Acupuncture-induced changes in locus coeruleus connectivity and memory improvement in patients with amnestic cognitive impairment

Abstract

Background

An increasing number of neuroimaging studies have consistently indicated that the locus coeruleus is associated with cognitive impairment in the early stages of Alzheimer's disease, and the locus coeruleus plays a critical role in cognition, including memory encoding, consolidation, and retrieval.

Objective

To investigate whether and how acupuncture modulates the functional connectivity patterns of the locus coeruleus, and offer a new perspective on the mechanism through which acupuncture exerts its efficacy.

Methods

Resting-state functional magnetic resonance imaging (fMRI) data were collected from 50 patients with amnestic cognitive impairment (aMCI) before and after verum or sham acupuncture. Seed-based whole-brain functional connectivity (FC) was calculated and compared to explore the changing patterns of the locus coeruleus in aMCI patients following acupuncture.

Results

Increased FCs were observed between the left locus coeruleus and the left inferior parietal lobule, and between the right locus coeruleus and the right posterior cerebellum in aMCI patients after verum acupuncture. Further analyses revealed a correlation between FC of the left locus coeruleus and the left inferior parietal lobule before acupuncture and improvement in immediate recall in aMCI patients.

Conclusions

These results suggest that acupuncture could enhance FC between the locus coeruleus and the inferior parietal lobule/the posterior cerebellum. These functional alterations appear to be linked to the efficacy of acupuncture, particularly in ameliorating memory deficits.

1. Introduction

Mild cognitive impairment (MCI) is considered a pathological state that lies between normal cognitive aging and Alzheimer’s disease (AD). Amnestic mild cognitive impairment (aMCI), characterized by memory deficits, is the main subtype of MCI. Research indicates that aMCI has a higher likelihood of progressing to AD as compared to non-aMCI (Glynn et al., 2021, Jungwirth et al., 2012). Therefore, early intervention in aMCI patients can effectively delay their progression to AD, which is of great significance for them (Lai et al., 2020).

Although the locus coeruleus (LC) is a small neural nucleus located deep within the pons (Berridge and Waterhouse, 2003), it is the primary source of norepinephrine in the brain, projecting widely and extensively throughout the whole brain (Aston-Jones and Waterhouse, 2016). Notably, LC plays a major role in several high-order cognitive functions, such as perception, attention, memory consolidation, and retrieval (Benarroch, 2009, Mather et al., 2015, Sara, 2009, Sterpenich et al., 2006). Previous studies suggested that the LC is the initial site of neurofibrillary tangles in AD(Andrés-Benito et al., 2017; Braak et al., 2011; Stratmann et al., 2015), and hyperphosphorylated tau protein in the LC can be quantitatively observed earliest in the disease progression (Ehrenberg et al., 2017). Structurally, neuromelanin-sensitive MRI showed that a higher contrast-to-noise ratio (LC_CNR) was positively associated with better memory performance in older adults (Dahl et al., n.d, Hämmerer et al., 2018). Conversely, lower LC_CNR is associated with poorer episodic memory and an increased likelihood of aMCI (Elman et al., 2021, Venneri and De Marco, 2020). This may be related to the involvement of the integrity of the LC in neural reserve and cognitive reserve(Robertson, 2013; Wilson et al., 2013). Functionally, MRI studies based on functional connectivity (FC) have provided evidence for the disruption of LC in cognitively impaired older adults. Particularly, FC between the LC and the frontoparietal cortex and cerebellum increases with advancing age (Zhang et al., 2016). Moreover, FCs of LC were reduced in aMCI patients as compared to healthy controls. Even in healthy adults with a parental history of AD, reduced FC between the LC and cerebellum was observed (Del Cerro et al., 2020, Jacobs et al., 2015). Further post hoc analyses revealed that in aMCI patients with higher total learning scores and delayed recall scores, stronger FC between the LC and the parahippocampal gyrus was associated with better memory performance (Jacobs et al., 2015). A 7 T MRI study showed that higher levels of FC between the LC and nucleus basalis of the Meynert or ventral tegmental area were associated with lower memory performance in those over 40 years old (Jacobs et al., 2018). In local brain functional networks, patients with cognitive impairment exhibited specific patterns of disrupted connectivity, characterizing by significantly reduced FC between the LC and the prefrontal cortex, thereby affecting working memory(Zhao et al. n.d.). Additionally, significant degeneration of the LC occurred in the early stages of aMCI(Grudzien et al., 2007; Lyness, 2003; Zarow et al., 2003). Together, these results suggested that the LC is associated with cognitive impairment in the aMCI.

