigbur

Ivanova

Stürmer

. Theta power as a marker for cognitive interference. Clin Neurophysiol 2011; 122: 2185–2194.

53.

Harper

Malone

Bernat

. Theta and delta band activity explain N2 and P3 ERP component activity in a go/no-go task. Clin Neurophysiol 2014; 125: 124–132.

54.

Falkenstein

Hoormann

Hohnsbein

. ERP Components in Go/Nogo tasks and their relation to inhibition. Acta Psychol 1999; 101: 267–291.

55.

Fallgatter

Strik

. The NoGo-anteriorization as a neurophysiological standard-index for cognitive response control. Int J Psychophysiol 1999; 32: 233–238.

56.

Cheng

C-H

Tsai

H-Y

Cheng

H-N

. The effect of age on N2 and P3 components: a meta-analysis of Go/Nogo tasks. Brain Cogn 2019; 135: 103574.

57.

Lockwood

Vaughn

Duffy

. Attentional ERPs distinguish aging and early Alzheimer’s dementia. Neurobiol Aging 2018; 70: 51–58.

58.

Cid-Fernández

Lindín

Díaz

. Effects of amnestic mild cognitive impairment on N2 and P3 Go/NoGo ERP components. J Alzheimers Dis 2013; 38: 295–306.

59.

Cid-Fernández

Lindín

Díaz

. Neurocognitive and behavioral indexes for identifying the amnestic subtypes of mild cognitive impairment. J Alzheimers Dis 2017; 60: 633–649.

60.

Chiang

H-S

Spence

Kraut

, et al. Age effects on event-related potentials in individuals with amnestic Mild Cognitive Impairment during semantic categorization Go/NoGo tasks. Neurosci Lett 2018; 670: 19–21.

61.

Cid-Fernández

. The importance of age in the search for ERP biomarkers of aMCI. Biol Psychol 2019; 142: 108–115.

62.

Correa-Jaraba

Lindín

Díaz

. Increased amplitude of the P3a ERP component as a neurocognitive marker for differentiating amnestic subtypes of mild cognitive impairment. Front Aging Neurosci 2018; 10: 19.

63.

Nguyen

Lydon

Shende

, et al. Disrupted value-directed strategic processing in individuals with mild cognitive impairment: behavioral and neural correlates. Geriatrics 2022; 7: 56.

64.

Nguyen

Mudar

Chiang

H-S

, et al. Theta and alpha alterations in amnestic mild cognitive impairment in semantic Go/NoGo Tasks. Front Aging Neurosci 2017; 9: 160.

65.

Lydon

Nguyen

Shende

, et al. EEG Theta and alpha oscillations in early versus late mild cognitive impairment during a semantic Go/NoGo task. Behav Brain Res 2022; 416: 113539.

66.

Molcho

Maimon

Regev-Plotnik

, et al. Single-channel EEG features reveal an association with cognitive decline in seniors performing auditory cognitive assessment. Front Aging Neurosci 2022; 14: 773692.

67.

Chen

, et al. The weakened relationship between prestimulus alpha oscillations and response time in older adults with mild cognitive impairment. Front Hum Neurosci 2020; 14: 48.

68.

Sharma

Kolekar

Jha

, et al. EEG And cognitive biomarkers based mild cognitive impairment diagnosis. IRBM 2019; 40: 113–121.

69.

L-H

Chen

Gao

L-J

, et al. The effect of apolipoprotein E ɛ4 (APOE ɛ4) on visuospatial working memory in healthy elderly and amnestic mild cognitive impairment patients: an event-related potentials study. Front Aging Neurosci 2017; 9: 145.

70.

Chen

Gao

, et al. Cognitive reserve modulates attention processes in healthy elderly and amnestic mild cognitive impairment: an event-related potential study. Clin Neurophysiol 2018; 129: 198–207.

71.

Zhang

Gesang

, et al. The changes of central auditory processing function and its electroencephalogram in patients with mild cognitive impairment and early Alzheimer’s disease. Heliyon 2023; 9: e15641.

72.

Liddell

Paul

Arns

, et al. Rates of decline distinguish Alzheimer’s disease and mild cognitive impairment relative to normal aging: integrating cognition and brain function. J Integr Neurosci 2007; 06: 141–174.

73.

Fraga

Mamani

Johns

, et al. Early diagnosis of mild cognitive impairment and Alzheimer’s with event-related potentials and event-related desynchronization in N-back working memory tasks. Comput Methods Programs Biomed 2018; 164: 1–13.

74.

Missonnier

Gold

Fazio-Costa

, et al. Early event-related potential changes during working memory activation predict rapid decline in mild cognitive impairment. J Gerontol A Biol Sci Med Sci 2005; 60: 660–666.

75.

Deiber

M-P

Meziane

Hasler

, et al. Attention and working memory-related EEG markers of subtle cognitive deterioration in healthy elderly individuals. J Alzheimers Dis 2015; 47: 335–349.

76.

Missonnier

Deiber

M-P

Gold

, et al. Working memory load–related electroencephalographic parameters can differentiate progressive from stable mild cognitive impairment. Neuroscience 2007; 150: 346–356.

