Is there a BOLD response of the visual cortex on stimulation of the vision-related acupoint GB 37?


  • Isabel K. Gareus,

    Corresponding author
    1. Department of Radiology, Section of Medical Physics, University Hospital Freiburg, Freiburg, Germany
    2. Institute of Environmental Medicine and Hospital Epidemiology, University Hospital Freiburg, Freiburg, Germany
    • Department of Radiology, Section of Medical Physics, University Hospital Freiburg, Hugstetterstrasse 55, D-79106 Freiburg, Germany
    Search for more papers by this author
  • Michael Lacour MD,

    1. Institute of Environmental Medicine and Hospital Epidemiology, University Hospital Freiburg, Freiburg, Germany
    Search for more papers by this author
  • Anja-Carina Schulte PhD,

    1. Department of Radiology, Section of Medical Physics, University Hospital Freiburg, Freiburg, Germany
    Search for more papers by this author
  • Juergen Hennig PhD

    1. Department of Radiology, Section of Medical Physics, University Hospital Freiburg, Freiburg, Germany
    Search for more papers by this author



To determine whether or not acupuncture of guangming (GB 37) produces a significant response of the visual cortex detectable by means of functional magnetic resonance imaging (fMRI).

Materials and Methods

This study investigates the activation of the visual cortex elicited by a soft and an intensified stimulation of GB 37, an acupoint documented to influence vision-related disorders. Three different paradigms were carried out to detect any possible modulation of the Blood Oxygenation Level Dependent (BOLD)-response in the visual cortex to visual stimulation through acupuncture.


The percentage signal changes in the visual stimulation cycles did not significantly differ before vs. during acupuncture.


Whereas no BOLD-response correlating with acupuncture was detected in the visual cortex, BOLD-signal-changes in response to needle twisting were detected in different cortical areas. Further studies are necessary to clarify whether these clusters correlate to inevitable somatosensory stimulation accompanying acupuncture or represent an acupuncture-specific response. J. Magn. Reson. Imaging 2002;15:227–232. © 2002 Wiley-Liss, Inc.

IN WESTERN COUNTRIES, acupuncture, a therapeutic method originating in prehistoric traditional theory and philosophical concepts of Chinese medicine, has recently gained increasing recognition as an effective complementary method in the therapy of organic diseases. According to traditional Chinese medicine, the insertion of acupuncture needles in characteristic points of the body surface regulates imbalances of the organism through stimulation or soothing of functional organic systems. The acupoints are arranged on so-called “meridians”, which represent a network of channels each connected to a functional organic system. The so-called “Foot-Shao-Yang”, or gallbladder meridian of the limbs, originates from the outer canthus of the eye (GB 1) and, after taking several turns, descends downward on the lateral aspect of the leg to the lateral side of the tip of the fourth toe (GB 44). In respect to this topographic relationship, points of the Foot-Shao-Yang meridian, even though they might be located on the leg, are described as effective acupoints directly influencing disorders related to the eyes (1, 2). Although the nature of the meridians is not well established in general, a functioning peripheral nervous system seems to be essential to induce the so-called “De Qi”-phenomenon, a local tingling sensation, which, according to classic literature, is essential for the effectiveness of acupuncture. The discovery that opioid peptides are released by acupuncture (3, 4) and modern neurophysiological evidence, especially in the field of acupuncture analgesia, suggest that the effect of acupuncture is transmitted through the neuronal system.

The mode of action of acupuncture analgesia has been investigated extensively. Evaluation of recent studies in acupuncture and new findings in the field of neurophysiology have led to the following main theories on the mechanisms of acupuncture (5):

  • 1Acupuncture stimulates small myelinated primary afferents (Aδ or group III) in skin and muscle, which conduct the impulse to the spinal gray matter.
  • 2The activation of enkephalin-containing interneurons in the substantia gelatinosa of the spinal gray matter by acupuncture blocks the conduction of afferent pain signals to the brain.
  • 3Acupuncture produces a generalized neurohormonal mechanism, where β-endorphin and met-enkephalin are released.
  • 4Acupuncture works through activation of two descending pain-control systems of the midbrain, the first one being serotonergic and the second one adrenergic.

Ever since its introduction to Western countries in the mid-1970s, where it was reputed to provide analgesia in surgery, research has tended to focus on analgetic mechanisms of acupuncture, thus leading to a very narrow perspective of acupuncture effects. Therefore, little is known about probable non-analgetic effects of acupuncture. In this field, the antiemetic efficacy of acupuncture, especially of the acupoint neiguan (PC 6), has recently been confirmed in several clinical trials, although the mode of action still remains unclear (6, 7). As non-analgetic effects of acupuncture might also be transmitted through the neuronal system, some authors believe that stimulation of neiguan might influence the median nerve and the area postrema (8).

