Evaluating Electroacupuncture Assessment Strategies for Mild Cognitive Impairment: A Scoping Review
Article information
Abstract
ABSTRACT
Objective:
This scoping review aimed to identify and analyze the range of assessment tools used in electroacupuncture trials for mild cognitive impairment (MCI) and examine emerging trends in study design and evaluation.
Methods:
In May 2025, seven electronic databases, the International Clinical Trials Registry Platform, and ClinicalTrials.gov were searched for clinical reports and protocols on electroacupuncture for MCI. Studies were screened for participant characteristics, intervention protocols, comparators, and outcome measures.
Results:
Seventeen reports and three protocols were reviewed. The Mini-Mental State Examination and Montreal Cognitive Assessment were the most frequently used tools. Recent studies incorporated broader neuropsychological batteries, daily functioning, quality of life, emotional state, and brain activity assessments. Electroacupuncture most often targeted GV20, EX-HN1, and GV24, with 30-min sessions across 24 treatments. Controls included pharmacological treatments, sham electroacupuncture, and cognitive rehabilitation.
Conclusions:
Electroacupuncture appears to be a promising intervention for mild cognitive impairment, particularly the amnestic subtype. However, considerable heterogeneity in outcome measures and protocol design underscores the need for standardized core outcome sets that incorporate cognitive, functional, emotional, and biological domains. Future research should also adopt rigorous sham controls, develop standardized electroacupuncture treatment protocols, ensure transparent safety reporting, and explore the neurophysiological mechanisms underlying electroacupuncture effects.
I. Introduction
With the increasing elderly population, cognitive decline has emerged as a major global public health challenge1. Mild Cognitive Impairment (MCI) represents an initial stage of dementia, underscoring the importance of early diagnosis and intervention. MCI is characterized by impairment in specific cognitive domains such as memory, attention, and executive function, while daily living activities remain relatively preserved2. Approximately 10-15% of individuals with MCI progress to Alzheimer’s disease annually, placing them in a high-risk group2.
Electroacupuncture (EA), a traditional East Asian medical intervention combining electrical stimulation with acupuncture, is proposed to modulate central nervous system excitability and neuroplasticity3. Animal models4,5 have demonstrated that EA may enhance cognitive function through mechanisms such as increased hippocampal blood flow, improved synaptic plasticity, and suppression of neuroinflammation.
In recent years, an increasing number of clinical trials have examined EA’s effects on MCI patients.
However, across studies, the outcome measures used to evaluate therapeutic efficacy have been markedly inconsistent. While cognitive scales such as the Mini-Mental State Examination (MMSE) and Montreal Cognitive Assessment (MoCA) are widely used, other domains-including functional ability, neuropsychiatric symptoms, and neurophysiological markers-are less consistently reported. Furthermore, objective imaging-based assessments, such as functional MRI (fMRI) or event-related potentials (ERP), have been rarely incorporated despite their potential to validate treatment mechanisms.
The lack of standardization in assessment tools makes it difficult to synthesize findings across trials and limits the interpretability of EA’s efficacy. To address this gap, this review was conducted to map the range and frequency of outcome measures used in EA trials for MCI. This review aimed to identify current trends and highlight the need for a standardized and multidimensional evaluation framework in future research.
II. Methods
The study adhered to established methodological frameworks for scoping reviews, incorporating the foundational approach outlined by Arksey and O’Malley6, further developed by Levac et al7., and guided by the reporting standards set forth by the PRISMA-ScR8 and JBI (Joanna Briggs Institute) guidelines9. A comprehensive mapping of clinical studies meeting the predefined inclusion criteria was undertaken, guided by the following research questions.
1. Research questions
The primary objective of this study was to map the existing evidence on the use of EA for MCI. In line with this aim, three research questions were developed to structure the investigation:
1) What types of outcome measures have been used to assess the effects of EA in clinical trials for MCI?