Acupuncture, a commonly used and safe non-pharmacological therapy, has been demonstrated to be potentially effective in improving cognitive function in patients with aMCI (Feng et al., 2012; He et al. n.d.; Lai et al., 2020, Zhang et al., 2022). Functional magnetic resonance imaging (fMRI) provides a method to explore the mechanisms of acupuncture. Resting-state fMRI studies indicated that acupuncture-induced brain alterations were mainly observed in the cingulate cortex, prefrontal cortex, and hippocampus(Yin et al., 2023). Similarly, task fMRI showed that acupuncture mainly modulated functional connectivity of the temporal lobe and the frontal lobe, which were closely related to memory and cognition (Zuo et al., 2012). Other cognitive and memory related brain regions, such as LC, which is the initial site of neurofibrillary tangles in AD, remains largely unknown. Investigating changes in LC-FC pre- and post-acupuncture treatment is vital for comprehending the mechanisms behind acupuncture's memory-enhancing effects in aMCI patients.

In the present study, we aimed to explore whether and how the LC-FC patterns were modulated by acupuncture in aMCI patients. Thus, we hypothesized that: 1) following verum acupuncture intervention, aMCI patients would demonstrate enhanced LC-FC compared to their baseline status; and 2) changes in cognitive function scores would show a significant correlation with LC-FC in aMCI patients.

2. Materials and methods

2.1. Participants

This study was registered on International Traditional Medicine Clinical Trial Registry identifier ITMCTR2025000418).The patients with aMCI were diagnosed based on criteria proposed by Petersen (2004)by two experienced doctors (Jinhuan Zhang and Xingxian Huang). Patients who met the following criteria were selected: (1) self-reported memory issues or informant complaints lasting over 3 months; (2) aged between 50 and 75; (3) Mini-Mental State Examination (MMSE) score between 24 and 28; (4) Montreal Cognitive Assessment (MoCA) score between 19 and 26; (5) Clinical Dementia Rating scale (CDR) score of 0.5; and (6) 17-item Hamilton Depression Rating Scale score < 7. All participants were Han Chinese and right-handed. The exclusion criteria were as follows and patients meeting any of these criteria were excluded: (1) history of intellectual disability or severe neurological disease; (2) evidence of brain injury (e.g., trauma, stroke, and hydrocephalus); (3) systemic diseases affecting cognitive cognition (e.g., epilepsy, Parkinson’s disease, severe anemia, anthrax, HIV, alcohol or drug misuse, syphilis or thyroid dysfunction); (4) inability to complete neuropsychological assessments, such as severe hearing loss, aphasia or visual impairment; (5) history of psychiatric illness or antipsychotic medication use; (6) major neurologic, renal, cardiovascular, or hepatic deficiency; (7) received acupuncture treatment or other cognitive function related treatment within one month; (h) intolerance to acupuncture; (8) localized infection at the acupoint, and (9) any contraindication to MRI.

We initially enrolled 111 patients, and only 54 patients completed all acupuncture treatments, two times of MRI scanning, and two times of clinical assessments. The trail flow and reasons for exclusion are illustrated in Figure S1. The trial was conducted following the principles of the Declaration of Helsinki 1975, revised in 2008. The study was approved by the ethics committees of Shenzhen Hospital of Traditional Chinese Medicine (approval number: K2021- 012–01), and written informed consent was obtained from each participant.

