Evidence integration of acupuncture for prevention and treatment of Alzheimer’s disease and mild cognitive impairment from a neuroimaging perspective

Abstract

Background

Acupuncture has clinical potential in treating Alzheimer's disease (AD) and mild cognitive impairment (MCI), but there is a lack of systematic review and presentation of clinical evidence from the perspective of neuroimaging in this field.

Objective

To conduct a systematic review of clinical studies on acupuncture for AD and MCI from the perspective of neuroimaging, and to comprehend the evidence distribution of relevant research.

Methods

This article retrieved all the neuroimaging clinical studies on acupuncture treatment for AD and MCI that were published and included in the seven databases from their establishment until February 22, 2025. It analyzed and organized the data based on the PICOS (Population, Intervention, Comparison, Outcome, Study design) principle, and presented the quality and distribution of evidence.

Results

A total of 58 studies were included. The diagnostic criteria for the research subjects mainly refer to the standards of Western medicine. The task design was mostly two-arm before-and-after comparisons and single-group immediate studies, with the intervention measures mainly including hand acupuncture and electroacupuncture. The study employed 8 neuroimaging techniques and 29 outcome measures, with a primary focus on brain functional activation regions and brain functional connectivity. Included studies had high bias risk in blinding design/implementation; overall evidence quality was acceptable.

Conclusions

Acupuncture for AD and MCI demonstrates clear efficacy, which is supported by imaging evidence. In the future, more large-sample, multi-center joint clinical studies using neuroimaging methods will be needed to further investigate AD and MCI, providing more high-quality evidence-based medical evidence in this field.

Keywords

acupuncture Alzheimer's disease mild cognitive impairment neuroimaging radiomics

Introduction

Alzheimer's disease (AD) is a neurodegenerative disease with a complex etiology and severe intellectual disability, and it is the most common type of dementia.¹ The clinical stages can progress from normal cognition to mild cognitive impairment (MCI), and then to dementia.² There are at least 50 million people with dementia worldwide, and this number is expected to reach 152 million by 2050, among which approximately 60%-70% are AD patients. However, the existing research on the clinical pathogenesis of AD still remains unclear, and most studies on preventive treatments aimed at improving the pathological mechanisms of AD are still in the clinical trial stage. The currently approved AD treatment drugs, cholinesterase inhibitors and N-methyl-D-aspartate receptor antagonists, are only symptomatic treatments for improving cognitive dysfunction. They do not have significant therapeutic effects on other neuropsychiatric symptoms and are associated with peripheral cholinergic side effects (such as leg cramps and intestinal discomfort) and other potential risks. There is an immediate requirement for clear and efficient new methods and strategies in the field of AD prevention and treatment. As a traditional Chinese medicine external therapy, acupuncture is characterized by simplicity in operation, low cost, quick effect, high safety and few side effects. Studies over the past few decades have shown that acupuncture has significant potential efficacy in treating brain diseases.³ In recent years, various studies have indicated that acupuncture exerts a good therapeutic effect on AD by participating in mechanisms such as amyloid-β (Aβ) deposition, inhibiting tau protein phosphorylation, and regulating synaptic and neurotransmitter transmission.⁴ Clinical studies have confirmed that acupuncture can improve the overall function, cognitive impairment and activities of daily living in AD patients. Acupuncture has the potential to become an alternative therapy for AD and shows developmental prospects. However, in current research in this field, there is a lack of objective, high-quality evidence regarding its clinical mechanism of action and efficacy. Neuroimaging techniques obtain images by utilizing various standard imaging equipment and extract quantifiable imaging features. They can reflect more subtle structural or functional differences or heterogeneities at a single or multiple time points. With the development of imaging and biomarker technologies, various imaging indicators have become a major boost to core research in this field, but there is still a lack of overall analysis of these research evidence. Based on the mining and collation of literature data, this paper takes radiomics as the anchor point to provide systematic and multi-angle imaging evidence for the efficacy of acupuncture in the prevention and treatment of AD and MCI, and at the same time offers more information references for subsequent clinical research design, so as to optimize research topics and clinical research design schemes.

Methods

Document retrieval

We searched seven major databases including CNKI, Wanfang Database, VIP Database, China Biomedical Literature Database, PubMed, Embase, and CENTRAL (Cochrane Controlled Trials Register). The retrieval time was from the establishment of the databases to January 22, 2025. A combination of subject terms and free words was used for retrieval. The specific retrieval strategies for each database are detailed in Supplemental Material 1.

Inclusion and exclusion criteria

Inclusion criteria: (1) Population characteristics: The main research subjects are patients with a clear diagnosis of AD or MCI. The study can include healthy subjects under the same conditions, with no restrictions on age, gender, region, or ethnicity. (2) Intervention measures: The observation group receives acupuncture treatment (or combined with other treatment methods), with no restrictions on specific needles, needling methods, moxibustion methods, treatment sites, or treatment duration; the control group has no restrictions on intervention measures. (3) Control type: The purpose of the experiment is to explore the efficacy of acupuncture. (4) Outcome indicators: At least one imaging result is included as an observation indicator. (5) Literature type: All are clinical studies published in journals.

Exclusion criteria: (1) Animal experiments, reviews, meta-analyses and systematic reviews, experimental protocols, scientific magazines, and dissertations and conference papers, among other types of literature research. (2) Duplicate literature in the database. (3) Literatures that do not meet the above inclusion criteria or provide insufficient information.

Literature screening and data extraction

In the initial screening process, we used Medical Literature Manager to remove duplicates of the retrieved literature. After reading the full text of each literature, we extracted and sorted out the following information from 58 articles: basic characteristics (journal name, publication year, authors, author's address, fund support); basic characteristics of the subjects (inclusion and exclusion criteria, diagnostic criteria, age, dominant hand, education level, research type and so on); intervention and control (Task design, image design, group design, acupuncture mode, intervention category, the frequency, cycle and number of acupuncture sessions and so on); outcome indicators (Radiological technology, Image mode, Image data indicators and so on). The quality of the included studies was assessed using both the Cochrane Risk of Bias Tool and the MINORS criteria. Meanwhile, excluded literature that did not meet the inclusion and exclusion criteria. Subsequently, the information of the finally included 58 articles was merged and sorted out.

Literature screening was conducted independently by two researchers, who initially screened and re-screened the literature according to the inclusion criteria. After screening, the two researchers compared their screening results. For inconsistent parts, joint discussions were held or the third researcher was invited to participate, and consensus was reached in accordance with the criteria to determine the final included literatures and the information to be entered. No additional statistical models were involved.

Statistical analysis

Data analysis was conducted using a combination of text and charts for multi-angle description. We used Microsoft Office Excel 2019 to merge and sort out the above collected information, see Supplemental Material 2; Origin Pro 2025 was used to draw heat distribution maps, bar charts, Sankey diagrams, bubble charts and other graphs to describe and analyze the clinical characteristics of the literature. Review Manager 5.4 was used to create the figures related to quality assessment.

Results

Literature screening results

A total of 1093 articles were initially retrieved in this study. After removing duplicates and excluding conference papers and dissertations, 933 articles were initially included for preliminary screening. By reading the titles and abstracts, 121 articles were included for downloading. Among them, 2 articles could not obtain the full text, and finally 119 articles entered the full-text reading and screening stage. In strict accordance with the inclusion and exclusion criteria, after independent screening and cross-checking by two researchers, a total of 58 eligible clinical studies were included. The screening process is shown in Figure 1.

Figure 1.