2) How have study designs, participant characteristics, and comparators varied across EA trials for MCI?
3) What trends and methodological patterns can be identified in the use of EA protocols, including acupoint selection, frequency, and treatment duration?
2. Literature searches
A comprehensive search was conducted across key international biomedical databases, including PubMed, Embase, and the Cochrane Library, as well as clinical trial registries such as ClinicalTrials.gov -managed by the U.S. National Library of Medicine under the National Institutes of Health-and the World Health Organization’s International Clinical Trials Registry Platform (ICTRP). To include Chinese-language literature, the China National Knowledge Infrastructure (CNKI) database was searched, while Korean Medical Database (KMBASE), Oriental Medicine Advanced Searching Integrated System (OASIS), and Research Information Sharing Service (RISS) were additionally searched to include Korean-language literature. The search terms used included the disease name “mild cognitive impairment” and the intervention “electroacupuncture.” The literature search was completed on May 15, 2025. All relevant clinical studies available in the databases and trial registries up to the date of search were retrieved. Additionally, reference lists of previously published systematic reviews on related topics were examined to identify any studies potentially missed in the database searches.
1) Selection criteria
Participants included individuals diagnosed with MCI according to internationally recognized diagnostic criteria (e.g., Petersen’s criteria, DSM-IV, NIA-AA), or those identified by clinicians as having MCI based on cognitive screening tools such as the MoCA, MMSE scores. When available, cutoffs were interpreted with adjustments for education level, in line with common practice in East Asian studies (e.g., MMSE ≥17 for illiterate, ≥20 for primary education, ≥24 for secondary education and above).
Interventions analyzed comprised both EA alone and EA combined with other treatment modalities such as cognitive rehabilitation, conventional medications, or herbal medicine.
Study designs included randomized controlled trials (RCTs), non-randomized controlled clinical studies, pre-post comparison studies, case-control studies, cohort studies, and retrospective medical record analyses involving human subjects. In addition to published full-text articles, study protocols and conference abstracts were also considered. There were no restrictions regarding publication year, country, or language.
Comparisons included no treatment, usual care, sham acupuncture or EA, manual acupuncture, and conventional pharmacological therapies.
Outcome measures comprised cognitive function assessment tools such as MoCA, MMSE, and Auditory Verbal Learning Test (AVLT), as well as neuroimaging indicators including fMRI, diffusion tensor imaging (DTI), and single-photon emission computed tomography (SPECT).
2) Exclusion criteria
Studies not conducted on human subjects were excluded, including animal models, in vitro experiments, and literature-based reviews. Additionally, studies involving healthy adults without a clear diagnosis of cognitive impairment, as well as those targeting patients with moderate to severe dementia were excluded. Furthermore, in order to focus on primary, idiopathic MCI, cases attributed to specific etiologies (e.g., vascular MCI, diabetes-related MCI, etc.) were excluded.
Studies that employed only manual acupuncture, pharmacopuncture, moxibustion, or herbal medicine without EA as the intervention were excluded. In cases where a single clinical trial was reported across multiple sources-such as clinical trial registries, study protocols, or result publications-only the most comprehensive publication was included, and the others were excluded as duplicates.
3. Data analysis
To identify the general characteristics of the clinical studies selected for final analysis based on the exclusion criteria, the following data were extracted. The age and sex of participants at the time of study design were examined to determine the demographic characteristics of the study population. If the study design included a distinction between the treatment and control groups, the number of enrolled participants and the number of individuals who completed the clinical trial in each group were assessed. To obtain detailed information on the EA procedures performed, data were collected on the total number of treatment sessions, duration of each treatment, and EA stimulation frequency. The acupuncture points applied in the EA treatment groups of each study were identified, and the frequency of each acupuncture point appearing in the selected literature was analyzed. Additionally, information on outcome assessment measures was investigated.
