Article Text

other Versions


Key regions of the cerebral network are altered after electroacupuncture at the Baihui (GV20) and Yintang acupuncture points in healthy volunteers: an analysis based on resting fcMRI
  1. Guifeng Zhang1,2,
  2. Shanshan Qu1,
  3. Yu Zheng1,
  4. Junqi Chen1,
  5. Guizhu Deng1,
  6. Canhong Yang3,
  7. Yong Huang1
  1. 1Traditional Chinese Medicine (TCM) School, Southern Medical University, Guangzhou, Guangdong, China
  2. 2Guangdong Xinxing Chinese Medicine School, Xinxing, Guangdong, China
  3. 3The First School of Clinical Medicine, Southern Medical University, Guangzhou, Guangdong, China
  1. Correspondence to Dr Yong Huang, Traditional Chinese Medicine (TCM) School, Southern Medical University, No 1023, Shatai Road, Guangzhou, Guangdong 510515, China; nanfanglihuang{at}


Objective To identify the key cerebral functional region affected by acupuncture point needling by examining cerebral networks using functional connectivity MRI (fcMRI) and analysing changes in the key regions of these brain networks at different time points after needle removal.

Methods Twelve healthy volunteers received 30 min of electroacupuncture (EA) at the Baihui (GV20) and Yintang acupuncture points and then underwent two fMRI scans, one each at 5 and 15 min after needle removal. Related brain networks were analysed centred at different ‘seeds’, centres which functionally connect the other cerebral regions in an organised network, such as the anterior frontal lobe, anterior cingulate gyrus, parahippocampal gyrus, amygdala, hypothalamus, head of the caudate nucleus and anterior lobe of the cerebellum. Networks were analysed based on the resting cerebral functional connection, and the differences in the activities of the brain networks between the two time points were compared.

Results At 5 min after needle removal, 12 brain functional regions were involved in organising the network centred at the caudate nucleus ‘seed.’ This number was greater than the number of related brain networks centred at the other ‘seeds’. At 15 min after needle removal, 15 and 14 brain functional regions were involved in organised networks centred at the parahippocampal and hypothalamus ‘seeds’, respectively; these numbers were greater than the numbers of other related brain networks centred at the other ‘seeds’.

Conclusions A brain network composed of a large number of cerebral functional regions was found after EA at GV20 and Yintang in healthy volunteers. The key brain ‘seed’ supporting the largest brain network changed between 5 and 15 min after needle removal.

  • Acupuncture
  • Neurophysiology

Statistics from


The relationship between the needling effect and cerebral functional regions has been demonstrated by previous studies.1 ,2 Several cerebral functional regions are activated/deactivated as a result of needling at an acupuncture point. This relationship has not been described as one acupuncture point–one brain region, but rather as one acupuncture point–one brain network. Acupuncture can activate/deactivate a cerebral functional network, indicating that the brain identifies and integrates the needling signal in a manner corresponding to the brain network model.

Investigations into the cerebral integration of the needling signal have primarily been based on resting functional connectivity MRI (fcMRI) analyses of the cerebral network. Hui et al3 ,4 and Fang et al5 published studies on regulatory models, the brain’s default mode and anticorrelated network and the limbic–paralimbic–neocortical network (LPNN).

Studies investigating the needling–brain network have compared needling at a true acupuncture point versus a sham point, needling versus non-needling, needling versus sham needling, needling one acupuncture point versus needling another acupuncture point, and de qi versus non-de qi. Qiu et al6 demonstrated network regulation through an analysis of the cerebral functional connection seen during needling at Taichong (LR3) versus a sham point. Fang et al5 and Hui et al4 investigated specific changes in the LPNN network through comparisons of true needling versus tactile stimulation and de qi versus abnormal pain, respectively. Qin et al7 found a corresponding relationship between needling at Zusanli (ST36) and a certain cerebral network centred at the amygdala ‘seed’; however, they found no effect by non-needling or sham needling. Bai et al8 also noted that needling at ST36 could activate a corresponding cerebral network centred at the ‘seed’ of the temporal lobe; non-needling and sham needling did not produce this effect. Ren et al9 reported that different cerebral networks could be activated as a result of needling at Neiguan (PC6), Daling (PC7) and Guangming (GB37). Recently, Fang et al10 compared the short- and long-distance connections after needling at Guanyuan (CV4) and Zhongwan (CV12) and reported that an increase was primarily seen in the short-distance connection of the cerebral functional network.