Flowchart of literature inclusion and screening process

Analysis of basic characteristics of included literature

The number of publications on clinical studies of acupuncture for the prevention and treatment of AD and MCI has generally shown an upward trend since the first report in 2000, with a trough in 2016 and a peak in 2022, indicating the continuous increase in research attention in this field in recent years. The brief trough period may be attributed to the publication of high-quality research and the preliminary application of new imaging technologies, as shown in Figure 2b. The 58 included articles showed regional aggregation in terms of geographical origin. Beijing (16 articles), Heilongjiang Province (11 articles), and Guangdong Province (10 articles) were the most concentrated regions, totaling 37 articles, accounting for 63.79%. The remaining articles were scattered in other provinces in China (19 articles, 32.86%) and Busan, South Korea (2 articles, 3.45%). The regional distribution of domestic literature is shown in Figure 2a. The majority of published papers were in journals ranked by SCI (19 papers, 32.76%) and Science and Technology Core journals (13 papers, 22.41%), as shown in Figure 2b. Among the studies that received funding support (44 papers, 75.86%), national-level fund support accounted for the highest proportion (29 papers, 50.00%), followed by provincial-level fund support (15 papers, 25.86%). The details of fund support are shown in Figure 2c.

Figure 2.

Basic characteristics of included literature. (a) Regional heat distribution map, (b) Journal level quantity map, (c) Fund level proportion map, (d) Research type distribution map. RCT: clinical randomized controlled trial; nRCT: non-clinical randomized controlled trial; pre-post: Pre-and-post control trial.

The 58 articles mainly included randomized controlled trials (33 articles, 56.9%) and pre-post experiments (23 articles, 39.66). The sample size of the former was mostly distributed in 41-60 cases (12/33), while the latter was mostly 21-40 cases (13/23). The overall sample size was mainly concentrated in 21-40 cases (25 articles, 43.1%) and 41-60 cases (14 articles, 24.14%), still dominated by small-sample studies. See Figure 2d.

Basic characteristics of the included literature

Population characteristics. Basic characteristics of research subjects. The main characteristics of the research subjects in the 58 included articles (including disease classification, education level, dominant hand, and age range) are shown in Table 1. Some studies did not report detailed population characteristics. Among the articles that reported relevant information: the education level of the research subjects was mainly concentrated in middle school and above (accounting for 29.31%). The research subjects were mainly right-handed (accounting for 48.28%). The age range of the research subjects was mainly 50-80 years old. It is worth noting that in the articles that clearly gave the age range, almost all studies included subjects aged 70-80 years, suggesting that this age group may be a high-incidence period of AD or MCI or a focus of clinical research.

Table 1.

Basic information of research objects

Type	Frequency	Proportion
Disease Classification
Cognitive impairment	34	58.62%
Alzheimer's disease	22	37.93%
Both	2	3.45%
Dominant hand
Unclear	30	51.72%
Right-handed	28	48.28%
Education Level
Primary school and above	10	17.24%
Middle school and above	17	29.31%
High school and above	5	8.62%
Unclear	26	44.83%
Age Range
50-80	31	53.45%
60-80	13	22.41%
60-90	5	8.62%
50-90	3	5.17%
40-80	2	3.45%
70-90	1	1.72%
Unclear	3	5.17%

Diagnostic criteria for diseases. This study counted all diagnostic criteria used in the 58 included articles and extracted the top seven Western diagnostic criteria with the highest citation frequency, all Chinese diagnostic criteria, and frequently cited evaluation scales. The Western diagnostic criteria with the highest citation frequencies were the NINCDS-ADRDA diagnostic criteria (17/58), Diagnostic and Statistical Manual of Mental Disorders (11/58), and mild cognitive impairment as a diagnostic entity (9/58). The evaluation scales were mainly the Mini-Mental State Examination (MMSE) (45/58), Clinical Dementia Rating (CDR) (30/58), and Montreal Cognitive Assessment (MoCA) (28/58), as shown in Table 2.

Table 2.

Frequency table of evaluation scales, Chinese and Western standards for subject diagnosis in included literature

Western Diagnosis	Frequency
NINCDS-ADRDA diagnostic criteria	17
Diagnostic and Statistical Manual of Mental Disorders	11
Mild cognitive impairment as a diagnostic entity	9
Chinese Guidelines for the Diagnosis and Treatment of Dementia and Cognitive Impairment	5
NIA-AA diagnostic criteria	4
Current concepts in mild cognitive impairment	4
Chinese Expert Consensus on the Prevention and Management of Cognitive Impairment	3
Others 1	13
Chinese Diagnosis	Frequency
Reference Standards for Syndrome Differentiation of Deficiency Syndromes in Traditional Chinese Medicine	3
Guidelines for Clinical Research of New Traditional Chinese Medicines	2
Chinese Medicine Clinical Diagnosis and Treatment Terminology - Syndrome Section	1
Stages of Dementia and Diagnostic Criteria in Clinical Trials of Traditional Chinese Medicine	1
Diagnostic, Syndrome Differentiation and Classification, and Efficacy Evaluation Criteria for Senile Dementia (Draft for Discussion)	1
Research on the Correlation between Cognitive Levels of the Elderly and TCM Syndromes	1
Evaluation Scale	Frequency
Mini-Mental State Examination (MMSE)	45
Montreal Cognitive Assessment (MoCA)	28
Clinical Dementia Rating (CDR)	30
Activity of Daily Living Scale (ADL)	12
Auditory Verbal Learning Test (AVLT)	8
Others 2	51

Others 1: Including 5 other diagnostic reference standards or literatures with low application frequency (1 time), and blank citations (8 articles did not mention the corresponding diagnostic reference standards or literatures). Others 2: Including 21 other evaluation scales with low citation frequency. NINCDS-ADRDA diagnostic criteria: American Society of Neurology, Language Disorder, and Stroke-Alzheimer’s Disease and Related Diseases Working Group criteria. NIA-AA diagnostic criteria: National Institute on Aging/Alzheimer’s Association criteria.

Intervention and control methods

Acupuncture intervention design and study group design. In the 58 articles, acupuncture intervention designs were categorized into two types: immediate acupuncture and multiple acupuncture, which had significant differences in frequency of use and specific acupuncture methods: immediate acupuncture (21 articles, 36.21%) was used less frequently, while multiple acupuncture (37 articles, 63.79%) was the more mainstream acupuncture intervention design; Regarding the choice of acupuncture methods, the former only used manual acupuncture and electroacupuncture, while the latter involved various acupuncture methods, as shown in Figure 3a. Overall, manual acupuncture had the highest application frequency (36 times, 62.07%), while acupoint injection (2, 3.45%) and acupoint catgut embedding (1, 1.72%) were rarely used in studies possibly due to their operational difficulty and poor controllability, as shown in Figure 3b. Meanwhile, for multiple acupuncture sessions, we further analyzed the intervention duration, treatment frequency, and total number of sessions, as shown in Table 3. Most studies had an intervention frequency mainly concentrated in the ranges of 3-4 times/week and 5-6 times/week, which is consistent with clinical acupuncture treatment protocols. The study durations were mostly 1-2 months, and only a very small number of studies had a trial duration of more than 6 months.

Figure 3.

Distribution of acupuncture task design, acupuncture methods, and characteristics of the intervention and the control. (a) Acupuncture task design diagram, (b) Frequency diagram of acupuncture intervention methods, (c) Group design (intervention control) diagram. A: acupuncture; A1, A2: Two different types of acupuncture methods; SA: sham acupuncture; Oth: Other therapies; All A: In the same study, the control group was different, but the acupuncture method was the same.

Table 3.

Frequency and cycle table of multiple acupuncture interventions

Total number of sessions	Number of articles (frequency)
1-20	4 (10.815%)
21-40	22 (59.46%)
41-60	9 (24.32%)
61-80	1 (2.70%)
81-100	1 (2.70%)
Frequency of interventions
1-2 times per week	2 (5.41%)
3-4 times per week	15 (40.54%)
5-6 times per week	11 (29.73%)
7 times per week	8 (21.62%)
Once per month	1 (2.70%)
Duration of interventions
1-30 days	8 (21.62%)
31-60 days	21 (56.76%)
61-90 days	5 (13.51%)
More than 90 days	3 (8.11%)

In terms of study group design, two-arm pre-post controlled studies (24/58, 41.38%) and single arm studies (17/58, 29.31%) were the main types. Multi-arm task designs were rare and all employed pre-post control. In terms of intervention control settings, single arm studies only used acupuncture intervention, possibly focusing on the direct effect of acupuncture; in two-arm studies, the “acupuncture VS sham acupuncture” control method was the most common (14 items in total); the intervention controls of three-arm and four-arm studies were more complex, including comparisons of multiple intervention methods, as shown in Figure 3c.