III. Results
1. Study selection
The total number of searched researches was 988, and the first selection was carried out by excluding 321 duplicated documents. After removing duplicates, 667 studies were screened and assessed based on inclusion and exclusion criteria, leading to the exclusion of 642 studies. Subsequently, five additional studies were excluded due to duplicate publication of the same research. As a result, a final selection of 20 studies (17 completed studies and 3 ongoing studies) was included for analysis (Fig. 1).
2. Characteristics of completed studies
Of the 17 studies, 16 were randomized controlled trials (RCTs)10-25, while one was a retrospective study26. The characteristics of each study are summarized in Table 1.
1) Participants
The retrieved studies enrolled individuals diagnosed with MCI, predominantly using MMSE, MoCA, and other criteria such as DSM-IV, Petersen’s criteria, or TCM pattern diagnosis. Many studies focused specifically on amnestic MCI with MMSE scores typically ranging from 24 to 27. Some Chinese studies10-12 applied TCM syndromic differentiation, such as kidney essence deficiency or liver-kidney yin deficiency, as inclusion criteria. Most participants were elderly, aged 60 and above, and the overall female participation rate was 50.7%.
2) Intervention
The most frequently used acupuncture points were mainly located on the head. GV20 (Baihui) was used 13 times, while EX-HN1 (Sishencong) and GB20 (Fengchi) were used 9 times, and GV24 (Shenting) was used 7 times. Additionally, distal acupuncture points, such as GB39 (Xuanzhong) five times, BL23 (Shenshu), and KI3 (Taixi) were each used four times. In all studies, low-frequency stimulation (≤100 Hz) was utilized. Specifically, the 3-15 Hz range was the most frequently applied, appearing four times. Additionally, alternating frequencies within the 2-100 Hz range were employed in one study25. Most treatments lasted 30 minutes per session, with the number of sessions ranging from 16 to 50. Despite variations in technique, 24 sessions over 30 minutes remained the most common protocol.
3) Comparator interventions
EA was compared against several control conditions, including pharmacotherapy (donepezil or nimodipine), sham acupuncture, basic acupuncture, cognitive training, or no intervention. In most cases, EA alone or in combination with pharmacotherapy produced equal or superior improvements in cognitive scores compared to pharmacological controls. Sham-controlled studies generally showed statistically significant superiority of real EA over sham EA. A few studies14,16 also compared syndrome-differentiated EA against non-syndrome-based EA, finding enhanced effects in the tailored interventions.
4) Outcome measures and safety
The outcome measures of each study were summarized in Table 2. MMSE was the most frequently used assessment tool, appearing 13 times. In studies22-24,26 where MMSE was not utilized, MoCA was used instead, with a total of nine studies incorporating MoCA (including MoCA-K). Several studies expanded the cognitive testing battery to include tools such as Clock Drawing Test (CDT) in 5 studies13,14,16,19,21, and the ADAS-cog in 4 studies22-24,26. In addition, various tools were used to assess cognitive function, including Pattern Recognition Test, Graphic Recognition Test, Clinical Memory Scale (CMS), Boston Naming Test (BNT), and Seoul Neuropsychological Screening Battery-II (SNSB-II).
Psychological and mood-related scales, such as Center for Epidemiologic Studies Depression Scale (CES-D), Hospital Anxiety and Depression Scale (HADS), short version of geriatric depression scale (SGDepS), Perceived Stress Scale (PSS), Beck Depression Inventory (BDI), and Insomnia Severity Index (ISI) were applied in a few studies22-26, though their use was less consistent compared to cognitive tests.