2.2. Interventions

Patients were randomly assigned to either the verum acupuncture group (VA) or the sham acupuncture group (SA). Both VA and SA underwent a 30-minute treatment three times per week for eight consecutive weeks. The manual acupuncture procedures were administered by experienced acupuncturists with an average of 10 years in the field, who received standardized training on the intervention before the study. Acupuncture was performed using single-use, sterile, stainless-steel needles (Hawto, Suzhou Medical Appliance Factory, Suzhou, China; 0.25 × 25 mm, 0.25 × 40 mm).

The acupoint prescription used in this study was standardized based on our previous clinical trials which confirmed its efficacy in improving cognition in aMCI patients(Zhang et al., 2022). In the VA, participants received acupuncture at Sishencong (EX-HN1), Guanyuan (RN4), Qihai (RN6), Baihui (DU20), Shenting (DU24), Yintang (EX-HN3), bilateral Hegu (LI4), Shenmen (HT7), Zusanli (ST36), Fenglong (ST40), Xuanzhong (GB39), Taixi (KI3), and Taichong (LR3). The acupoints were identified following the National Standard of the People’s Republic of China (GB/T 12346–2006), established in 2006 (Fig. 1). After insertion, needles were manually manipulated until the de qi sensation(Yang et al., 2013) was elicited at the acupoints, and each acupoint was manipulated three times during the needle retention period.

Fig. 1.

Location of selected acupoints.

For the SA, non-acupoints were superficially pierced without manipulation to prevent de qi sensation. The non-acupoints were located near actual acupoints to maintain blinding. Detailed information on the locations of non-acupoints can be found in Table S1.

2.3. Clinical measurements

Demographic and disease-related information of the participants including previous medical history and family history were recorded. The primary outcome measure was the MoCA (Beijing version) score, as it has demonstrated high sensitivity and specificity for MCI screening in various countries, including China(Li et al., 2018).

Furthermore, the Auditory Verbal Learning Test-Huashan version (AVLT-H) and the MMSE were the secondary outcomes. Episodic memory impairment is a characteristic clinical feature of aMCI(Dai et al. 2023). Besides, AVLT-H evaluated both short-term delayed free recall and long-term delayed free recall and performed well in assessing the specificity, sensitivity, validity, and reliability of episodic memory in patients with aMCI(Guo et al., 2009; Zhao Q Fau - Guo et al., n.d). The MMSE, similar to the MoCA, is utilized for assessing global cognitive function and is regarded as the standard for cognitive assessment (Malek-Ahmadi et al. 2015).

2.4. MRI data acquisition

To ensure high-quality images, during the pre-scanning preparation, we informed the patients beforehand to keep their eyes closed and remain relaxed, without falling asleep. Simultaneously, rubber earplugs and foam padding were used to reduce scanner noise and minimize head movement.

The fMRI images were acquired using a 3.0 Tesla MRI scanner (Siemens MAGNETOM Prisma). The functional images were collected transversely using an echo planar imaging sequence with the following settings: repetition time (TR) = 2000 ms, echo time (TE) = 30 ms, flip angle (FA) = 90◦, field of view (FOV) = 220 mm × 220 mm, slices = 37, in-plane matrix = 64 × 64, thickness = 3.2 mm, and voxel size = 3.75 mm × 3.75 mm× 4.0 mm. Each functional run consisted of 240 volumes and lasted for 480 s.