Immediate acupuncture: refers to only one complete acupuncture operation during the experiment, while observing changes in brain neural activity and other imaging indicators of the subjects; Multiple acupuncture: refers to multiple systematic acupuncture treatments during the experiment, and periodically observe changes in brain neural activity and other imaging indicators of the subjects at the baseline stage and during or after the treatment.

Immediate test: Image data is collected and compared immediately; Before-and-after control test: Image data is collected twice/multiple times before and after the baseline period and the experimental period for immediate comparison.

Acupoint prescriptions. This study counted all acupoints used in acupuncture treatments in the 58 articles, and summarized the top 15 acupoints with the highest frequency of use in Table 4, Figure 4a, and Figure 4b. Baihui (GV20) (27 times, 46.55%) and Taixi (KI3) (26 times, 44.83%) had the highest frequency of use, ranking at the top. Sishencong (EX-HN1) (25.86%), Hegu (LI4) (24.14%), Shenmen (HT7) (22.41%), Taichong (LR3) (25.86%), and Zusanli (ST36) (25.86%) were also frequently used acupoints.

Figure 4.

Human display and frequency of high-frequency acupoints in included literature. (a) Human display of acupoints, (b) Frequency of acupoint use.

Table 4.

Frequency, location, and meridian of high-frequency acupoints

Imaging Task Design

Location	Acupoint	Meridian	Frequency
Head and Neck	Baihui (GV20)	Du Meridian	27
	Sishencong (EX-HN1)	Extraordinary Points	15
	Shenting (GV24)	Du Meridian	9
	Fengchi (GB20)	Gallbladder Meridian	7
	Yintang (EX-HN3)	Du Meridian	6
	Dazhui (GV14)	Du Meridian	5
Upper Limbs	Hegu (LI4)	Large Intestine Meridian	14
Upper Limbs	Shenmen (HT7)	Heart Meridian	13
Trunk	Guanyuan (CV4)	Ren Meridian	10
Lower Limbs	Taixi (KI3)	Kidney Meridian	26
	Taichong (LR3)	Liver Meridian	15
	Zusanli (ST36)	Stomach Meridian	15
	Xuanzhong (GB39)	Gallbladder Meridian	12
	Fenglong (ST40)	Stomach Meridian	8
	Sanyinjiao (SP6)	Spleen Meridian	5

The imaging task designs in the 58 included articles were mainly divided into four types: Resting state mode (33 articles, 56.90%), Non-Repetitive Event-Related (NRER) design (10 articles, 17.24%), single-BLOCK design (6 articles, 10.34%), and BLOCK design (7 articles, 12.06%). The Resting state mode was the mainstream design mode, focusing on changes in brain imaging-related indicators in AD or MCI patients in the resting state during the baseline period and after experimental intervention.

During the statistical analysis, we discovered that the group design corresponding to the Resting state mode primarily consisted of acupuncture versus sham acupuncture control, as well as other therapy combined with acupuncture versus the therapy in two-arm studies. NRER design, single-BLOCK design, and BLOCK design, in terms of the control design, mainly adopt acupuncture (with pre-and post-control). The former is mostly found in studies involving multiple acupuncture sessions, while the latter three dominate in studies of immediate acupuncture, as shown in Figure 5 and Table 5.

Figure 5.

Distribution of imaging task design and corresponding group design. A: acupuncture; A1, A2: Two different types of acupuncture methods; SA: sham acupuncture; Oth: Other therapies; All A: In the same study, the control group was different, but the acupuncture method was the same; Three-arm 1, 2, 3, Four-arm 1, 2 are detailed in Supplemental Material 2.

Table 5.

Distribution of imaging task design and corresponding acupuncture intervention design

Acupuncture Intervention Design	Imaging Task Design	Frequency
Immediate Acupuncture	NRER design	10
	single-BLOCK design	6
	BLOCK design	3
	Others	2
Multiple Acupuncture	BLOCK design	4
Multiple Acupuncture	Resting state mode	33

BLOCK design: Intermittent stimulation, alternating cycles of stimulation and rest with the same duration, recording all data; single-BLOCK design: Only one period of stimulation and one period of non-stimulation, recording all data; NRER design: Only one stimulation, collecting one piece of data before and after the stimulation; Resting state mode: Collect resting state data once at the baseline period and once after the treatment ends.

Outcome characteristics

Update of imaging technology. The 58 articles included a total of 8 imaging technologies: magnetic resonance imaging (MRI) (including functional MRI and resting-state MRI in this paper) (36 articles), event-related potentials (ERP) (6 articles), magnetic resonance spectroscopy (MRS) (5 articles), transcranial doppler ultrasound (TCD) (4 articles), electroencephalography (EEG) (4 articles), functional near-infrared spectroscopy (fNIRS) (3 articles), 3D arterial spin labeling (3D-ASL) (2 articles), and diffusion tensor imaging (DTI) (3 article).

The use of imaging techniques was dominated by MRI, with multiple imaging techniques used in parallel. The distribution by year was roughly divided into three stages: before 2008, MRI and ERP were the mainstream imaging methods in this field; from 2010, EEG and MRS began to be applied; since 2017, DTI, TCD, 3D-ASL, and fNIRs have also been used in research in this field, as shown in Figure 6.

Figure 6.

Yearly distribution of various imaging methods and their frequencies.

Imaging methods and their indicators. Among the 58 related studies, MRI was the most commonly used imaging method, and the imaging indicator with the highest frequency was Brain functional activation areas or images (11 times, 18.97%). The three high-frequency outcome indicators of MRI: Brain functional activation areas or images, Functional connectivity analysis (10 times,17.24%), and ALFF or fALFF values or images (10 times, 17.24%) were all functional indicators of local activation. Brain function activation areas and images focus on the localization of task-related activation areas; ALFF or fALFF values or images focuses on detecting the intensity of local spontaneous brain activity; The functional connectivity analysis focuses on the functional coordination and integration of neuronal activities between different brain regions. The three aspects—structure, function, and connectivity—respectively cover the specific changes in brain activity manifestations.

ERP (6 times, 10.34%) and MRS (5 times, 8.62%) were the next most frequently used. The former focuses on detecting the cognitive processing process of the brain and can accurately reflect the time sequence in the cognitive process; the latter focuses on the metabolic status of brain neurons at the cellular and molecular levels and reflects the functional state of neurons through the ratio of related metabolites. The key indicators of focus were P3 latency and amplitude (6 times, 10.34%) and related metabolite ratios (the frequency of metabolite ratios was not directly shown in Table 6, and only the frequency of all metabolites was counted separately, see Supplemental Material 2). In addition, TCD, fNIRS, 3D-ASLand DTI were also involved in a small number of studies, as shown in Table 6.

Table 6.

Frequency of imaging methods and their indicators

Radiological Technology (N)	Image data indicators	Frequency
MRI (36)
	Brain functional activation areas or images	11
	Functional connectivity analysis	10
	ALFF/fALFF values or images	10
	ReHo values or images	5
	Brain network pattern analysis	4
	Diffusion coefficient (DC) or image	1
	Functional connectivity density (lrFCD) or image	1
	Volume of the hippocampal subregion	1
ERP (6)
	The latency and amplitude of P3	6
	The latency and amplitude of N2	2
MRS (5)
	N-Acetylaspartic Acid (NAA)	5
	Creatine (Cr/tCr)	5
	Choline (Cho)	2
	Inositol (mI)	2
EEG (4)
	Electroencephalogram abnormality degree/abnormality rate	2
	Correlation indicators of each frequency band (α band, θ band, β1 band, β2 band, γ band, δ band)	2
	Correlation dimension values (DC)	1
TCD (4)
	Average blood flow velocity (Vm)	3
	Cerebral blood flow velocity (Vp, Vd)	1
	Cerebral vascular pulsation and resistance index (PI, RI)	1
DTI (3)
	Anisotropic fractional value (FA) or image	3
fNIRS (3)
	Oxygenated hemoglobin concentration (Oxy⁃Hb)	3
	Brain functional activation regions and image	2
	Functional connectivity analysis	2
	Graph theory indicators (global efficiency, local efficiency, small-world characteristics)	1
3D-ASL (2)
	Relative cerebral blood flow (rCBF) and its images	1
	Arterial spin labeling signal intensity	1

The 58 articles used different numbers of imaging methods, and the corresponding numbers of literature-imaging methods-imaging indicators were different. “N” in the table is the frequency of imaging methods used.

