Brain imaging study on acupuncture:where we are and where we are headed

Jian Kong,Sharon Sun

Psychiatry Department,Massachusetts General Hospital,Harvard Medical School,MA,U S A

哈佛大學醫(yī)學院麻省總醫(yī)院，美國

Technical advances in PET imaging not only provide tools for investigating brain metabolism,blood fl ow changes,and other non-selective markers of neural activity,but also whole brain determinants of specific receptor-binding distributions in fully conscious humans.Such progress enables us to indirectly assess neurotransmitter changes associated with placebo analgesia.For example,it allows us to indirectly measure the release of endogenous opioids in the brain[1-3].

Most importantly,technology even allows us to collect fMRI and PET data simultaneously,which provides a new tool when investigating brain activity and neurotransmitters[2].In addition,there are other brain imaging tools such as high resolution EEG,and MEG,which can provide accurate temporal resolution information.

The literature on brain imaging studies in acupuncture research is rich and extensive.In this paper,we will only focus on some key points critical to acupuncture research,i.e.acupoint specificity,dissociating acupuncture from placebo effects,and using imaging as a biomarker / measurement and predictor for acupuncture treatment.

1 fMRI study on acupoint specifi city

A salient feature of Traditional Chinese Medicine(TCM)is that specific acupuncture points are theorized to have salubrious effects on distant target organ systems.When fMRI studies on acupuncture were beginning,many investigators attempted to use fMRI(typically,a block or event related design)to test the existence(correction)of acupoint specifi city[1,4-22].One typical example of such research is the fMRI study of the vision points,acupoints that can be used to treat eye disorders.

In an early study published in 1998,Cho and colleagues[23]proposed a conceptual relationship whereby acupoint specifi city was mediated via central neural networks that included corresponding brain regions.Using fMRI to test their model,Cho’s group reported that acupuncture manipulation at acupoints on the leg known to benefit visual system disorders(UB 67-UB 60)could produce specific fMRI signal changes within the occipital lobes of the brain,whereas sham non-acupoints(NAP)could not.

Cho’s study received much attention and triggered a rush of experimental replications.Several research groups[9-10,13,24-28]attempted to replicate this work in various ways.Although most drew similar conclusions,their results sometimes varied in fMRI signal change directionality.The work of Gareus et al[10]was a notable exception,reporting in contrast to Cho et al.that acupuncture at a vision-related acupoint(GB 37)could neither directly produce activation in the visual cortex and associated areas,nor modulate fMRI signal changes in the visual cortex evoked by calibrated visual stimulation.In another study[26],we found electroacupuncture stimulation at both vision related acupoints(UB60 and GB37)and the NAP produced modest,comparable fMRI signal decreases in the occipital cortex,including the bilateral cuneus,calcarine fissure,and surrounding areas.There is no significant difference among the three points UB 60,GB 37 and NAP.We speculate that cross modal inhibition,produced by needlingevoked somatosensory stimulation,may account for our finding of BOLD signal decreases in the occipital cortex.More recently,a research group from China[27-28]using a different experimental paradigm,found that the brain’s neural response to acupuncture stimulation at GB37 and a nearby NAP was signifi cantly different.

Despite their continuous investigation of this topic,Cho and colleagues published a retraction of their earlier results[23],reporting that,“there is no point specifi city,at least for pain and analgesia effects,and that we no longer agree with results in our PNAS article”[29].

Rethinking the concept of the vision point could prove interesting.Based on Traditional Chinese Acupuncture theory,the meridians that are connected to the eyes should all be usable in treating eye disorders.In this case,the more than one hundred acupoints on the urinary bladder,gallbladder,small intestine,heart and liver meridians could all be used to treat eye disorders.

In addition,it is worth noting that GB 37(Chinese translation Guangming,brightness)has only been used to treat eye disorders in the past century.A literature search did not find any evidence of GB 37 use for treating eye disorders in ancient classic acupuncture books,including The Spiritual Axis(Ling Su); Systematic Classic of Acupuncture(Zhen jiu jia yi jing,Jin Dynasty 265-420); The Illustrated Classic of Acupuncture Points as Found on the Bronze Model(Tong ren shu xue zhen jiu tu jing,Song Dynasty 960-1279); Classic of Nourishing Life with Acupuncture and Moxibustion(Zhen jiu zi sheng jing,Song Dynasty 960-1279); Gatherings from Eminent Acupuncturist(Zhen jiu ju ying,Ming Dynasty 1368-1644); and Great Compendium of Acupuncture and Moxibustion(Zhen jiu da cheng,Ming Dynasty 1368-1644).This lack of support may imply that using acupoint GB 37 to treat eye disorders is a modern concept that we speculate derived from the Chinese name of the acupoint,brightness.