The above studies indicate that needling can act on a cerebral network. Needling specifically regulates the cerebral functional network compared with non-needling, sham needling, needling at a sham point and non-de qi. In addition, the affected network differs when different acupuncture points are needled.

Our previous results are in agreement with these studies on the network, and further indicate the presence of a ‘key cerebral region’ that is the ‘core’ of the network, whereas the other cerebral regions involved in the network are peripheral. The acupuncture signal specifically targeted the key cerebral region and then regulated the entire cerebral functional network.11 In this study, we attempted to identify the ‘key cerebral region’.

We have taken account four additional considerations.

  • Clinically, needling is not administered at one acupuncture point but a combination of at least two acupuncture points. Thus, investigation into the relationship between needling and the cerebral network should be modified to investigate needling at a combination of acupuncture points. The combination of GV20 and Yintang is commonly used in clinical practice and was investigated in this experiment. This combination has a positive (primary or complementary) effect on mental disorders with an unclear mechanism, such as depression and Alzheimer's disease.12–15 To reflect clinic practice, a combination of two acupuncture points was used in the experiment.

  • Electroacupuncture (EA) is widely used because the stimulation parameters are easily controlled and stimulation models are correspondingly stable. Hence, it was used in this experiment.

  • Less research has focused on the needling–cerebral network relationship after needle removal than during retention of the needles. The effects of acupuncture can continue after needle removal, the post-acupuncture effect. Therefore, we focused on the post-acupuncture effect, selecting 5 min and 15 min after needle removal as the observation points. We then attempted to identify changes in the key brain region and the brain network as time progressed.

  • A ‘seed’-related analysis is the most widely used approach for resting fcMRI data.16 ,17 Because the subjects were healthy volunteers, it was unclear which cerebral functional area would be affected by EA at GV20 and Yintang; therefore, we analysed the related cerebral functional regions one by one. The seed, which connects the largest number of cerebral functional regions, was regarded as the centre, or key, of the network affected by EA at GV20 and Yintang. The anterior frontal lobe, anterior cingulate gyrus, parahippocampal gyrus, amygdala, hypothalamus, head of the caudate nucleus and anterior lobe of the cerebellum have been linked to psychiatric diseases such as depression and Alzheimer's disease18–20 and were selected as the ‘seeds’ in this experiment; the cerebral network was analysed based on each of the above-mentioned seeds. The ‘seed’ that connected the largest number of cerebral regions and organised the network in this experiment was regarded as the key cerebral region of each time point.

We hypothesised that EA at GV20 and Yintang might target a cerebral functional network centred at a certain cerebral region and that the activated central cerebral region, the key region, was linked to the combination of the acupuncture points used. We also hypothesised that the activated key region, which connects the entire brain network, would change depending on the time of observation.

Data and methods

Volunteer data

Twelve healthy 20–24-year-old right-handed university students (six male and six female) were enrolled in this study. All the subjects had a normal body weight (range 52–71 kg, mean±SD 64.32±6.17) and height (158–178 cm, 167.53±5.63), were not addicted to coffee, cigarettes or alcohol and passed the screening test for under- and over-responders. None of the subjects had received acupuncture for 4 weeks before the experiment. All the subjects voluntarily signed the consent form before the experiment. The trial was approved by the Chinese ethical review committee (ChiERCT-201100 3) and is registered on the clinical trials website (trial registration number: ChiCTR-TNRC-11001311).