Imaging focus.

AD or MCI-imaging methods: MRI was frequently used in separate/combined studies of both types of patients, followed by TCD, EEG, and MRS. However, there were differences in the selection and focus of specific methods: studies with AD as the subject included three imaging methods: TCD, ERP, and 3D-ASL; MCI included two imaging methods: fNIRs and DTI, as shown in Figure 7a.

Figure 7.

(a) Frequency distribution of subject disease types and corresponding imaging methods. (b) Yearly distribution frequency of MRI imaging indicators.

MRI mining analysis: We extracted a total of 36 studies using MRI and further sorted out their imaging indicators, as shown in Figure 7b. In the early stage, the corresponding indicators of MRI were concentrated in brain activation areas and images, and brain functional connectivity; in recent years, three indicators: amplitude of low-frequency fluctuation (ALFF), fractional amplitude of low-frequency fluctuation (fALFF), and Regional Homogeneity (ReHo) have gradually become research hotspots. Compared with the former two, the latter three indicators, which have received more attention in recent years, are more detailed in reflecting brain functional connectivity. ALFF can reflect the intensity of spontaneous brain function activity, and FALFF value and ReHo have higher relative accuracy in displaying local functional connectivity and can avoid noise interference, which to a certain extent reflects the progress of imaging technology.

Quality evaluation

In this study, the Cochrane Risk of Bias Tool for Randomized Controlled Trials was used to assess the reporting quality of 33 randomized controlled trial (RCT) studies. The Methodological Index for Non-Randomized Studies (MINORS) criteria were applied to evaluate the methodological quality of 23 before-after self-controlled studies (only the first eight criteria were assessed) and 2 non-RCT studies (all criteria were assessed). The specific results are as follows.

Quality evaluation of RCTs. Among the 33 studies, 14 studies had a low risk in terms of the generation of the random sequence (described specific and reliable random methods, including 10 studies using random number tables and 4 studies using computer-generated random numbers). The remaining 19 studies were uncertain regarding the specific generation process (only mentioning randomization); in terms of the random allocation scheme, only 4 studies had a low risk (described reliable allocation schemes, including 3 studies using opaque envelopes method and 1 study using random number remainder allocation method). The allocation concealment aspect was not mentioned in the remaining studies. Regarding the blinding aspects of the implementers and participants, only 3 studies were classified as low risk (as they only included imaging indicators as outcome measures, which did not affect the objective conclusions). In terms of the blinding aspect of outcome assessment, 8 studies were included in the low-risk category (mentioning reliable implementation of blinding or the fact that not setting blinding would not affect the outcome indicators). Regarding the completeness of outcome data, only 1 article was classified as high risk (with excessive dropout and no specific reasons provided). In terms of selective reporting, 2 articles had inconsistent observation indicators with the original trial registration plan. Other biases were unclear. The bias risk assessment for the included RCTs is shown in Figure 8.

Figure 8.

Results of bias risk assessment in RCT studies.

Quality evaluation of non-RCT and self-preceding control trials. Among the 25 studies, only 3 studies mentioned sample size estimation, and only 1 study mentioned the implementation of blinding in the evaluation of endpoint indicators. Most of the literature had clear research purposes and collected data according to the expected design. For the last 4 items (applicable to 2 non-RCT studies), only one study showed a relatively large gap in some data at the baseline between the groups. The selection of the control group was relatively appropriate and carried out simultaneously with the experimental group, the baseline of the groups was comparable, and an appropriate statistical method was selected. The bias risk evaluation of the self-preceding control studies and non-RCTs included in this study is shown in Figure 9.

Figure 9.

Results of bias risk assessment for self-contrast studies and non-RCT studies

Discussion

The standards for research objects in this field need to be further improved

The report of the basic information of the subjects is not standardized

The basic information in the existing 58 studies mainly reported three aspects: age (55/58 articles), dominant hand (28/58 articles), and the level of education (32/58 articles).

The increase in age is a significant factor contributing to the rising incidence of AD.⁵ This may be related to the formation of amyloid plaques and neurofibrillary tangles, the loss of synapses, neuroinflammation, and changes in mitochondrial function.⁶ The dominant hand affects the macroscopic and microscopic functions and structures of multiple brain regions, and these differences may be related to spatial perception ability, the role of neural junctions, and the influence of daily behaviors on the plasticity of brain white matter development.⁷ The higher the level of education, the slower the decline in cognitive ability.⁸ This may be related to the reduced rate of tau protein accumulation.⁹ The differences in these population characteristics may lead to pathological variations in the brain tissue structure or neuronal functions, resulting in a decline in the quality of the research due to the varying baseline conditions of the subjects. Therefore, this has become a commonly reported issue in the current research in this field.

However, we also found that the relevant studies did not provide a unified and standardized report on other risk factors related to AD. Studies have shown that gender, genetic factors, living habits, cardiovascular risk factors, and bad living habits are all related to an increased risk of AD.^2,10–12- These differences in basic information may cause individual differences in research objects—differences in TCM syndromes or physiques. In the same group of patients, there are differences in symptoms or constitutions among different individuals. This can lead to significant variations in the therapeutic effects of the same acupuncture treatment, thereby reducing the reliability of the research conclusions. In the future, scholars could conduct in-depth mining and review of relevant risk factors for AD and MCI to provide standardized references for limiting population standards in future research in this field, and minimize the impact of potential risk factors of the disease on clinical research.

Insufficient participation of traditional Chinese medicine (TCM) diagnostic criteria

Among the 58 studies, the reference diagnostic criteria for AD or MCI patients were classified into 18 categories, mainly relying on three Western medical diagnostic criteria: NINCDS-ADRDA diagnostic criteria, Diagnostic and Statistical Manual of Mental Disorders, and mild cognitive impairment as a diagnostic entity (37/58, total proportion 63.79%); while the citation of TCM diagnostic criteria was very few (9/58, total proportion 15.52%).

In Western medicine, AD is diagnosed through clinical assessment scales (such as the neuropsychological test MMSE) and clear pathological changes. At the same time, other brain diseases that can cause progressive memory and cognitive dysfunction, as well as systemic conditions known to cause dementia, are clearly excluded. For AD, clear requirements have been set in terms of the objective evidence supporting the pathological changes as well as in the differential diagnosis from similar diseases.

TCM diagnosis collects references to the external “symptoms” of the disease through four diagnostic methods (inspection, auscultation and olfaction, inquiry, and palpation), and then makes a diagnosis of the patient's core “syndrome type”, focusing on syndrome differentiation rather than disease differentiation. In TCM theory, AD or MCI falls under the categories of “forgetfulness syndrome”, “dementia”, and “intellectual decline syndrome”. The core pathogenesis lies in insufficient nourishment of the marrow sea, obstruction of meridians by phlegm and blood stasis, and deficiency of qi and blood. Therefore, from a Chinese medicine perspective, AD or MCI may also include other types of dementia such as vascular dementia, Lewy body dementia, and frontotemporal dementia. It is difficult to make differential diagnosis of AD or MCI from similar diseases based on TCM diagnostic criteria alone. To some extent, this has led to a lower recognition rate of the diagnostic criteria for related traditional Chinese diseases, and their application frequency in related research is also relatively low.