2.5. Resting-state fMRI data pre-processing

Data pre-processing was performed using Data Processing Assistant for Resting-State fMRI (DPARSF, http://resting-fmri.sourceforge.net/)(Yan et al., 2016). The first 10 time points were removed for the reduction of the non-equilibrium effects of magnetization. The remaining 230 scans of each participant underwent slice timing, realignment, co-registration with the participants' structural images, and segmentation. Then, the resulting images were normalized spatially with the standard Montreal Neurological Institute (MNI) EPI template in DARTEL and resampled to 3 × 3 × 3 mm³. To balance the need for spatial specificity with the necessity of improving signal-to-noise ratio and accommodating inter-subject anatomical variability in group-level analyses, we utilized a 4-mm Gaussian smoothing kernel to process the generated images. Removing linear and quadratic trends, and regressing out Friston 24 motion parameters(Friston et al., 2011), white matter, and cerebrospinal fluid signals to reduce noises, no global signal regression was performed, followed by filtering with a temporal band-pass of 0.01–0.1 Hz to minimize the effects of low-frequency drift and high-frequency physiological noise (Cordes et al. 2001). Finally, a scrubbing procedure(Power et al., 2014) was performed to eliminate the distance-dependent artifact of head motion according to frame displacement (FD) > 0.5. One image before and two images after the bad image were deleted. To reduce the impact of head motion on FC(Power et al., 2012; Satterthwaite et al., 2012), only data from participants with head movement < 3.0 mm of translation and 3.0° of rotation in each axis were included. Therefore, four of the participants were excluded due to excessive head movement.

2.6. Seed-based functional connectivity

We employed a meta-mask for LC, developed by Dahl et al. (2022)(Figure S2). First, we resampled the seed regions’ masks and all participants’ images into a voxel size of 2 × 2 × 2 mm³ and extracted the mean time series to check whether LCs’ signals were extracted. Second, we applied Pearson's correlation coefficient to estimate functional connectivity between the seed region's averaged time series and remaining brain voxels. Then Fisher's z transformation was performed to convert functional connectivity to the z value. Then, 2 × 2 mixed design repeated ANOVA with a between-subjects factor of the group and a within-subjects factor of treatment was performed using DPABI software to identify the FC with the interaction effect of group × time and the main effect of time. The significance was determined using the Gaussian random field (GRF) theory multiple comparison corrections (voxel p < 0.001, cluster p < 0.05, corresponding to a minimum cluster extent of k ≥ 19 voxels). To elucidate the functional connectivity of the LC with other brain regions more clearly, we conducted separate analyses for the left and right locus coeruleus.

2.7. Regional-based functional connectivity

The signals extracted from the region of interest computed above were analyzed by the general linear model repeated measure analysis using SPSS software package version 26.0 (SPSS Inc, Chicago, IL, USA). The statistically significant level was set at p < 0.05 (Bonferroni corrected for multiple comparisons)

2.8. Correlation analysis

The correlation analysis was used to explore the association between changes in FC and clinical measurements in the aMCI. Changed scores were defined as the scores of patients after VA minus those of patients before VA, Since MMSE and MoCA showed no correlation with FC values, we focused on the correlation between FC and the changed AVLT scores. Correlation was assessed by Spearman correlation analyses while controlling for age, sex, and education. The significant level was set at p < 0.05.

2.9. Statistical analysis

SPSS 26.0 was used for the following statistical analysis. Two-sample t-tests were performed to identify group differences in age and neuropsychological test scores pre and post acupuncture. Due to heteroscedasticity in education, the Mann-Whitney test was used to compare the differences between groups. Sex was analyzed using the chi-square test (χ²) or Fisher’s exact test.

3. Results

3.1. Demographics and clinical measurements

Finally, a total of 26 patients with aMCI in the VA and 24 in the SA were included (Table 1). These patients were well-matched in terms of age, gender, education level, and baseline cognitive function assessments. Both verum and sham acupuncture treatments showed significant positive effects on cognitive functions. However, patients in the VA showed significantly better clinical assessment scores compared to those in the SA after treatments.

Table 1.

Demographics and clinical measurements of aMCI patients.