Functional assessment was primarily conducted through evaluations of quality of life. Although three studies17,19,25 stated that they used Activities of Daily Living (ADL) as an assessment tool, it appears that they utilized self-developed instruments to measure ADL. Other studies specified the names of the assessment tools, Korean version of the Activities of Daily Living (K-ADL), Korean version of the Instrumental Activities of Daily Living (K-IADL) in 2 studies22,23 and five-level version of the EuroQol 5-Dimension scale (EQ-5D) in 3 studies22,23,25 used to assess the real-world applicability of cognitive gains. Only one study17 included the Berg Balance Scale (BBS) to examine motor-cognitive interaction. Additionally, Patient Global Impression of Change (PGIC) in one study24, Simplified Nutritional Appetite Questionnaire (SNAQ) and Individualized Medicine Patient Satisfaction Scale (IMPSS) in the other study26 were used to assess functional domains.
Notably, several trials incorporated objective physiological measures, including P300 ERP in three studies12,21,25, NAA/Cr in one study11, FA (Fractional anisotropy) derived from DTI, ReHo (Regional Homogeneity) obtained from resting-state fMRI (rs-fMRI) and serum inflammatory or neurodegenerative biomarkers (e.g., IL-6, hs-CRP) in one study21. These outcomes were included to explore the mechanistic impact of EA on brain function and neural integrity.
Adverse event reporting was mentioned in eleven studies11,13-15,18,20,22-26 out of seventeen studies. In five studies11,20,22,23,25, the adverse event reporting section stated that there were no events to report. However, six studies13-15,18,24,26 provided specific details on adverse events. Yu (2010)13 reported one case of syncope and one case of headache with dizziness. Chen (2011)14 documented seven cases of adverse events, including headaches and hematomas. Zhao (2012)15 recorded five cases of hematomas and three cases of syncope. Luo (2013)18 reported one case of hematoma, with hematomas being relatively common but generally mild. Choi (2021)24 described one serious adverse event, but it was due to a car accident unrelated to EA. Lee (2024)26 noted that adverse event reporting should consider the potential relevance to the herbal medicine intervention rather than acupuncture alone.
5) Main findings
Across the 17 completed studies, EA consistently led to improvements in global cognitive scores. Most trials reported significant increases in MMSE and MoCA scores following EA treatment, either alone or in combination with pharmacotherapy. When compared to donepezil or nimodipine, EA showed equal or superior cognitive outcomes, particularly in memory and pattern recognition tasks (e.g., Liu (2009)10; Yu (2010)13; Chen (2011)14).
Combination therapies (EA+donepezil) showed additive benefits in several trials12,20, where cognitive gains were significantly higher than either intervention alone. In addition, syndrome-based EA protocols, tailored to individual TCM diagnoses, demonstrated superior outcomes compared to fixed acupoint methods in two studies17,18.
Objective physiological indicators showed favorable changes following EA. For example, Liu (2010)11 found improved neuronal integrity through 1H-MRS imaging, and Xu (2017)21 reported enhanced P300 responses and reduced inflammatory markers, supporting the neurobiological plausibility of EA’s cognitive effects.
Functionally, studies such as Li (2013)17 and Yu (2021)25 reported improvements not only in cognitive scales but also in daily functioning (ADL) and self-reported health (EQ-5D). However, not all studies showed additive effects. Kim (2020)22 found no significant difference between EA plus cognitive rehabilitation and cognitive rehabilitation alone.
3. Characteristics of ongoing and planned trials
1) Participants
Three ongoing RCTs27-29 are currently recruiting or underway in East Asia. Participants are generally diagnosed with MCI using Petersen’s criteria, K-MMSE (20-23), and MoCA-K (<23 or <26). The sample sizes in each study27-29 were 64, 24, and 112 participants, respectively, indicating that these studies were not large-scale, similar to previous research.
2) Electroacupuncture intervention
Two trials27,28 employed standardized EA protocols using GV20 (Baihui), EX-HN1 (Sishencong), and GV24 (Shenting) in both cases, with one trial27 additionally incorporating GB20 (Fengchi) and the other28 utilizing GB13 (Benshen). The frequencies used were 3 Hz and 4/20 Hz, while the remaining study29 did not provide specific details. Notably, the Kim 2024 study27 includes 24 sessions over several timepoints (week 0, 7, 13, and 24), allowing for longitudinal outcome tracking. Li (2024)28 uses a relatively shorter protocol with 8 sessions over 6 weeks.