However, acupuncture itself is based on TCM theory and relies on TCM syndrome differentiation to provide prescription ideas for clinical acupoint selection. Corresponding TCM diagnostic criteria are essential in relevant studies of acupuncture treatment (especially in multiple acupuncture treatments). Therefore, in the future, multi-center and large-sample epidemiological studies on AD/MCI based on traditional Chinese medicine syndromes need to be carried out. This will help establish a set of derived TCM syndrome clusters to promote more reliable TCM diagnostic standards in this field, aiming to increase the frequency of combined application of Western and TCM diagnostic standards, and make related clinical research more standardized.

Clinical trial design for acupuncture treatment of AD or MCI from the perspective of neuroimaging

Acupuncture design. A total of 58 studies employed four types of acupuncture methods: manual acupuncture, electroacupuncture, moxibustion, and acupoint injection or acupoint catgut embedding, with multiple acupuncture sessions being the main approach. Acupuncture and electroacupuncture may be widely used in this field due to their universal clinical therapeutic effects and the ability to be quantitatively adjusted during the experiments. The design of acupuncture groups and research directions can be divided into the following categories:

Single-arm trials (23/58): In this type of study, acupuncture was used as the sole intervention method. Neuroimaging techniques were employed to detect the brain changes of the two groups of patients (either AD or MCI patients or healthy individuals) during the immediate acupuncture process or before and after acupuncture. The study explored and verified that acupuncture, by improving the cerebral hemodynamic responses of AD or MCI patients, regulating brain neural activities, enhancing functional connections in brain regions, and modulating brain network activities, may play an effective role in improving cognitive functions of AD or MCI patients.

In addition, Ma Kai's team explored the characteristic differences in electroencephalogram signals between AD patients and normal people under resting and task states after acupuncture intervention.¹³ If the correlation between the differences in electroencephalogram signals and the development stage of AD is further verified, this characteristic difference will become a new method for diagnosing and differentiating AD.

Two-arm trials (35/58): Such studies are mainly divided into three categories: 1. Acupuncture VS sham acupuncture (14/58): Using non-acupoint acupuncture or non-acupuncture operation at the same acupoint to explore the real efficacy of acupuncture; 2. Acupuncture combined with other therapies VS other therapies (8/58), acupuncture VS other therapies/blank control (4/58): using acupuncture combined with Western medicine or rehabilitation therapy to explore the effectiveness of acupuncture combined therapy; 3. Other acupuncture control designs: Differences in efficacy between the same acupuncture method in different populations (Patients with MCI from different ethnic groups or Patients with Alzheimer's disease and those with MCI) (2/58) or differences in efficacy between different acupuncture methods under the same conditions of subjects (1/58).

Multi-arm trials: The number of such studies is small (6/58), see the original Figure 3 for details.

Furthermore, we found that only approximately one-third of the studies were single-session acupuncture experiments aimed at exploring the immediate effects of acupuncture. For multiple acupuncture-related experiments, the majority of the included studies had an acupuncture treatment period of 1-term follow-up on the relevant outcome indicators (only the Lai Ziyan team conducted subsequent follow-up on the relevant scales¹⁴). We believe that this might be due to the limitations of imaging detection sites, personnel, and funds in the follow-up process; on the other hand, the design of a longer trial period would increase the time and financial costs for the subjects, thereby significantly reducing their compliance. Current studies have shown that the subsequent effects still influence brain nerve activities for a considerable period after acupuncture cessation.^15,16 Increasing the observation of long-term follow-up can further enhance the completeness and reliability of the experimental conclusions. Therefore, we hope that future researchers can refer to the diagnostic guidelines for AD and MCI and related studies as much as possible, and add at least one evaluation scale (such as the MESS scale) and imaging detection follow-up part to the relevant experimental design of multiple acupuncture studies to improve the research quality and enhance the reliability of the various conclusions.

Imaging design

In all 58 articles, we divided the neuroimaging task design modes into five categories with reference to the magnetic resonance task imaging design mode: Resting state mode (we also temporarily classified imaging designs with two tests before and after treatment and no reported stimulation), task-state imaging design (NRER design, BLOCK design, and single-BLOCK design).

Resting state mode (33/58) was the most common, used in almost all types of intervention designs. Its detection method: before and after treatment, the patient is instructed to stay awake without any thinking activity for neuroimaging detection. Compared with the other three imaging mode designs: ① High completion rate: The cognitive function decline in AD patients varies greatly from individual to individual, making it difficult to have all patients in the same group complete a unified complex task. The resting-state design does not require active participation and can be adapted to various acupuncture experiment designs. It can also avoid data loss or reduced effectiveness caused by complex task designs, and is more conducive to obtaining reliable data (especially suitable for patients with moderate to severe AD). ② More in line with the natural state: One of the pathological features of AD is abnormal functional connectivity of brain intrinsic networks such as the default mode network.¹⁷ The imaging data obtained under the resting state mode design is close to the pathological state of patients under natural rest, reducing bias caused by subjective psychological factors of subjects.

Task-state imaging designs are divided into three categories: NRER design (10/58), BLOCK design (7/58), and single-BLOCK design (6/58) : BLOCK design can be divided into two different task types—acupuncture stimulation task and working memory task.^18-20 According to statistics, studies using acupuncture as the experimental stimulation task all used MRI detection, and the selection of acupoints mainly focuses on those located on the limbs (such as Neiguan (PC6), Shenmen (HT7)), focusing on the immediate changes in brain function activation under acupuncture stimulation. Studies using working memory task design mostly used fNIRs or MRI detection (in the case of moxibustion intervention). This avoids being restricted by the limitations of detection equipment and allows for the selection of acupoints not only in the limbs but also in the head (such as Baihui (GV20), Fengchi (GB20), and the chest and abdomen (such as Zhongwan (CV12)). Such studies pay more attention to the changes in brain functional networks after acupuncture treatment. NRER design and single-BLOCK design are often used in conjunction with MRI examinations in studies of immediate acupuncture. The former can capture transient responses and reduce adaptation effects, which is suitable for the immediate neural mechanism of acupuncture; the latter is simpler to operate, focuses on sustained stimulation effects, and is more suitable for patients with poor endurance.

Similarities and differences in acupoint selection

Current studies have shown that the core prescription for acupuncture treatment of AD cognitive impairment is Baihui (GV20)-Sishencong (EX-HN1)-Shenting (GV24)-Fengchi (GB20).²¹ The five most commonly used acupoints are Baihui (GV20), Sishencong (EX-HN1), Shenting (GV24), Fengchi (GB20), and Guanyuan (CV4). In relevant studies in the field of neuroimaging, Baihui (GV20) and Taixi (KI3) were used most frequently (27/58, 26/58), and Sishencong (EX-HN1), Zusanli (ST36), Hegu (LI4), and Taichong (LR3) were also in the list of commonly used acupoints.

The most commonly used acupoints in both belong to the Governor Vessel. The Classic of Difficult Issues records that “The Governor Vessel (Du Mai) originates at the point below the coccyx, runs along the interior of the spinal column, ascends to Fengfu (GV16), and enters to connect with the brain.” “All the marrow (sui) is related to the brain.” The brain marrow relies on the Governor Vessel to transport the essence of the kidney upward into the brain. The smooth flow of the Governor Vessel enables the brain marrow to be nourished, and the vitality to be vigorous and the spirit to be prosperous²². Therefore, the Governor Vessel is closely related to brain diseases and mental illnesses. Baihui (GV20) belongs to the Governor Vessel and plays an important role in regulating the qi of the Governor Vessel, so it is most commonly used in acupoint selection for acupuncture treatment. In addition, Baihui (GV20) and other frequently used acupoints such as Sishencong (EX-HN1), Shenting (GV24), and Fengchi (GB20) all belong to local acupoint selection. “Where the acupoint is located, its treatment effect is there.” They are frequently used in clinical research treatment due to their proximal therapeutic effect on brain diseases. Shujun Shao and other studies found that acupuncture at Baihui (GV20) and Sishencong (EX-HN1) can improve cognitive function by enhancing cerebral blood flow;²³ Fangfang Pan and other studies found that the lesions of AD rats were significantly inhibited after scalp acupuncture intervention.²⁴ These studies confirming the efficacy of local acupoint selection in the treatment of AD-related brain mental diseases.