In addition to the development on the function of acupoints,the brain activity itself is very complex,and often diffi cult to control in experimental settings.For this reason,subtle alterations in the experimental paradigm including instruction to the subjects and their cooperation,as well as details of experimental procedures may all infl uence study fi ndings.It is well known in the field of fMRI that the attention level,general arousal,mood of each individual,and deqi sensation evoked during an acupuncture treatment[30-31]while scanning can each signifi cantly affect the fi nal results of a study.

Accordingly,subtle differences between fMRI signal changes evoked by stimulation at acupoints and NAPs may be masked by psychological conditions such as attention.Taking the case of vision-related acupoints as an example,the subtle fMRI signal difference between vision-related and non visionrelated acupoints / non-acupoints,if the differences exist,may be obscured by other physiological phenomena such as cross-modal interaction.

In summary,it has been more than a decade since Cho and colleagues fi rst attempted to investigate acupuncture point specifi city using fMRI[23].However,validation of the phenomenon through the brain remains undetermined.This suggests that the field faces many challenges due to the complexities of acupuncture,the brain,and fMRI.

2 Powerful placebo effect in acupuncture research and choices of controls for the study

Acupuncture,as defined on the website of the National Center for Complementary and Alternative Medicine,refers to a family of procedures involving the stimulation of specific points on the body using a variety of techniques(http://nccam.nih.gov/health/acupuncture?nav=gsa).

Despite its wide application in Eastern countries for thousands of years,clinical trials evaluating the effi cacy of acupuncture treatment have yielded rather contradictory results due to large placebo effects[32-38].In part,these findings reflect the complexity of identifying the “active ingredients” of acupuncture so as to enable design of an appropriate sham condition[39].

Because acupuncture and placebo may work in part by activating self-healing / regulation mechanisms,it is not surprising that studies find the two share common pathways.Using pain as an example,studies show that both endogenous opioids and cholecystokinin are involved in acupuncture analgesia[40-41]and placebo analgesia[42-43].Under the umbrella of self-healing,the distinction between the two could be blurred too far.A more useful construct holds that acupuncture,which is based on the stimulation of acupuncture points in the human body,should be regarded for “bottom-up” modulatory effects,and the placebo for “top-down” modulatory effects based on previous learning,expectation,therapeutic alliance and other factors common to all placebo treatments.

Brain imaging has the potential to illuminate these contributory mechanisms.We found in previous studies that sham acupuncture combined with positive expectancy can produce a significant placebo analgesia effect[44]and that sham acupuncture combined with negative expectancy can produce significant nocebo hyperalgesia[45].In addition,we also found that enhanced expectancy can signifi cantly enhance the acupuncture analgesia effect whereas diminished positive expectancy appeared to inhibit acupuncture analgesia[46].Although placebo and acupuncture analgesia are comparable in behavioral measurements as indicated by subjective pain rating reduction,the two are associated with different brain networks.More specifically,verum acupuncture primarily involved lower signal intensity changes in brain regions associated with pain intensity signaling[46-47].In contrast,sham acupuncture involved lower activity in brain regions associated with painrelated cognitive-affective signaling.This finding is consistent with a recent study which reported that verum acupuncture,but not sham acupuncture,produced short-term increases in the binding potential of μ-opioid receptors(MOR)in multiple pain and sensory processing regions,and long-term increases in MOR binding potential in some of the same structures[3].

It is worth noting that this large expectancy effect is not something unique to acupuncture.Studies found that expectancy can significantly modulate the analgesic effect of pharmacological drugs as well[48].In a previous study[49],investigators explored how expectancy could change the analgesic effi cacy of a potent opioid in healthy volunteers.Their results showed that positive treatment expectancy substantially enhanced the analgesic benefit of remifentanil and negative treatment expectancy abolished remifentanil analgesia.

In parallel,in light of the contradictory nature of clinical trials on the efficacy of acupuncture treatment,it is important to advance the investigation of the efficacy of acupuncture and the mechanism by which it functions.Double-blinded randomized controlled trials(RCTs)serve as a gold standard when comparing the effect of treatment with the effect of an inert control.However,determining the proper inert control for an RCT designed to evaluate the effi cacy of acupuncture is methodologically challenging due to:(1)Difficulty of mimicking both the visual appearance of the acupuncture treatment device and the method of needle insertion(2)Challenge of controlling for all nonspecific factors involved in acupuncture treatment including the ritual effect(3)Double-blinding the acupuncturist[39].