3) Comparator interventions
The ongoing trials include a variety of control conditions. Kim (2024)27 uses computerized cognitive rehabilitation (CCR) as the control group. Li (2024)28 compares EA to a no-EA group, while Zhang (2024)29 uses sham EA. These designs reflect a continued interest in isolating the specific effects of EA and evaluating it against both active and placebo-like interventions.
4) Outcome measures
Compared to earlier studies, ongoing trials adopt a broader array of assessment domains. Cognitive outcomes include MMSE, MoCA-K, ADAS-K-cog, CDR, and domain-specific tools like Auditory Verbal Learning Test (AVLT), Trail Making Test (TMT) Part -A/B, and BNT. There is a clear increase in neuropsychological test battery depth, particularly in Zhang (2024)29. Mood assessments are also more systematically included, with use of Geriatric Depression Scale (GDS), CES-D, HADS, and SGDepS. Indicators such as EQ-5D-5L, ADL, and Pittsburgh Sleep Quality Index (PSQI) were utilized to assess functional outcomes. Physiological measures, such as ERP, rs-fMRI, Alzheimer’s-related biomarkers, and inflammatory factors, are also incorporated, reflecting a multimodal evaluation trend.
Among the three studies, two27,28 explicitly outlined plans for collecting safety information related to the intervention.
IV. Discussion
1. Summary of key findings and implications
This study presents an analysis of clinical research trends regarding the application of EA for MCI. The distribution of publication years across the reviewed literature suggests sustained interest in EA treatment for MCI. While earlier studies primarily focused on comparisons between EA and pharmacological treatments, recent research has increasingly examined the combined effects of EA with cognitive rehabilitation therapy or computer-based rehabilitation therapy. Additionally, the diversity of control groups has expanded, with studies incorporating sham acupuncture and sham EA. A review of ongoing clinical trial plans further reflects these trends, as one study compares EA with computer-based rehabilitation therapy, another utilizes a sham EA group as a control, and the third designates a non EA group as a comparator.
A notable shift has occurred in patient selection criteria-from traditional pattern differentiation based on kidney essence or liver-kidney yin deficiency, as seen in earlier studies, toward the use of standardized diagnostic frameworks. This trend suggests a growing emphasis on clinical standardization, which may serve as a foundation for future guideline development.
A consistent trend across the reviewed studies is the frequent selection of specific acupoints associated with traditional theories of cognitive function and organ deficiency. The combination of GV20 (Baihui), EX-HN1 (Sishencong), GB20 (Fengchi), and GV24 (Shenting) was the most commonly employed, appearing in a total of nine studies13-16,18,20-23 This consistency suggests a shared clinical rationale rooted in the acupoints’ purported ability to regulate brain function and promote mental clarity. In addition, the recurring use of BL23 (Shenshu), GB39 (Xuanzhong), and KI3 (Taixi) in several studies10,11,20 reflects the enduring influence of the traditional view20 that kidney deficiency is a fundamental pathophysiological mechanism underlying cognitive decline in MCI.
The exact mechanism of EA has not yet been fully clarified, but animal studies30,31 have confirmed that EA treatment reduces inflammation, oxidative stress, and apoptosis in models of septic brain injury and ischemia-reperfusion-induced cognitive impairment. Furthermore, in animal models of dementia32, EA has been shown to regulate inflammatory cytokines, modulate glial cell activity, and potentially enhance hippocampal function. However, the optimal frequency parameters remain unclear. Different studies have adopted varying approaches, reflecting a lack of consensus in the field. For instance, Yu (2021)25 employed a mixed-frequency protocol alternating between 2 Hz and 100 Hz, a setting that, according to recent study33, may increase blood-brain barrier (BBB) permeability via modulation of tight junctions. Feng (2016)19 on the other hand, used high-frequency stimulation (80-100 Hz), which has been associated in animal models with enhanced synaptic transmission in the hippocampus34. However, standardizing the frequency settings appears to be necessary for establishing a clinical guideline for EA applications.