It is worth mentioning that from a neuroimaging perspective, Taixi (KI3) is a frequently selected acupoint, with nearly one-third of related studies using Taixi as a single acupoint for intervention (8/26). The core pathogenesis of AD is mostly kidney essence deficiency and phlegm and blood stasis blocking the orifices. TCM believes that the kidney is the congenital foundation, and the deficiency or excess of kidney essence directly affects the abundance of the brain marrow. Taixi (KI3) is the Yuan-Source Point of the Kidney Meridian. Lingshu (Miraculous Pivot) states: “Each of the Twelve Yuan-Source Points has its specific location; understanding these Yuan-Source Points clearly… Among the yin (organs), the Taiyin (meridian corresponds to) the Kidneys, whose Yuan-Source Point emerges at Taixi (KI3)”, clarifying the role of Taixi (KI3) in regulating kidney storage (especially kidney yin), nourishing yin and tonifying the kidney, strengthening the root and promoting diuresis, and regulating the triple energizer. The high-frequency single-acupoint studies of Taixi (KI3) reflect the active application and exploration of the “treating the root cause for chronic diseases” acupoint mechanism in the field of neuroimaging.

In addition, similar single-acupoint or fixed combination acupoint studies also include Renzhong (GV26), Neiguan (PC6), Shenmen (HT7), Zusanli (ST36), and Siguan (Hegu (LI4) and Taichong (LR3), mostly used in combination and classified as one category here). The frequent occurrence of such studies (23/58) may suggest that acupuncture treatment for AD or MCI mainly focuses on the clinical efficacy of acupuncture point combinations in the macroscopic aspect. In contrast, the related studies from the perspective of neuroimaging tend to emphasize exploring the mechanism of action of single points or fixed combination points. At the same time, due to the limitation of article length, the specific rules of acupoint combination in neuroimaging-related studies of acupuncture treatment of AD or MCI need to be further summarized by other scholars.

Development of neuroimaging technology application in acupuncture treatment of AD or MCI

Classification of neuroimaging technologies. Through statistical analysis, we divided the imaging technologies used in the 58 studies into three categories according to their focus: neuroelectrophysiological technology; blood flow and perfusion imaging technology; structural or functional imaging technology.

Neuroelectrophysiological technology mainly includes: ERP and EEG. ERP focuses on detecting the immediate electrical response time and intensity of the brain to specific stimuli (such as cognitive tasks), focusing on the temporal dynamic changes of neural activities, and evaluating the efficiency and integrity of cognitive processing processes (such as attention, memory, decision-making). It is often used as an electrophysiological marker to identify, diagnose, and predict cognitive impairment,²⁵ including contingent negative variation, event-related potential P300 and so on; EEG focuses on the electrical activity rhythms and patterns of the whole or local brain, reflecting the functional state of brain networks through power ratios of different frequency bands (such as increased θ or β ratio indicating attention deficit) and functional connectivity (synchronization of brain region electrical signals). Electroencephalography can monitor neural activity on the millisecond scale, enabling us to track the rapid and dynamic processes of neuronal groups.²⁶

From the time distribution of the number of studies, research in this field has developed from the early use of ERP alone to the addition of EEG technology in recent years. Neuroelectrophysiological technology has gradually refined and deepened its indicators from the time dimension, providing stronger support for the objective data reflection of brain function activities in this field.

Blood flow and perfusion imaging technology mainly includes: TCD and 3D-ASL. TCD is an ultrasonic imaging technology that performs Doppler detection on intracranial blood vessels through thin areas of the skull (such as the temporal window), focusing on the blood flow velocity and dynamic parameters of major intracranial blood vessels (such as the middle cerebral artery), reflecting the macro blood flow perfusion state. It is widely used in clinical practice due to its non-invasiveness, no radiation, portability, and real-time monitoring.²⁷ 3D-ASL is a new branch of MRI, focusing on the blood perfusion volume of brain tissue (regional cerebral blood flow), reflecting the real-time blood supply of brain regions. In addition, this technology has fewer artifacts and higher signal-to-noise ratio of collected images, so it can clearly evaluate the perfusion of intracranial tissues.^28,29 At present, 3D-ASL has been applied in the diagnosis, condition, and prognosis evaluation of central nervous system diseases (cerebral arteriovenous malformation, epilepsy, stroke, brain tumor, etc.).^30-32

In comparison, TCD pays more attention to the overall supply trend of cerebral blood flow, while 3D-ASL focuses on the micro differences of cerebral blood flow in specific brain regions.

Structural or functional imaging technology mainly includes: MRI, DTI, MRS, and fNIRS. MRI is a traditional radiological imaging technology. With the continuous development of functionality in recent years, it has derived various technical factions such as structural MRI, functional MRI, diffusion imaging technology, and perfusion imaging technology. In this paper, MRI includes ① structural MRI, which can perform high-resolution imaging of brain structure using strong magnetic fields and radiofrequency pulses. Clinically, it is commonly used for the localization diagnosis of brain tissue structural disorders. ② functional MRI, which measures neuronal activity by monitoring changes in blood oxygen levels in the brain, can measure and map brain activity during specific mental or physical tasks. It is widely used in neuroscience research to explore the relationships between brain functional regions and behaviors, perception, emotion, etc. Both DTI and MRS are derived technologies of MRI. The former evaluates the structure and integrity of white matter fiber bundles based on the characteristics of water molecule diffusion movement; the latter analyzes metabolite concentrations by detecting the chemical shift of metabolites in brain tissue, reflecting neuron function and brain metabolic status. Although fNIRS does not belong to MRI, statistical analysis shows that it is often used in combination with MRI (consistent with clinical experience), evaluating brain function activities by detecting changes in oxygenated hemoglobin (Oxy-Hb) concentration to reflect cerebral hemodynamics in brain regions. Compared with traditional MRI, fNIRS does not require large equipment, sensors can be made into head-mounted devices, and subjects can complete detection in natural states (such as sitting, standing, or even slight movement); at the same time, it has low environmental requirements, no need for magnetic field shielding or radiation protection, and simple operation, so it has great potential in future imaging research in this field.

These techniques and their associated indicators provide objective and quantifiable evidence for verifying the effectiveness of acupuncture in treating AD/MCI from the perspectives of function, metabolism, structure, and hemodynamics³³, confirming that acupuncture has the advantage of multi-target effects in treating AD/MCI. However, we also found that most existing studies only use a single neuroimaging technique to detect brain function abnormalities in patients with these two conditions. In fact, the three existing types of neuroimaging techniques each have their own focuses; a multimodal imaging strategy can enable multi-angle analysis of brain function in terms of task processing, tissue structure, and activity intensity. Within the same research group, diverse datasets of the same type or different types can also provide more evidential support for the research results of acupuncture in treating AD/MCI, and even open up new directions for in-depth analysis. We also hope that in the future, more diverse multimodal imaging strategies will be actively applied in research in this field, in order to provide more new references for revealing the changes in brain function caused by acupuncture in the treatment of AD/MCI.

Development characteristics of neuroimaging technologies. Through statistical analysis, neuroimaging technologies in this field have developed in three stages: early stage (2000-2010) : mainly MRI and ERP, focusing on basic structural/electrophysiological analysis; middle stage (2010-2016): EEG, DTI, MRS and other technologies integrated, expanding brain functional connectivity and white matter structure research; recent stage (2016-2024) : the application of subdivided technologies such as TCD, 3D-ASL, and fNIRS increased, reflecting the gradual development of research towards multimodal integration (structure, function, blood flow, metabolism) and multi-technical integration research. It not only relies on classic technologies such as MRI as the fundamental support, but also continuously expands the application of new technologies like 3D-ASL and fNIRS to meet the more complex and diverse exploration needs of neuroscientific research.