Placebo controls for acupuncture studies generally fall under one of two categories:(1)sham acupuncture,in which the skin is punctured with real acupuncture needles at nonacupoint locations,shallowly at acupoint locations,or both and(2)placebo acupuncture,which utilizes nonpenetrating stimuli.The latter includes the Streitberger device,Park device,Japanese double blinded device,the foam device and other non-penetrating devices such as toothpicks.Each of these devices has their pros and cons.Please see reference[39]for a detailed description and comparison of these devices.

Taken together,the applications of fMRI can significantly enhance our ability to distinguish between the neurological effects underlying verum and sham acupuncture effects.Choosing an appropriate control is crucial for acupuncture research as there is no single placebo method / device that can be universally applied for all acupuncture studies; the choice of an acupuncture control must therefore be determined by the specifi c aim of the study.

3 Use brain imaging measurements as an objective biomarker

In medical practice,many patients complain of the highly subjective experience,frequently using phrases such as “I am not feeling well” and“I feel a lot of pain on my lower back”.Under most circumstances,the physician / medical practitioner can only make judgments based on the reports of the patients,and there is no objective measurement for these personal experiences.

Using pain experience as an example,although there are more patients who visit clinicians for pain related disorders than almost any other illness,the most commonly used metric for pain measurements is the patient’s subjective report using the visual analogue scale between 0-100.As a result,there exists a demand in both clinical practice and research for identification of the neural correlates associated with pain.Since the subjective experience depends on the central processes of the brain,brain imaging such as fMRI has been applied toward developing a biomarker for pain.

Over the course of the past few decades,investigators have identified a brain network underlying the experience of pain and pain intensity encoding.This network involves brain regions including the primary and secondary somotasensory cortices,insula,and dorsal anterior cingulate cortex[50-52].Wager and colleagues[53]recently developed an fMRI-based measure for pain on an individual level.First,they used machine-learning analyses to identify a pattern of fMRI activity across regions associated with heat-induced pain.They then tested the sensitivity and specificity of the brain response pattern to painful stimuli versus nonpainful warm stimuli in a new sample.Finally,they assessed specifi city relative to social pain and the responsiveness of the measure to the analgesic agent remifentanil.This study demonstrated the feasibility of using fMRI to assess pain elicited by noxious heat stimuli in healthy subjects.

In addition to a task-related fMRI study,another recent application of brain imaging to the investigation of the neural correlates associated with pain is the study of resting state functional connectivity and brain structure[54-56].It is believed that low-frequency components of spontaneous fMRI signals during rest can provide information about the intrinsic functional and anatomical organization of the brain.With these tools,investigators have found significant brain functional connectivity and structural changes in chronic pain patients as compared to matched healthy controls[54-56].Previous studies also suggested that acupuncture can modulate the functional connectivity[57-64]as well as cortical thickness(after longitudinal treatment)[65].

In a more recent study[65],we investigated brain cortical thickness and the functional connectivity changes after acupuncture treatment in knee osteoarthritis patients.We found that after longitudinal treatment,cortical thickness in the left posterior medial prefrontal cortex(pMPFC)decreased significantly in the sham group across treatment sessions as compared with the verum acupuncture group.Resting state functional connectivity analysis using the left pMPFC as a seed showed functional connectivity between the left pMPFC and the key regions of the descending pain modulatory system(rostral anterior cingulate cortex,and periaqueductal gray)[66]are significantly enhanced after verum acupuncture when compared with the sham acupuncture group.Taken together,the fMRI/MRI measurement can be used as marker to evaluate the treatment effect and predict the development of disease / treatment[67-68].

In summary,brain imaging tools have been used extensively in acupuncture research.On the one hand,the application of these tools has enhanced our understanding of acupuncture’s mechanism,but on the other hand,we should be very cautious of our interpretation of the results due to the unique and complicated nature of acupuncture as a medical system,which includes many branches and schools.In particular,studies or clinical trials can only test a specifi c acupuncture treatment protocol for a specifi c population cohort.Over-interpretation of a positive or negative result is not scientifi cally appropriate and can also damage the development of acupuncture research in the long term.We believe future studies should focus on 1)translational research that can utilize the findings from brain imaging studies to enhance acupuncture treatment effects,2)elucidating the association between brain activity/connectivity/structural changes and the clinical outcomes,and 3)examining the causal relationship between the brain and behavioral effects.