Experimental methods

EA: All the subjects received EA at GV20 (the intersection point of the line connecting the tips of the ears and the midline) and Yintang.21 Needles (single-use, 25 mm×0.30 mm, Hwato Br.) were inserted horizontally, and an EA machine (Han's acupuncture point and nerve stimulator, LH-202H) was connected after eliciting de qi. The stimulation was continued for 30 min at a frequency of 2/100 Hz and at a strength meeting the tolerance of the volunteer. All the needling, including the prescreening test, was performed by the same qualified acupuncturist.

Resting-state fMRI scan: A 3.0 T MRI scanner (GE Signa Excite system, USA) and an eight-channel head coil were used. All the subjects received three brain scans: before needling and at 5 and 15 min after needle removal. The resting-state fMRI scan was obtained by blood oxygenation level-dependent scanning with a GR-EPI sequence (TR 3000 ms/TE 20 ms, flip angle 90°, field of view 240 mm×240 mm, layer thickness 5 mm, gap 1.0 mm and image matrix 96×96).22

Data analysis: Statistical parametric mapping (SPM8) software ( was used to process the imaging data. The Montreal Neurological Institute template was used for spatial standardisation. A 4 mm×4 mm×4 mm full-width at half maximum Gaussian kernel was used for the spatially smoothed procession. The data for those subjects whose head moved more than 1.5 mm in three dimensional (3D) translation and/or more than 1.5° in 3D rotation were excluded. Ultimately, the data for two of the subjects were rejected, and 10 volunteers were included in the final statistics.

The anterior frontal lobe, anterior cingulate gyrus, parahippocampal gyrus, amygdala, hypothalamus, head of the caudate nucleus and anterior lobe of the cerebellum were selected as the different seeds (MarsBar, Each cube-like seed with a size of 2 voxels was analysed for functional connections with the other regions of the brain at the voxel level (voxels with a threshold of p<0.001; cluster size ≥10 voxels).23 ,24

The Kruskal–Wallis method in SPSS V.13.0 for Windows was also used.


All the volunteers finished the experiment. The cerebral regions connected by different seeds at 5 min and 15 min after EA at GV20 and Yintang are shown in the online supplementary tables S1 and S2.

The results indicate that the number of cerebral functional regions that were connected to each seed differed (p<0.05) 5 min after removing the needles. The largest number of connected regions was seen in the network centred on the caudate nucleus. Fifteen minutes after removing the needles, the number of cerebral functional regions that were connected to the different seeds was also different (p<0.05). In this case, the largest networks were centred on the parahippocampal region and hypothalamus (table 1).

Table 1

A comparison of the number of clusters present 5 and 15 min after electroacupuncture (EA)


Acupuncture treatment produces both immediate and delayed effects. This delayed effect is stronger, more extensive and longer lasting, can be cumulative and may reflect the basic characteristics of needling therapy.25 In this experiment, we focused on the effects of needling that occur 5 min and 15 min after needle removal.

The anterior frontal lobe, anterior cingulate gyrus, parahippocampal gyrus, amygdala, hypothalamus, head of the caudate nucleus and anterior lobe of the cerebellum have been linked to cognition, consciousness, the expression of affection, emotion processing and mental disorders26–30 and were used as ‘seeds’ to analyse the connecting brain network.

EA at GV20 and Yintang targeted the caudate nucleus, which connected the greatest number of clusters (12) making an organised brain network 5 min after removing the needles. Among the brain areas connected to the caudate nucleus, Brodmann area (BA) 9, 10, 13, 38, 47 and 46 primarily function in connecting to other cerebral regions, while BA 7, 24, 32, 33, 40 and 44 function in spatial orientation, emotion, cognition and language.