However, for acupuncture treatment of AD/MCI, existing neuroimaging still stays in the detection of abnormal brain function in both types of patients, and the changes in deep pathological characteristics need to be explored. Based on the progress in the field of biomarkers, some scholars have proposed the “ATN” diagnostic framework for Alzheimer's disease, namely: three biomarkers A (amyloid), T (phosphorylated tau), and N (neurodegeneration, measured by total tau when applicable).² Using radiomics and biomarker technology, the changes of acupuncture on the pathological characteristics of AD can be intuitively displayed at the cellular and molecular levels. In May 2020, the U.S. Food and Drug Administration approved the tau tracer flortaucipir for clinical use. At present, there are many ongoing studies on biomarkers and PET technology in the AD field. In the subsequent research process, we look forward to integrating biological tracer technology with existing neuroimaging technologies to provide more reliable evidence-based medical evidence for acupuncture treatment of AD or MCI and even more brain diseases.

Differences in imaging focus between AD and MCI patients. We found that in the included studies, there were certain differences in the use of imaging methods according to different types of subjects. Studies on MCI patients used fNIRs and DTI to focus on differences in brain structure and tissue, while studies on AD patients more used TCD, 3D-ASL, and EEG to focus on changes in brain function before and after acupuncture treatment.

MCI is characterized by subjective cognitive decline in patients (including memory, executive function, attention, language, and visuospatial skills), but they can still remain independent in daily life without other obvious dementia symptoms,³⁴ and may be accompanied by a certain degree of brain atrophy (bilateral medial temporal lobe, frontotemporal lobe, and other neocortical regions).³⁵ Because its diagnosis relies on subjective cognition, it is difficult to unify strong objective criteria for confirmation, which to a certain extent leads to the research focus shifting to the brain neuron structure and tissue environment (blood supply) of such patients, providing effective references for the unified diagnosis and preventive treatment of early MCI. As a non-invasive functional detection method, fNIRS can monitor changes in oxygenated and deoxygenated hemoglobin (HbO and HbR) concentrations in the cerebral cortex;³⁶ DTI can study the direction and integrity of WM bundles by measuring the diffusion of water molecules in neural tissue,³⁷ and both can provide reliable objective data support for research in this field, thus being applied.

Compared with MCI, the diagnostic criteria for AD are more objective. Therefore, in the current situation where the pathological mechanism is not yet clear and there are no radical treatment methods, clinical research is more inclined towards improving the brain function of patients. EEG and other methods can reflect the characteristics of brain activity itself. It can help explain the impact of acupuncture on brain activity through different brain wave rhythms triggered by acupuncture,³⁸ providing objective electrophysiological indicators for the assessment of treatment efficacy.

At the same time, studies that incorporate both AD and MCI are relatively rare, and most of these studies did not report the efficacy tracking in the later stage of the trial. This may suggest that the underlying disease progression pathways and the long-term effects on brain tissue structure of acupuncture intervention for both conditions still need to be explored.

Quality evaluation of neuroimaging studies on acupuncture for AD/MCI. In the 58 included studies, the high-risk or moderate-risk items were mainly concentrated in the criteria related to blinding implementation and allocation concealment. Particularly in terms of blinding implementation, among the 33 RCT studies, only 3 were rated as low-risk. For the remaining 25 non-RCT studies and before-after self-controlled studies, only 1 study mentioned the implementation of blinding in the evaluation of endpoint indicators. If blinding and allocation concealment protocols are not implemented, or if the implemented protocols are incorrect, researchers may intentionally or unintentionally select and allocate participants based on their disease conditions.³⁹ This undermines the randomization of the allocation protocol and leads to the generation of bias. In most studies in this field, besides imaging-related indicators, outcome indicators such as cognitive function and activities of daily living are still affected by the subjective behaviors of participants and outcome assessors. Therefore, the standardized design, implementation, and reporting of blinding are essential. It is recommended that in the future, when designing trial protocols for acupuncture in the treatment of AD/MCI, researchers should strictly adhere to the requirements of the Cochrane Risk of Bias criteria and MINORS criteria, standardize protocol design and report writing, ensure the scientificity and rigor of studies, and provide high-quality trial data for clinical practice.

Research advantages and limitations

This study has some limitations. Firstly, there are relatively few clinical trials that meet the inclusion criteria at present, resulting in only 58 studies being included, and the majority of these were conducted in China. On one hand, the small number of included literatures results in a weak data basis for supporting the conclusions of the article. On the other hand, the authors believe that since acupuncture originates from traditional Chinese medicine, compared with international studies, local participants and researchers may have a relatively higher cultural recognition of acupuncture therapy. This may, to a certain extent, lead to potential bias risks in local studies, thereby limiting the generalizability and international representativeness of the research. We hope that more international, multicenter, diverse, and large-sample studies will be conducted in the future to reduce such risks and improve the level of research evidence. Secondly, most studies have issues such as incomplete reporting and inconsistent standards regarding the baseline information of research subjects and diagnostic criteria (detailed above, not repeated here). Additionally, there are high risks of bias in the reporting of blinding implementation and allocation concealment, which reduces the evidence quality of such studies and further affects the overall quality of this research. Finally, this paper points out that the relationship between AD and MCI is very close. However, due to certain reasons, there are corresponding biases and gaps in the application of neuroimaging techniques. Therefore, more research is still needed to enhance the reliability of the results of this study.

Despite these limitations, the current imaging evidence indicates that acupuncture may be a useful alternative therapy for AD and MCI. It shows significant effects in improving the brain function activation and brain function connectivity of patients. This study fills the gap in the existing literature regarding the imaging evidence of acupuncture treatment for AD and MCI, and proposes methodological improvement suggestions for future research. This study enhances the scientific understanding and clinical application of acupuncture in managing AD, providing valuable insights for scholars in this field.

Supplemental Material

sj-docx-1-alz-10.1177_13872877261420235 - Supplemental material for Evidence integration of acupuncture for prevention and treatment of Alzheimer’s disease and mild cognitive impairment from a neuroimaging perspective

Supplemental material, sj-docx-1-alz-10.1177_13872877261420235 for Evidence integration of acupuncture for prevention and treatment of Alzheimer’s disease and mild cognitive impairment from a neuroimaging perspective by Weiguo Zhu, Honghui Li, Kean Wang, Mengzi Sun, Ke Xiang, Shengtao Shan and Chao Ke in Journal of Alzheimer's Disease

Supplemental Material

sj-xlsx-2-alz-10.1177_13872877261420235 - Supplemental material for Evidence integration of acupuncture for prevention and treatment of Alzheimer’s disease and mild cognitive impairment from a neuroimaging perspective

Supplemental material, sj-xlsx-2-alz-10.1177_13872877261420235 for Evidence integration of acupuncture for prevention and treatment of Alzheimer’s disease and mild cognitive impairment from a neuroimaging perspective by Weiguo Zhu, Honghui Li, Kean Wang, Mengzi Sun, Ke Xiang, Shengtao Shan and Chao Ke in Journal of Alzheimer's Disease

Supplemental Material

sj-xlsx-3-alz-10.1177_13872877261420235 - Supplemental material for Evidence integration of acupuncture for prevention and treatment of Alzheimer’s disease and mild cognitive impairment from a neuroimaging perspective

Supplemental material, sj-xlsx-3-alz-10.1177_13872877261420235 for Evidence integration of acupuncture for prevention and treatment of Alzheimer’s disease and mild cognitive impairment from a neuroimaging perspective by Weiguo Zhu, Honghui Li, Kean Wang, Mengzi Sun, Ke Xiang, Shengtao Shan and Chao Ke in Journal of Alzheimer's Disease

Footnotes

Acknowledgements

We thank the Home for Researchers editorial team () for the language editing service.

ORCID iDs

Weiguo Zhu

Chao Ke

Author contribution(s)

Weiguo Zhu: Conceptualization; Data curation; Formal analysis; Methodology; Visualization; Writing – original draft; Writing – review & editing.

Honghui Li: Data curation; Formal analysis; Methodology; Writing – original draft.

Kean Wang: Formal analysis; Investigation; Visualization; Writing – original draft.

Mengzi Sun: Formal analysis; Investigation; Writing – original draft.

Ke Xiang: Formal analysis; Writing – original draft.