[1]Dougherty DD,Kong J,Webb M,et al.A combined(11C)diprenorphine PET study and fMRI study of acupuncture analgesia.Behav Brain Res,2008,193(1):63-68.

[2]Wey HY,Catana C,Hooker JM,et al.Simultaneous fMRI-PET of the opioidergic pain system in human brain.Neuroimage,2014,102(P2):275-282.

[3]Harris RE,Zubieta JK,Scott DJ,et al.Traditional Chinese acupuncture and placebo(sham)acupuncture are differentiated by their effects on mu-opioid receptors(MORs).Neuroimage,2009,47(3):1077-1085.

[4]Wu MT,Hsieh JC,Xiong J,et al.Central nervous pathway for acupuncture stimulation:localization of processing with functional MR imaging of the brain- preliminary experience.Radiology,1999,212(1):133-141.

[5]Hui KK,Liu J,Makris N,et al.Acupuncture modulates the limbic system and subcortical gray structures of the human brain:evidence from fMRI studies in normal subjects.Hum Brain Mapp,2000,9(2):13-25.

[6]Wu MT,Sheen JM,Chuang KH,et al.Neuronal specificity of acupuncture response:a fMRI study with electroacupuncture.Neuroimage,2002,16(6):1028-1037.

[7]Kong J,Ma L,Gollub RL,et al.A pilot study of functional magnetic resonance imaging of the brain during manual and electroacupuncture stimulation of acupuncture point(LI-4 Hegu)in normal subjects reveals differential brain activation between methods.J Altern Complement Med,2002,8(4):411-419.

[8]Li G,Liu HL,Cheung RT,et al.An fMRI study comparing brain activation between word generation and electrical stimulation of language-implicated acupoints.Hum Brain Mapp,2003,18(2):233-238.

[9]Siedentopf CM,Golaszewski SM,Mottaghy FM,et al.Functional magnetic resonance imaging detects activation of the visual association cortex during laser acupuncture of the foot in humans.Neurosci Lett,2002,327(5):53-56.

[10]Gareus IK,Lacour M,Schulte AC,et al.Is there a BOLD response of the visual cortex on stimulation of the vision-related acupoint GB 37?J Magn Reson Imaging,2002,15(4):227-232.

[11]Liu WC,Feldman SC,Cook DB,et al.fMRI study of acupunctureinduced periaqueductal gray activity in humans.Neuroreport,2004,15(8):1937-1340.

[12]Yoo SS,The EK,Blinder RA,et al.Modulation of cerebellar activities by acupuncture stimulation:evidence from fMRI study.Neuroimage,2004,22(2):932-40.

[13]Litscher G,Rachbauer D,Ropele S,et al.Acupuncture using laser needles modulates brain function:first evidence from functional transcranial Doppler sonography and functional magnetic resonance imaging.Lasers Med Sci,2004,19(1):6-11.

[14]Li G,Huang L,Cheung RT,et al.Cortical activations upon stimulation of the sensorimotor-implicated acupoints.Magn Reson Imaging,2004,22(3):639-644.

[15]Napadow V,Makris N,Liu J,et al.Effects of electroacupuncture versus manual acupuncture on the human brain as measured by fMRI.Hum Brain Mapp,2004,24(2):193-205.

[16]Hui KK,Liu J,Marina O,et al.The integrated response of the human cerebro-cerebellar and limbic systems to acupuncture stimulation at ST 36 as evidenced by fMRI.Neuroimage,2005,27(3):479-496.

[17]Yan B,Li K,Xu J,t al.Acupoint-specific fMRI patterns in human brain.Neurosci Lett,2005,383(3):236-240.

[18]Pariente J,White P,Frackowiak RS,et al.Expectancy and belief modulate the neuronal substrates of pain treated by acupuncture.Neuroimage,2005,25(4):1161-1167.

[19]Kong J,Gollub RL,Webb JM,et al.Test-retest study of fMRI signal change evoked by electroacupuncture stimulation.Neuroimage,2007,34(3):1171-1181.

[20]Fang J,Jin Z,Wang Y,et al.The salient characteristics of the central effects of acupuncture needling:Limbic-paralimbic-neocortical network modulation.Hum Brain Mapp,2009,30(4):1196-1206.

[21]Chae Y,Chang DS,Lee SH,et al.Inserting needles into the body:a meta-analysis of brain activity associated with acupuncture needle stimulation.J Pain,2013,14(3):215-222.