Subjects	VA (mean ± SD)			SA (mean ± SD)			P value
Sample size	26			24			-
Age (years)	61.19 ± 7.19			61.63 ± 7.60			0.161a
Sex (male/female)	9/17			8/16			0.244b
Education (years)	11.00 ± 3.46			10.83 ± 3.56			0.127c
	Before VA	After VA	P valued	Before SA	After SA	P valuee	P valuea	P valuef
MMSE	26.31 ± 1.52	29 ± 0.98	< 0.001	26.63 ± 1.58	27.83 ± 1.71	0.002	0.473	0.012
MoCA	22.92 ± 1.98	27.08 ± 1.79	< 0.001	22.54 ± 2.62	24.13 ± 2.25	0.017	0.562	< 0.001
AVLT_IR	15.96 ± 3.76	24.15 ± 4.20	< 0.001	15.38 ± 3.84	21.25 ± 4.78	< 0.001	0.781	0.026

Notes: a, two-sample t-test before VA and SA; b, chi-square test of VA and SA; c, Mann-Whitney test of VA and SA; d, paired t-test for before and after VA; e, paired t-test before and after SA; f, two-sample t-test after VA and SA. Abbreviations: SD, Standard deviation; MMSE, Mini-Mental State Examination; MoCA, Montreal Cognitive Assessment; AVLT, auditory speech Learning Test; IR, immediate recall; VA, verum acupuncture; SA, sham acupuncture.

3.2. Results of mixed effect analysis of functional connectivity

We analyzed whole-brain FC using the left and right locus coeruleus as seed points. The whole-brain interaction effects were found between the LC and inferior parietal lobule (IPL) and cerebellum (Fig. 2 and Table S2).

Fig. 2.

Mixed effect results of locus coeruleus functional connectivity. (A) and (B) display the interaction effects of the left locus coeruleus functional connectivity with the remaining brain regions. (C) illustrates the interaction effects of the right locus coeruleus functional connectivity with the remaining brain regions. The above results were obtained using DPABI.

We found that FC between the left locus coeruleus(LC.L) and the left inferior parietal lobule(IPL.L) was significantly increased after verum acupuncture (P = 0.002, corrected), whereas no significant change was found after sham acupuncture. Meanwhile, the FC between the left locus coeruleus and the left cerebellum inferior was increased after sham acupuncture (P = 0.001, corrected), whereas no significant change was found after verum acupuncture (Figs. 3A and B).

Fig. 3.

Changes in LC-FC in aMCI patients before and after acupuncture treatment. The regions and connections were shown in the left and middle column, the mean RSFC for VA and SA before and after acupuncture were shown in the middle column, (*P < 0.05, ** P < 0.001).

For the right locus coeruleus (LC.R), we found that FC between the LC.R and right cerebellum posterior lobe VI (Cere6.R) was decreased after sham acupuncture (P = 0.018, uncorrected) but significantly increased after verum acupuncture (P = 0.004, corrected). The connectivity values between the right locus coeruleus and the right cerebellum superior after verum acupuncture were significantly higher than those after sham acupuncture (Fig. 3C).

3.3. Correlation results

AVLT-IR scores (Fig. 4A) were significantly increased after both VA and SA,with a greater improvement observed in the VA group. Moreover, the FC value of IPL.L-LC.L before acupuncture was negatively correlated with the changed AVLT_IR scores among the aMCI patients of VA (Fig. 4B). No similar correlation was found in the sham acupuncture group.

Fig. 4.

Correlation analysis between FC of IPL.L and LC.L before acupuncture and changed AVLT_IR in the VA. (A) shows the AVLT_IR changes after acupuncture, and (B) shows the correlations between mean FC value before treatment and changed AVLT_IR in the VA. (** P < 0.001).