77.

Fodor

Sirály

Horváth

, et al. Decreased event-related beta synchronization during memory maintenance marks early cognitive decline in mild cognitive impairment. J Alzheimers Dis 2018; 63: 489–502.

78.

Karrasch

Laine

Rinne

, et al. Brain oscillatory responses to an auditory-verbal working memory task in mild cognitive impairment and Alzheimer’s disease. Int J Psychophysiol 2006; 59: 168–178.

79.

Hogan

Swanwick

GRJ

Kaiser

, et al. Memory-related EEG power and coherence reductions in mild Alzheimer’s disease. Int J Psychophysiol 2003; 49: 147–163.

80.

Nimmy John

Puthankattil

Menon

. Analysis of long range dependence in the EEG signals of Alzheimer patients. Cogn Neurodyn 2018; 12: 183–199.

81.

Nimmy John

Subha Dharmapalan

Ramshekhar Menon

. Exploration of time–frequency reassignment and homologous inter-hemispheric asymmetry analysis of MCI–AD brain activity. BMC Neurosci 2019; 20: 38.

82.

Mazzon

De Dea

Cattaruzza

, et al. Memorization test and resting state EEG components in mild and subjective cognitive impairment. Curr Alzheimers Res 2018; 15: 809–819.

83.

B-Y

Tang

H-D

Chen

S-D

. Retrieval deficiency in brain activity of working memory in amnesic mild cognitive impairment patients: a brain event-related potentials study. Front Aging Neurosci 2016; 8: 54.

84.

Broster

Jicha

, et al. A cognitive electrophysiological signature differentiates amnestic mild cognitive impairment from normal aging. Alzheimers Res Ther 2017; 9: 3.

85.

Jiang

Schmitt

, et al. Memory-related frontal brainwaves predict transition to mild cognitive impairment in healthy older individuals five years before diagnosis. J Alzheimers Dis 2021; 79: 531–541.

86.

Broster

Jenkins

Holmes

, et al. Electrophysiological repetition effects in persons with mild cognitive impairment depend upon working memory demand. Neuropsychologia 2018; 117: 13–25.

87.

Broster

Wagner

, et al. Spared behavioral repetition effects in Alzheimer’s disease linked to an altered neural mechanism at posterior cortex. J Clin Exp Neuropsychol 2018; 40: 761–776.

88.

Fodor

Horváth

Hidasi

, et al. EEG Alpha and beta band functional connectivity and network structure mark hub overload in mild cognitive impairment during memory maintenance. Front Aging Neurosci 2021; 13: 680200.

89.

Tyrer

Gilbert

Adams

, et al. Lateralized memory circuit dropout in Alzheimer’s disease patients. Brain Commun 2020; 2: fcaa212.

90.

Pijnenburg

v d Made

van Cappellen van Walsum

, et al. EEG Synchronization likelihood in mild cognitive impairment and Alzheimer’s disease during a working memory task. Clin Neurophysiol 2004; 115: 1332–1339.

91.

Tautvydaitė

Adam-Darqué

Andryszak

, et al. Deficient novelty detection and encoding in early Alzheimer’s disease: an ERP study. Brain Topogr 2022; 35: 667–679.

92.

Bagattini

Mazza

Panizza

, et al.

Neural dynamics of multiple object processing in mild cognitive impairment and Alzheimer’s disease: future early diagnostic biomarkers?

J Alzheimers Dis 2017; 59: 643–654.

93.

Xun

Wei

, et al.

Is the electroencephalogram power spectrum valuable for diagnosis of the elderly with cognitive impairment?

Int J Gerontol 2015; 9: 196–200.

94.

Güntekin

Hanoğlu

Aktürk

, et al. Impairment in recognition of emotional facial expressions in Alzheimer’s disease is represented by EEG theta and alpha responses. Psychophysiology 2019; 56: e13434.

95.

Mazaheri

Segaert

Olichney

, et al. EEG Oscillations during word processing predict MCI conversion to Alzheimer’s disease. Neuroimage Clin 2018; 17: 188–197.

96.

McBride

Zhao

Munro

, et al. Spectral and complexity analysis of scalp EEG characteristics for mild cognitive impairment and early Alzheimer’s disease. Comput Methods Programs Biomed 2014; 114: 153–163.

97.

Plaza-Rosales

Brunetti

Montefusco-Siegmund

, et al. Visual-spatial processing impairment in the occipital-frontal connectivity network at early stages of Alzheimer’s disease. Front Aging Neurosci 2023; 15: 1097577.

98.

Iliadou

Paliokas

Zygouris

, et al. A comparison of traditional and serious game-based digital markers of cognition in older adults with mild cognitive impairment and healthy controls. J Alzheimers Dis 2021; 79: 1747–1759.

99.

Abe

Onoda

Takamura

, et al. Altered feedback-related negativity in mild cognitive impairment. Brain Sci 2023; 13: 203.

100.

Karimi

Marefat

Khanbagi

, et al. Temporal dynamics of animacy categorization in the brain of patients with mild cognitive impairment. PLoS One 2022; 17: e0264058.