Shengtao Shan: Investigation; Writing – original draft.

Chao Ke: Methodology; Project administration; Supervision; Writing – review & editing.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by Youth Program of National Natural Science Foundation of China (82405580), Natural Science Foundation of Hunan Province (2025JJ60638), Natural Science Foundation of Changsha City (kq2403097), the General Program of the Hunan Provincial Health Commission (W20243005), and the Key Program of the School and Hospital with Hunan University of Traditional Chinese Medicine Scientific Research Fund (2023XYLH006).

Declaration of Conflicting Interests

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

Supplemental material

Supplemental material for this article is available online.

References

Lin

Wang

, et al. Chinese guidelines for early prevention of Alzheimer’s disease. Chin J Alzheimers Dis Relat Disord 2024; 168–175.

Scheltens

De Strooper

Kivipelto

, et al. Alzheimer's disease. Lancet 2021; 397: 1577–1590.

Gao

Wang

Meng

, et al. Acupuncture for brain diseases: Conception, application, and exploration. Anat Rec (Hoboken) 2023; 306: 2958–2973.

Xing

Zhou

Xiao

, et al. Research progress of acupuncture for Alzheimer's disease. China J Mod Med 2025; 35: 36–41.

Hao

Zhang

Sun

, et al. Analysis on trend and prediction of mortality of Alzheimer's disease in elderly population in China, 1990-2021. Dis Surveill 2025; 40: 943–948.

Vaiasicca

Balietti

Bevilacqua

, et al.Convergence between brain aging and Alzheimer's disease: Focus on mitochondria. Mech Ageing Dev 2024; 222: 112001.

Yao

Brain structure and function in left- and right-handedness: a multi-modal MRI study. Master's Thesis, Tianjin Medical University, China, 2020.

Luo

Chen

, et al. Parental education, own education, and cognitive function in middle-aged and older adults. JAMA Netw Open 2025; 8: e2513036.

Cai

Fang

Yang

, et al. Higher educational attainment and accelerated tau accumulation in Alzheimer disease. JAMA Neurol 2025; 82: 848–858.

10.

Jia

Exploring the pathogenic mechanism of apolipoprotein E4 in Alzheimer’s disease using induced pluripotent stem cell technology. China Biotechnol, 2025; 1–11.

11.

Behl

Kaur

Sehgal

, et al. The link between Alzheimer's disease and stroke: A detrimental synergism. Ageing Res Rev 2024; 99, 102388.

12.

Dai

Luo

Expert consensus on early diagnosis and treatment of Alzheimer's disease-associated cognitive impairment. Chin J Med Imaging 2025; 33: 337–346.

13.

Liu

Research of amygdaloid nucleus EEG from acupuncture treatment on Alzheimer's disease. Chin J Gerontol 2013; 33: 4108–4111.

14.

Lai

Zhang

Liang

, et al. Efficacy and mechanism of moxibustion treatment on mild cognitive impairment patients: an fMRI study using ALFF. Front Mol Neurosci 2022; 15, 852882.

15.

Liang

Liu

Luo

, et al. Investigation on post-effects of different needling methods on β-endorphin in the hypothalamus in the ratof acute adjuvant arthritis. Chin Acupunct Moxibust 44-46.

16.

Long

Liu

, et al. Resting-state functional MRI evaluation of after-effect of acupuncture at Zusanli point. Chin J Med Imaging Technol 25: 373–376.

17.

Fleisher

, et al. Altered default mode network connectivity in Alzheimer's disease–a resting functional MRI and Bayesian network study. Hum Brain Mapp 2011; 32: 1868–1881.

18.

Rui

, et al. Moxibustion modulates working memory in patients with amnestic mild cognitive impairment: a functional magnetic resonance imaging study. J Tradit Chin Med 2023; 43: 801–808.

19.

Ghafoor

Lee

Hong

, et al. Effects of acupuncture therapy on MCI patients using functional near-infrared spectroscopy. Front Aging Neurosci 2019; 11: 237.

20.

Khan

Ghafoor

Yoo

, et al. Acupuncture enhances brain function in patients with mild cognitive impairment: evidence from a functional-near infrared spectroscopy study. Neural Regen Res 2022; 17: 1850–1856.

21.

Zhang

Yang

Zhang

. Exploring the acupoint selection patterns in acupuncture treatment for cognitive impairment in Alzheimer’s disease based on data mining. J Guangzhou Univ Tradit Chin Med 2025; 42: 1187–1194.

22.

Wang

Jiang

Shi

, et al. Theoretical analysis of “tongdu qishen” acupuncture therapy for psychosis. Mod Tradit Chin Med Materia Medica-World Sci Technol 2020; 22: 2641–2646.

23.

Shao

, et al. Effects of Du meridian electroacpuncture on memory-learning ability and deposition level of hippocampal Aβ in APP/PS1 double transgenic dementia mice. Chin Arch Tradit Chin Med 2020; 38: 101–104.

24.

Pan

Tang

Xiong

, et al. Effect of scalp acupoint penetrating acupuncture on Alzheimer's disease rats based on NF-κB/NLRP3/caspase-1 pathway. Chin Arch Tradit Chin Med 2024; 42: 34–37.

25.

Zhao

Liu

, et al. Event-related potential for cognitive function research: a visual analysis. Chin J Rehabil Theory Pract 2022; 28: 69–78.

26.

Berger

Über das elektrenkephalogramm des menschen. Arch Psychiatr 1929; 87: 527–570.

27.

Jiang

Han

, et al. Diagnostic value of transcranial color doppler ultrasound in ischemic stroke. J Jilin Med Univ 2021; 42: 372–374.

28.

Koolstra

Staring

de Bruin

, et al. Subject-specific optimization of background suppression for arterial spin labeling magnetic resonance imaging using a feedback loop on the scanner. NMR Biomed 2022; 35: e4746.

29.

Wang

Zhao

Wang

, et al. Study on the auxiliary diagnostic value of MRI combined with 3D-ASL perfusion in cerebral infarction and the relationship of that with hemodynamics. China Med Equip 2023; 20: 54–57.

30.

Zhu

Gong

Wang

, et al. Three-dimensional arterial spin labeling-guided dose painting radiotherapy for non-enhancing low-grade gliomas. Jpn J Radiol 2023; 41: 335–346.

31.

Shen

Ren

, et al. The value of three-dimensional arterial spin labeling MRI in predicting the early neurological improvement after intravenous thrombolysis in patients with ischemic stroke. J Clin Radiol 2023; 42: 1076–1080.

32.

Wen

, et al. The study of SWI, 3D-ASL and IVIM in distinguishing high and low grade glioma. Chin J CT MRI 2023; 21: 9–11.

33.

Zhang

Zhao

Wang

Application and considerations of artificial intelligence and neuroimaging in the study of brain effect mechanisms of acupuncture and moxibustion. Zhongguo Zhen Jiu 2025; 45: 428–434.

34.

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 2011; 7: 270–279.

35.

Saykin

Wishart

Rabin

, et al. Older adults with cognitive complaints show brain atrophy similar to that of amnestic MCI. Neurology 2006; 67: 834–842.

36.

Chen

Tang

Chen

, et al. Amplitude of fNIRS resting-state global signal is related to EEG vigilance measures: a simultaneous fNIRS and EEG study. Front Neurosci 2020; 14: 560878.

37.

Sexton

Kalu

Filippini

, et al. A meta-analysis of diffusion tensor imaging in mild cognitive impairment and Alzheimer's disease. Neurobiol Aging 2011; 32: 2322.e5-18.

38.

Wang

Bian

, et al. Multi-scale order recurrence quantification analysis of EEG signals evoked by manual acupuncture in healthy subjects. Cogn Neurodyn 2013; 7: 79–88.

39.

Huang

Xiong

Based on CONSORT + STRICTA electroacupuncture on cervical spondylotic radiculopathy domestic RCT evaluation report quality research. Chin Arch Tradit Chin Med, 35, 1804-1808.

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

1.71 MB

0.03 MB

0.02 MB