4. Discussion

Acupuncture is recognized as an effective non-pharmacological therapy for improving cognition in aMCI (Feng et al., 2012, He et al., n.d, He et al., 2021, Lai et al., 2020). Previously, we reported that clinical improvements following acupuncture correlated with functional changes in the left dorsolateral prefrontal cortex(Zhang et al., 2022). In the present study, while both VA and SA significantly enhanced cognitive, VA demonstrated superior efficacy. Given the pivotal role of the LC in early cognitive decline(Tang et al., 2025), and emerging evidence that acupuncture modulates the LC-prefrontal circuit in dementia models (Wan et al., 2025)，we hypothesized that the therapeutic mechanism of acupuncture might involve the restoration of LC function. To test this, we employed seed-based FC analysis in a cohort of 50 patients with aMCI to explore whether and how LC-FC patterns are modulated by acupuncture. The findings indicated that FCs between the LC.L and the IPL.L, as well as between LC.R and the Cere6.R exhibited a significant increase after VA in the aMCI. Additionally, the FC between the LC.L and the left cerebellar Crus II region was tend to increase following SA. Further correlation analyses revealed that the elevated scores on the AVLT-IR were significantly correlated with the pre-acupuncture FC of LC.L and IPL.L.

Previous studies found that alongside the impairment of LC, there is a decline in various cognitive domains, with a particularly notable impact on episodic memory. This disruption also elevates the risk of progressing to aMCI(Bell et al., 2023, Elman et al.). Resting-state FC studies indicated a reduction in FC between the LC and other brain regions among individuals with aMCI(Jacobs et al., 2015). Furthermore, in aMCI patients who have sleep disorders simultaneously, there is an exacerbated reduction in FC between the LC and several key brain regions, including the frontal, temporal, insular, and hippocampal areas(Um et al., 2022). Additionally, parametric empirical Bayesian models revaled reduced effective connectivity within the LC in the aMCI patients(Maturana-Quijada et al., 2025). Recognizing that the degeneration of the LC is a feature of MCI and may play a role in disease progression(Mercan and Heneka, 2022), the LC has emerged as a promising target for therapeutic intervention(Bostanciklioğlu, 2020). A recent meta-analysis of randomized clinical trials, encompassing 1300 patients with cognitive impairment, supported that interventions targeting the LC noradrenergic system can significantly enhance global cognitive function(David et al., 2022). Our findings further demonstrate the efficacy of acupuncture in augmenting memory function in aMCI patients, concurrent with an observed increase in FC between the LC and both the IPL and cerebellum.

In particular, the FC between LC.L and IPL.L was found to increase subsequent to VA. The parietal lobe receives significant innervation from noradrenergic neurons that originate in the LC(Mather et al., 2020), and postmortem studies demonstrated a high concentration of norepinephrine in the frontoparietal regions(Javoy-Agid et al., 1989). In line with these findings, resting-state functional MRI exhibited robust FC of the LC and frontoparietal regions (Jacobs et al., 2018, Zhang et al., 2016). The LC and the IPL.L jointly participate in memory consolidation and retrieval(Sterpenich et al. 2006). Cumulative evidence of in-vivo neuroimaging studies observed a reliable association between the LC and episodic memory, particularly in older individuals(Bell et al., 2023; Ciampa et al., 2022; Dahl et al., 2019; Jacobs et al., 2021). Additionally, functional MRI studies highlighted the critical role of the IPL.L in episodic retrieval processes (Rugg et al., 2002, Smith et al., 2004). Importantly, a series of studies indicated LC significantly correlated with emotional memory, however, regardless of emotional context(Sterpenich et al. 2006). Furthermore, accordance with our results, by activating the LC-noradrenergic system, improvements in free-recall test performance can be observed in patients with aMCI(Segal et al., 2012). Animal studies provided further evidence that these improvements are dependent on LC activity(Lima et al., 2023). On the other hand, LC contributes to the interaction and integration of two attentional systems: dorsal frontoparietal system and ventral frontoparietal system(Sara, 2009). Functional MRI showed that LC may send signals in attentional tasks thereby modulating the response of the frontoparietal system(Raizada, 2008), and it has been observed that FC between the LC and frontoparietal regions is diminished in older adults (Lee et al., 2018). Consequently, acupuncture might enhance the initial recall scores of aMCI patients by bolstering memory retrieval and by facilitating attention during the administration of the AVLT. However, further caution should be exercised in interpreting these results, given the absence of attention test scores that could corroborate our findings.