In contrast to the variability in frequency settings, the duration and frequency of EA sessions have shown relative consistency, typically 30 minutes per session over 24 sessions. In upcoming studies, one study27 maintains this established protocol (30 minutes, 24 sessions), while another29 does not specify the treatment duration, and a third study28 applies only eight sessions, raising concerns that the treatment period may be insufficient for optimal therapeutic outcomes.
The design and application of sham EA also warrant careful consideration. Approaches vary, with one study25 using non-penetrating needles on classical acupoints, while the other24 apply sham stimulation to non-acupoints. Given the potential for cutaneous contact alone to induce physiological responses35, selecting an appropriate sham protocol is crucial. The methodological variability in sham interventions underscores the need for clearly defined, purpose-driven control conditions that align with the specific aims of each study.
Regarding outcome assessment, the MMSE was a key criterion for participant recruitment and was the most frequently used assessment tool in the analyzed studies. The MoCA also emerged frequently, with one study36 suggesting that MoCA is more sensitive than MMSE. This review confirms that both tools are widely adopted and clinically relevant for identifying MCI and highlights that EA contributes to measurable improvements in these cognitive scores.
Although these tools are efficient and well-validated for general cognitive screening, they have limited sensitivity in detecting subtle or domain-specific changes. To address this limitation, a subset of studies19,23-25 incorporated additional neuropsychological tests-such as ADAS-cog, TMT, AVLT, and BNT-particularly in more recent or ongoing trials27,29. This shift highlights the growing recognition of the need for more comprehensive assessment tools.
Moreover, recent studies have begun to incorporate neuropsychiatric and mood assessments, such as the HAMD, HAMA, NPI, and others. This broader scope of assessment represents a significant step toward a more comprehensive evaluation of therapeutic effects. It also reflects an emerging understanding that cognitive impairment in MCI often coexists with emotional and behavioral symptoms, which may significantly impact patients’ quality of life.
Following this trend, increasing attention has been given to neurophysiological and neuroimaging markers to elucidate the mechanisms underlying cognitive improvement. Although multiple factors contribute to the pathophysiology of cognitive decline, studies have consistently highlighted reduced cerebral blood flow and impaired functional connectivity as key mechanisms. Earlier investigations-such as Liu (2010)11, who examined NAA/Cr via magnetic resonance spectroscopy (MRS), and Sun (2010)12 and Yu (2021)25, who explored event-related potentials (P300)-laid the groundwork for this line of inquiry. More recent studies, including that of Xu (2017)21, employed a combination of ERP, DTI, and resting-state fMRI to assess the effects of EA on cerebral blood flow and connectivity. This trend continues with Zhang29’s upcoming study, which incorporates a comprehensive set of biomarkers-including Alzheimer’s-related proteins, neuroinflammatory factors, neurotrophins, neurotransmitters, and advanced imaging techniques-to further investigate EA’s effects. These advancements point toward a deeper integration of neuroimaging and molecular diagnostics in EA research and suggest a promising direction for future studies exploring early-stage cognitive decline.
Finally, the safety profile of EA has been generally favorable across both completed and ongoing trials. Among the 17 completed studies, 1111,13-15,18,20,22-26 reported safety outcomes, with only minor adverse events such as transient headaches and subcutaneous hematomas. Overall, EA was found to be a relatively safe intervention.
2. Comparison with previous literature
Our findings align with prior systematic reviews37,38 suggesting that EA may offer cognitive benefits for people with MCI or early dementia, although much of the prior work has focused more on efficacy than methodology. What distinguishes this review is its focus on how those effects are measured-and the implications of that diversity. The heterogeneity in outcome selection complicates direct comparison across trials and highlights the lack of standardized core outcome sets (COS) in EA-related cognitive research.