[22]Huang W,Pach D,Napadow V,et al.Characterizing acupuncture stimuli using brain imaging with FMRI--a systematic review and metaanalysis of the literature.PLoS One,2012,7(4):e32960.

[23]Cho ZH,Chung SC,Jones JP,et al.New fi ndings of the correlation between acupoints and corresponding brain cortices using functional MRI.Proc Natl Acad Sci U S A,1998,95(5):2670-2673.

[24]Li G,Cheung RT,Ma QY,et al.Visual cortical activations on fMRI upon stimulation of the vision-implicated acupoints.Neuroreport,2003,14(3):669-673.

[25]Parrish TB,Schaeffer A,Catanese M,et al.Functional magnetic resonance imaging of real and sham acupuncture.Noninvasively measuring cortical activation from acupuncture.IEEE Eng Med Biol Mag,2005,24(2):35-40.

[26]Kong J,Kaptchuk TJ,Webb JM,et al.Functional neuroanatomical investigation of vision-related acupuncture point specificity:a multisession fMRI study.Hum Brain Mapp,2009,30(1):38-46.

[27]Li L,Qin W,Bai L,et al.Exploring vision-related acupuncture point specificity with multivoxel pattern analysis.Magn Reson Imaging,2010,28(3):380-387.

[28]Liu J,Nan J,Xiong S,et al.Additional evidence for the sustained effect of acupuncture at the vision-related acupuncture point,GB37.Acupunct Med,2013,31(2):185-194.

[29]Cho ZH,Chung SC,Lee HJ,et al.New findings of the correlation between acupoints and corresponding brain cortices using functional MRI.Proc Natl Acad Sci U S A,2006,103(27):10527.

[30]Kong J,Fufa DT,Gerber AJ,et al.Psychophysical outcomes from a randomized pilot study of manual,electro,and sham acupuncture treatment on experimentally induced thermal pain.J Pain,2005,6(1):55-64.

[31]Kong J.Gollub R,Huang T,et al.Acupuncture de qi,from qualitative history to quantitative measurement.J Altern Complement Med,2007,13(10):1059-1070.

[32]Kaptchuk TJ.Powerful placebo:the dark side of the randomised controlled trial.Lancet,1998,351(9117):1722-1725.

[33]Kaptchuk TJ.The placebo effect in alternative medicine:can the performance of a healing ritual have clinical signifi cance? Ann Intern Med,2002,136(11):817-825.

[34]Kaptchuk TJ.Acupuncture:theory,effi cacy,and practice.Ann Intern Med,2002,136(5):374-383.

[35]Vickers AJ,Cronin AM,Maschino AC,et al.Acupuncture for chronic pain:individual patient data Meta-analysis.Arch Intern Med,2012,172(19):1444-1453

[36]Linde K,Niemann K,Schneider A,et al.How large are the nonspecifi c effects of acupuncture? A Meta-analysis of randomized controlled trials.BMC Med,2010,8(1):75.

[37]Linde K,Niemann K,Meissner K.Are sham acupuncture interventions more effective than(other)placebos? A re-analysis of data from the cochrane review on placebo effects.Forsch Komplementmed,2010,17(5):259-264.

[38]MacPherson H,Maschino AC,Lewith G,et al.Characteristics of acupuncture treatment associated with outcome:an individual patient meta-analysis of 17922 patients with chronic pain in randomised controlled trials.PLoS One,2013,8(10):e77438.

[39]Zhu D,Gao Y,Chang J,et al.Placebo acupuncture devices:considerations for acupuncture research.Evid Based Complement Alternat Med,2013,2013:628907.

[40]Mayer DJ,Prince DD,Rafi i A.Antagonism of acupuncture analgesia in man by the narcotic antagonist naloxone.Brain Research,1977,121(2):368-372.

[41]Han JS.Acupuncture:neuropeptide release produced by electrical stimulation of different frequencies.Trends Neurosci,2003,26(1):17-22.

[42]Levine JD,Gordon NC,Fields HL.The mechanism of placebo analgesia.Lancet,1978,2(8091):654-657.

[43]Benedetti F,Arduino C,Amanzio M.Somatotopic activation of opioid systems by target-directed expectations of analgesia.J Neurosci,1999,19(9):3639-3648.

[44]Kong J,Gollub RL,Rosman IS,et al.Brain activity associated with expectancy-enhanced placebo analgesia as measured by functional magnetic resonance imaging.J Neurosci,2006,26(2):381-388.