In addition, we observed an enhancement in FC between the LC.R and Cere6.R following VA. The cerebellum, alongside the prefrontal and parietal cortices, engages in high-level cognitive processes, with cerebellar lobules VI significantly activated during working memory tasks(Stoodley et al., 2012). It is of particular interest that a robust and consistent FC between the LC and the bilateral cerebellum has been observed across various atlases, whether based on the MNI space coordinates, anatomical atlases, the average location of the locus coeruleus from neuromelanin-sensitive MRI, or the individual locus coeruleus cell group maps(Liebe et al., 2020; Song et al., 2017; Wagner et al., 2017; Zhang et al., 2016). The posterior cerebellum is a region that receives direct noradrenergic projections from the LC, which influences the consolidation of cerebellum learning(Cartford et al., 2004), and a reduction in the positive connectivity between the LC and cerebellum was observed after taking Methylphenidate, a drug that supplies an external norepinephrine (Kline et al., 2016). While the connectivity between the LC and cerebellum tends to increase in the healthy older individuals(Zhang et al., 2016), but significantly decreases in the offspring of AD patients compared to cognitively healthy controls, with a positive correlation to AVLT scores (Del Cerro et al., 2020). Similarly, during neutral stimulus memory tasks, aMCI patients exhibit lower FC between LC and cerebellar regions compared to healthy controls (Jacobs et al., 2015). Thus, the observed increase in FC between the LC and the cerebellum as a result of acupuncture may offer a potential therapeutic benefit by mitigating cognitive impairments in individuals with aMCI.

Out of our expectation, we also identified increased FC between LC.L with the left cerebellum Crus II after SA, while no significant changes were observed after VA. Naturally, similar findings were reported in the existing literature. It is plausible that superficial acupuncture administered at non-traditional acupoints could elicit therapeutic responses(Li et al., 2023). A multicenter randomized controlled study conducted by Xu et al. indicated that the beneficial effects of VA over SA did not manifest until weeks 13(Xu et al., 2020). However, our observation period was limited to three weeks, which may not be sufficient for the full expression of VA's distinct therapeutic efficacy. Importantly, cerebellum Crus II located in the posterior cerebellum and showed significantly reduced volume in the aMCI(Kim et al., 2021). The traditional view of the cerebellum is that it plays an important role in motor behavior. However, emerging research revealed cerebellum also participates in non-motor functions such as cognition and emotion(Adamaszek and Kirkby, 2022, Van Overwalle et al., 2020b). Notably, repetitive transcranial magnetic stimulation targeted at the cerebellum crus II has been observed to ameliorate cognitive dysfunction in older adults(Yao et al., 2022). A meta-analysis of approximately 148 functional MRI studies highlighted that the cerebellum crus II is specialized for social mentalizing and self-related emotional cognition(Van Overwalle et al., 2020a).Given the crucial role of the cerebellum crus II in cognitive processes, this partially elucidates why SA also yield therapeutic effects. The efficacy of SA may also to be partially derived from the expectational contributions of the enrolled patients, influenced by the unique cultural context of Guangdong province in China(Sondermann et al., 2021). As a result, participants approached the treatment with a predisposition toward positive outcomes, reflecting the potent interplay between cultural beliefs and perceived health benefits.