101.

Khatun

Morshed

Bidelman

. A single-channel EEG-based approach to detect mild cognitive impairment via speech-evoked brain responses. IEEE Trans Neural Syst Rehabil Eng 2019; 27: 1063–1070.

102.

Olichney

. Abnormal verbal event related potentials in mild cognitive impairment and incipient Alzheimer’s disease. J Neurol Neurosurg Psychiatry 2002; 73: 377–384.

103.

Olichney

Taylor

Gatherwright

, et al. Patients with MCI and N400 or P600 abnormalities are at very high risk for conversion to dementia. Neurology 2008; 70: 1763–1770.

104.

Xia

Mazaheri

Segaert

, et al. Event-related potential and EEG oscillatory predictors of verbal memory in mild cognitive impairment. Brain Commun 2020; 2: fcaa213.

105.

Zhang

Xia

Liu

, et al. Machine learning on visibility graph features discriminates the cognitive event-related potentials of patients with early Alzheimer’s disease from healthy aging. Brain Sci 2023; 13: 770.

106.

Xue

, et al. Screening for mild cognitive impairment with speech interaction based on virtual reality and wearable devices. Brain Sci 2023; 13: 1222.

107.

Xiao

Liu

, et al. Electrophysiological evidence of impaired cognitive reappraisal in amnestic mild cognitive impairment: an event-related potential study. Behav Brain Res 2022; 427: 113800.

108.

Song

Xie

, et al. The auditory P50 gating in mild cognitive impairment: a case-control study. Am J Alzheimers Dis Other Demen 2022; 37: 153331752110689.

109.

Kamal

Morrison

Campbell

, et al. Event-related potential measures of the passive processing of rapidly and slowly presented auditory stimuli in MCI. Front Aging Neurosci 2021; 13: 659618.

110.

Eyamu

Kim

W-S

Kim

, et al. Prefrontal event-related potential markers in association with mild cognitive impairment. Front Aging Neurosci 2023; 15: 1273008.

111.

Kurt

Emek-Savaş

Batum

, et al. Patients with mild cognitive impairment display reduced auditory event-related delta oscillatory responses. Behav Neurol 2014; 2014: 268967.

112.

Van Deursen

Vuurman

EFPM

Smits

, et al. Response speed, contingent negative variation and P300 in Alzheimer’s disease and MCI. Brain Cogn 2009; 69: 592–599.

113.

Sams

Hari

Rif

, et al. The human auditory sensory memory trace persists about 10 s: neuromagnetic evidence. J Cogn Neurosci 1993; 5: 363–370.

114.

Mcevoy

Levänen

Loveless

. Temporal characteristics of auditory sensory memory: neuromagnetic evidence. Psychophysiology 1997; 34: 308–316.

115.

Goodman

Kumar

Zomorrodi

, et al. Theta-gamma coupling and working memory in Alzheimer’s dementia and mild cognitive impairment. Front Aging Neurosci 2018; 10: 101.

116.

Safi

SMM

. Early detection of Alzheimer’s disease from EEG signals using Hjorth parameters. Biomed Signal Process Control 2021; 65: 102338.

117.

Rodriguez

Arnaldi

Picco

. Brain functional network in Alzheimer’s disease: diagnostic markers for diagnosis and monitoring. Int J Alzheimers Dis 2011; 2011: 481903.

118.

Babiloni

Del Percio

Bordet

, et al. Effects of acetylcholinesterase inhibitors and memantine on resting-state electroencephalographic rhythms in Alzheimer’s disease patients. Clin Neurophysiol 2013; 124: 837–850.

119.

Pfurtscheller

. Event-related synchronization (ERS): an electrophysiological correlate of cortical areas at rest. Electroencephalogr Clin Neurophysiol 1992; 83: 62–69.

120.

Klimesch

Schimke

Pfurtscheller

. Alpha frequency, cognitive load and memory performance. Brain Topogr 1993; 5: 241–251.

121.

Cavanagh

Frank

. Frontal theta as a mechanism for cognitive control. Trends Cogn Sci 2014; 18: 414–421.

122.

Klimesch

. EEG Alpha and theta oscillations reflect cognitive and memory performance: a review and analysis. Brain Res Rev 1999; 29: 169–195.

123.

Dauwels

Vialatte

Cichocki

. Diagnosis of Alzheimer’s disease from EEG signals: where are we standing? Curr Alzheimers Res 2010; 7: 487–505.

124.

Vecchio

Miraglia

Piludu

, et al. Small world” architecture in brain connectivity and hippocampal volume in Alzheimer’s disease: a study via graph theory from EEG data. Brain Imaging Behav 2017; 11: 473–485.

125.

Wolf

Ueda

. Contribution of eye-tracking to study cognitive impairments among clinical populations. Front Psychol 2021; 12: 590986.

126.

Crook-Rumsey

Howard

Doborjeh

, et al. Spatiotemporal EEG dynamics of prospective memory in ageing and mild cognitive impairment. Cogn Comput 2023; 15: 1273–1299.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.11 MB