[45]Kong J,Gollub RL,Polich G,et al .A functional magnetic resonance imaging study on the neural mechanisms of hyperalgesic nocebo effect.J Neurosci,2008,28(49):13354-13362.

[46]Kong J,Kaptchuk TJ,Polich G,et al.fMRI study on the interaction and dissociation between expectation of pain relief and acupuncture treatment.Neuroimage,2009,47(3):1066-1076.

[47]Kong J,Kaptachuk TJ,Polich G,et al.Expectancy and treatment interactions:A dissociation between acupuncture analgesia and expectancy evoked placebo analgesia.Neuroimage,2009,45(3):940-949.

[48]Colloca L,Lopiano L,Lanotte M,et al.Overt versus covert treatment for pain,anxiety,and Parkinson’s disease.Lancet Neurol,2004,3(11):679-684.

[49]Bingel U,Wanigasekera V,Wiech K,et al.The effect of treatment expectation on drug efficacy:Imaging the analgesic benefit of the opioid remifentanil.Sci Transl Med,2011,3(70):70ra14.

[50]Tracey I.Nociceptive processing in the human brain.Curr Opin Neurobiol,2005,15(4):478-487.

[51]Tracey I,Mantyh PW.The cerebral signature for pain perception and its modulation.Neuron,2007,55(3):377-391.

[52]Kong J,Loggia ML,Zyloney C,et al.Exploring the brain in pain:activations,deactivations and their relation.Pain,2010,148(2):257-267.

[53]Wager TD,Atlas LY,Lindquist MA,et al.An fMRI-based neurologic signature of physical pain.N Engl J Med,2013,368(15):1388-1397.

[54]Davis KD,Moayedi M.Central mechanisms of pain revealed through functional and structural MRI.J Neuroimmune Pharmacol,2013,8(3):518-534.

[55]Apkarian AV,Baliki MN,Geha PY,et al.Towards a theory of chronic pain.Prog Neurobiol,2009,87(2):81-97.

[56]Apkarian AV,Bushnell MC,Treede RD,et al.Human brain mechanisms of pain perception and regulation in health and disease.Eur J Pain,2005,9(4):463-484.

[57]Zyloney CE,Jensen K,Polich G,et al.Imaging the functional connectivity of the periaqueductal gray during genuine and sham electroacupuncture treatment.Mol Pain,2010,6:80.

[58]Dhond RP,Yeh C,Park K,et al.Acupuncture modulates resting state connectivity in default and sensorimotor brain networks.Pain,2008,136(3):407-418.

[59]Liu P,Qin W,Zhang Y,et al.Combining spatial and temporal information to explore function-guide action of acupuncture using fMRI.J Magn Reson Imaging,2009,30(1):41-46.

[60]Bai L,Qin W,Tian J,et al.Acupuncture modulates spontaneous activities in the anticorrelated resting brain networks.Brain Res,2009,1279(1):37-49.

[61]Qin W,Tian J,Bai L,et al.fMRI connectivity analysis of acupuncture effects on an amygdala-associated brain network.Mol Pain,2008,4(1):55.

[62]Liu B,Chen J,Wang J,et al.Altered small-world effi ciency of brain functional networks in acupuncture at ST36:a functional MRI study.PLoS One,2012,7(6):e39342.

[63]Sun R,Yang Y,Li Z,et al.Connectomics:a new direction in research to understand the mechanism of acupuncture.Evid Based Complement Alternat Med,2014,2014:568429.

[64]Zhang Y,Qin W,Liu P,et al.An fMRI study of acupuncture using independent component analysis.Neurosci Lett,2009,449(1):6-9.

[65]Chen X,Spaeth RB,Retzepi K,et al.Acupuncture modulates cortical thickness and functional connectivity in knee osteoarthritis patients.Scientifi c Reports,2014,4:6482.

[66]Kong J,Tu PC,Zyloney C,et al.Intrinsic functional connectivity of the periaqueductal gray,a resting fMRI study.Behav Brain Res,2010,211(2):215-219.

[67]Baliki MN,Petre B,Torbey S,et al.Corticostriatal functional connectivity predicts transition to chronic back pain.Nat Neurosci,2012,15(8):1117-1119.

[68]Hashmi JA,Kong J,Spaeth R,et al.Functional network architecture predicts psychologically mediated analgesia related to treatment in chronic knee pain patients.J Neurosci,2014,34(11):3924-3936.