The parahippocampal gyrus and hypothalamus were identified as the key cerebral regions activated by EA 15 min after needle removal andwere connected to 15 and 14 clusters in the brain network, respectively. Compared with those of the caudate nucleus’ network, the involved cerebral regions in the latter network express more regulation of function than connection. Among the brain areas involved in this effect, BA 10, 11, 13, 38, 46 and 47 connect these regions to other cerebral regions, BA 4 and 6 function in movement, BA 8 functions in eye movement, BA 1, 2 and 3 function in somatic sensation, BA 18, 20, 21, 22, 37, 41, 42 and 45 function in vision, hearing and language, and BA 5, 7 and 40 function in spatial orientation.

Our study illustrated that the centre of the cerebral network changed from the caudate nucleus to the parahippocampal gyrus and hypothalamus from the fifth to the 15th minute after completing EA at GV20 and Yintang. At 5 min, the network centred on the caudate nucleus primarily participated in establishing a relationship with the other brain functional areas. In contrast, at 15 min the network centred on the parahippocampal gyrus and hypothalamus primarily functioned in somatic movement, sensation, vision, hearing and language. The integration and regulation of the brain that occurs after acupuncture is based on cerebral networks and is relatively focused on key areas. These key areas changed with time after needle removal, indicating that the persistent effect of acupuncture involves the response from different key cerebral regions, and that the possible course of needling involves two phases: setting up the connected network and then functional regulation (figure 1).

Figure 1

The different cerebral networks found to be active 5 and 15 min after electroacupuncture (EA), indicated by lines connecting each seed. The locations in the boxes were active at both time points.

Qin et al7 reported that a certain brain network centred on the amygdala is the basis for the analgesic effect of acupuncture. Hui et al, Dhond et al, Fang et al, Lai et al and Bai et al hypothesised that the regulation of the acupuncture effect is related to default and sensorimotor brain networks and anticorrelated brain networks.2 ,4 ,5 ,8 ,31 We used an acupuncture point combination that is effective in treating mental disorders and analysed the possible networks centred on different seeds. We then compared the ranges of the different networks and considered the network with the largest number of brain regions to be the brain's response to the acupuncture signal; the seed, or the centre of the network, was the key region targeted by acupuncture.

The acupuncture point combination in this experiment is used to treat mental disorders in clinical practice. However, the experiment only included a limited number of healthy volunteers. The time points of 5 and 15 min after EA were not chosen based on previous references but were derived from clinical experience and require further proof that they represent important time points after needling.

In conclusion, this experiment observed a delayed effect of EA at GV20 and Yintang in healthy volunteers and analysed the possible brain networks centred on the different ‘seeds’ using a resting-state cerebral fMRI scan. The results of this study indicate that the brain's response to EA manifests as the development of a cerebral network, with certain key regions being identified from the network. The key cerebral region changed with time after needle removal. This finding may indicate a mechanism for treating mental disorders using EA at GV20 and Yintang.

Summary points

  • In healthy volunteers, we studied the brain networks that were activated by electroacupuncture treatment for mental disorders

  • MRI taken at 5 and 15 min after acupuncture showed a clear evolution in the pattern of activated networks


We thank Professor Yanping Chen, chief physician of the MR Department, First Affiliated Hospital of Southern Medical University, who provided the technical support for the fMRI scans. We also thank the volunteers for their understanding and cooperation.


View Abstract
  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • Contributors YH conceived and coordinated the study. YH, GZ and SQ participated in the design of the study. JC performed the acupuncture stimulation and organised the fMRI scans. YZ, GD and CY participated in the fMRI data analysis. GZ and SQ wrote the manuscript. YH edited the manuscript.

  • Funding The research was supported by the national 973 program of China (No 2006CB504505, No 2012CB518504), the 3rd key construction program of the ‘211 project’ of Guangdong Province, the Guangdong provincial innovation program for undergraduate students (No 1212110042), and the project of the TCM administration bureau of Guangdong Province (No.2010095).

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval Chinese ethical review committee.

  • Provenance and peer review Not commissioned; internally peer reviewed.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.