The negative correlation between baseline LC.L–IPL.L connectivity and cognitive improvement (changes in AVLT-IR) may reflect a state-dependent effect of acupuncture and could be interpreted in light of the locus coeruleus–norepinephrine (LC-NE) adaptive gain theory(Aston-Jones and Cohen, 2005). In this framework, LC-NE–mediated modulation of neural gain is often described as following an inverted-U relationship with cognitive performance; accordingly, our aMCI cohort might be situated on the right-hand side of the curve, beyond the optimal peak.In early stages of neurodegeneration, including aMCI, converging evidence suggests that the LC system may show maladaptive hyperactivity, sometimes conceptualized as a compensatory response that can also be associated with noisier signaling and altered functional integration(Weinshenker, 2018). From this perspective, the observation that individuals with lower baseline LC–IPL connectivity tended to show larger improvements after VA might indicate that these participants were further from an optimal operating range at baseline and therefore had greater potential for normalization. We further speculate that acupuncture may exert a modulatory, potentially homeostatic influence on LC-related circuitry rather than acting as a purely stimulatory input. If so, VA could contribute to a re-tuning of LC dynamics—possibly attenuating maladaptive hyperactivity and supporting a shift toward a more optimal functional range—thereby yielding larger apparent benefits in those with more pronounced baseline dysconnectivity. Nevertheless, these mechanistic interpretations remain tentative, and future studies incorporating LC-sensitive imaging and physiological indices of noradrenergic function will be necessary to test these hypotheses directly.

However, there are several limitations in our present study. Firstly, constrained by the trial's duration, we were only able to assess the efficacy over a three-week period following the treatment regimen's completion, without the capacity to delineate the long-term effects in the cognitive domain. Secondly, the absence of a corresponding healthy control group impeded us to discern the FC patterns of the LC between aMCI patients and healthy individuals; nonetheless, we have contextualized our findings with the outcomes of prior studies. Thirdly, the sample size in our study was modest, and future research should encompass a larger cohort to bolster the validity of our results.

In conclusion, we utilized resting-state functional MRI to investigate the alterations in FC patterns between the LC and other brain regions before and after acupuncture intervention. Our findings indicate that acupuncture could enhance the FC between the LC both the IPL and the posterior cerebellum in patients with aMCI. This study provides a novel perspective on the mechanisms by which acupuncture improves cognitive function in the aMCI. Furthermore, we identified a correlation between the pre-treatment FC of the LC.L and the IPL.L with the degree of improvement in immediate memory among aMCI patients, suggesting the potential of these FC measures as imaging biomarkers for the prediction of treatment responsiveness in future studies.

Author contributions

Haibo Yu, Nan Yang, and Jinhuan Zhang designed the study. Xingxian Huang, Xinbei Li, and Juan Ou collected the data. Liyu Hu and Yingqi Lu processed the data. Jinhuan Zhang and Ting Liu helped with interpretation of the results clinically. Liyu Hu and Jinhuan Zhang wrote the manuscript. Haibo Yu helped with revising the manuscript. All authors reviewed the paper and approved the final version to be published.

CRediT authorship contribution statement

Xinbei Li: Validation, Supervision, Software, Investigation, Funding acquisition, Formal analysis. Juan Ou: Validation, Resources, Methodology. Xingchen Liu: Validation, Resources, Methodology, Investigation, Formal analysis. Yingqi Lu: Validation, Supervision, Software, Resources, Project administration, Data curation. Ting Liu: Validation, Supervision, Resources, Investigation. Xingxian Huang: Resources, Methodology, Investigation. Jinhuan Zhang: Writing – review & editing, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Nan Yang: Writing – review & editing, Visualization, Validation, Supervision, Software, Methodology, Investigation. Liyu Hu: Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization.

Funding

This work was supported by National Natural Science Foundation of China (82305388), Ningxia Key Research Program (No. 2022BEG03158), Shenzhen Excellent Scientific and Technological Innovation Talent Training Program (RCBS20231211090814025), the Shenzhen Science and Technology Research Program (No.JCYJ20200109114816594), Sanming Project of Medicine in Shenzhen (SZZYSM202311002), Construction Project of Guangdong Famous Traditional Chinese Medicine Inheritance Studio (Document No. 108 [2023] issued by the Guangdong Traditional Chinese Medicine Administration) and Program of the Guangdong Provincial Administration of Traditional Chinese Medicine (No. 20251316,20252035).

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary material

Supplementary material

Data availability

The data are in-house datasets and are available from the corresponding author upon reasonable request.