While Western trials of cognitive therapies (e.g., pharmaceuticals or computerized cognitive training) have increasingly adopted standardized measurement protocol, EA studies remain methodologically inconsistent, particularly in outcome domains beyond cognition. However, the ongoing trials reviewed here demonstrate a promising shift toward more standardized and multidimensional outcome strategies.
3. Limitations
This review has several limitations. First, despite efforts to be comprehensive, some relevant unpublished or non-indexed trials may have been overlooked, introducing potential publication bias. Second, while outcome measures were the primary focus, factors such as intervention fidelity, practitioner expertise, and adherence to EA protocols were not systematically evaluated. Third, the inclusion of both completed and ongoing trials may conflate intended effects with observed outcomes, making interpretations of ongoing studies speculative. Moreover, most of the included studies were small-scale (N<100), which limits statistical power and undermines the reliability and generalizability of the findings. The absence of meta-analysis or quantitative synthesis further restricts the ability to estimate effect sizes and assess consistency across studies. Additionally, no formal assessment of study quality-such as the use of risk-of-bias (RoB) tools-was conducted, which weakens the methodological rigor of the review. As a result, the findings should be interpreted with caution and remain largely qualitative.
4. Suggestions for future research
To address these limitations, future EA research on MCI should prioritize greater standardization in outcome selection. Developing and adopting a core outcome set-including cognitive, functional, emotional, and biological domains-would enhance comparability and facilitate evidence synthesis across studies. Incorporating neuroimaging and biomarker-based outcomes could further strengthen causal inferences regarding EA’s mechanisms of action. Moreover, integrating patient-reported outcomes and longitudinal follow-up measures would provide valuable insights into the sustainability and functional relevance of cognitive gains. Studies should also improve transparency in reporting adverse events and protocol fidelity to enhance replicability and clinical translation. In addition, refining sham control techniques is essential to minimize placebo effects and ensure the internal validity of EA trials. Along with randomized controlled trials, future research should include mechanistic investigations to elucidate how EA contributes to cognitive function improvement.
V. Conclusion
This scoping review synthesized the clinical research landscape surrounding the use of EA in the management of MCI, with a focus on the diversity and evolution of outcome measures. The reviewed studies demonstrate that EA, whether used alone or in combination with pharmacological or cognitive therapies, shows potential in improving cognitive function, particularly in individuals with amnestic MCI.
Earlier trials emphasized comparisons with pharmacological treatments, whereas more recent and ongoing studies have adopted broader and more standardized assessment strategies, integrating cognitive, emotional, functional, and neurophysiological domains. In parallel, the design of comparator groups has become more rigorous, incorporating sham EA and computerized cognitive training, thereby enhancing the methodological quality of current research.
Despite promising findings, substantial heterogeneity remains in EA intervention protocols and outcome selection. Core cognitive tools such as MMSE and MoCA remain dominant, but increasing incorporation of multidomain neuropsychological testing and neuroimaging techniques suggests a shift toward comprehensive evaluation frameworks. However, variation in stimulation frequency, session number, and sham design continues to challenge inter-study comparability.
To establish EA as a reliable and evidence-based intervention for early cognitive decline, researchers should prioritize the development of standardized EA protocols, refine sham control techniques and adopt core outcome sets that encompass cognitive, behavioral, emotional, and biological endpoints. These efforts are essential to ensure comparability across studies, enhance clinical applicability, and eliminate placebo effects, thereby securing scientific validity. Additionally, incorporating patient-reported outcomes, ensuring long-term follow-up, and maintaining transparent safety reporting will further strengthen the scientific credibility and clinical utility of EA.
In summary, EA represents a promising adjunctive treatment for MCI, but methodological standardization and broader assessment approaches are critical to advancing its clinical application and scientific validation.