In this study, we investigated the stimulation of guangming (GB 37), an acupoint of the gallbladder channel, which, as its name “brightness” implies, is described as a very effective acupoint influencing multiple vision-related disorders, such as cataract, night blindness, and optic atrophy (9). Modern neurophysiology could explain this possible functional correlation by junctions to the central nervous system (CNS), e.g., the visual cortex and associated areas. Cho et al recently reported brain activation in the visual cortex through stimulation of acupoints of Foot-Shao-Yang (10). Other studies attempting to characterize the CNS pathway for acupuncture stimulation in the human brain demonstrated the potential of functional magnetic resonance imaging (fMRI) to clarify the CNS mechanisms of acupuncture (11–15).

The purpose of this study was to elucidate the interaction of the cortical response in the visual cortex by acupuncture of the point GB 37. As a hypothesis, we focused on the following questions: 1. Does acupuncture of GB 37 lead to a specific cortical activation, especially in the visual cortex and associated areas, measurable by means of fMRI? 2. Does acupuncture of GB 37 modulate the response of the visual cortex to visual stimulation when applied simultaneously?

In order to clarify those two questions we carried out three paradigms: 1. Visual stimulation and soft acupuncture (seven subjects were studied). 2. Visual stimulation and intensified acupuncture (eight subjects were studied). 3. Intensified acupuncture without any visual stimulation (six subjects were studied).


Imaging Technique

All imaging was performed on a 2 T whole body system (BRUKER S200F Avance) equipped with a head gradient insert (30 mT/m/second). Single-shot gradient-echo planar imaging (GE-EPI) with a matrix size of 64 × 64 and a square field of view (FOV) of 25 cm was used. Twelve slices with a slice thickness of 4 mm and an interslice distance of 6 mm were acquired. Following an initial Rapid Acquisition with Relaxation Enhancement (RARE) (16) scan for positioning of the axial slices, single-shot GE-EPI with an effective echo time (TEeff) of 35 msec and a repetition time (TR) of 3 seconds was performed. Over a period of 32 minutes, 640 time frames were obtained, supplemented by additional T1-weighted axial images as anatomical references.

Subjects and Experimental Protocol

The local medical faculty's ethics committee at the University of Freiburg, Germany approved the experimental protocol described in this study.

In this study, 21 healthy, normal-sighted volunteers (four females, 17 males) between the ages of 24 and 51, with an average age of 36 years, were investigated using an identical acupuncture protocol. All of the volunteers were Europeans (non-Asians), most of them without any previous acupuncture experience. There was no significant difference in respect to the composition of the groups.

The basic experimental design is shown in Figure 1. It was composed of eight repetitions of an alternate block design for visual stimulation characterized by eight rest periods consisting of 40 (off) time frames and eight visual-stimulation periods (S1–S8), consisting of 10 (on), 10 (off), 10 (on), 10 (off) time frames. In the center of repetition period 3, acupuncture was applied, and it was terminated in the center of the visual-stimulation period 7 (S7). Visual stimulation was skipped in period 3 in order to allow observation of any direct activation effect due to needle insertion.

Figure 1.

Arrows symbolize the bilateral needle insertion and removal. The periods of needle twisting are represented by black tick bars. Visual-stimulation-cycles with goggles on and off are shown in black and white (S1–S8). Non-stimulation cycles are colored in light gray. Time frames before and after needle twisting used for cc analysis are shown as dark gray overlaid boxes.

Following an introductory relaxation exercise, acupuncture of GB 37, located in the lateral aspect of the lower leg, was performed perpendicularly, the needle being inserted on the anterior border of the fibula. Two sterile stainless-steel acupuncture needles of 0.22 × 25 mm (1 cun) were used. In group I (N = 7), only soft unilateral acupuncture (left side) without any further needle manipulation was performed. For intensified acupuncture in groups II and III (II: N =8, III: N =6), the needles were inserted bilaterally during the 200th timeframe and removed during the 540th timeframe, the left needle always being handled first. In order to intensify the De Qi-phenomenon in those subjects, a periodic manual twisting manipulation of the needles was performed during acupuncture in intervals of one minute. Each needle was rotated in both directions until the volunteer had experienced a De Qi-sensation or pain that he/she indicated to the acupuncturist with a short finger tap.

For group III (N = 6), the experiment was performed without visual stimulation in order to elucidate any direct activation effects from acupuncture. For visual stimulation, binocular goggles with red LEDs in each eyepiece were used. A single checkerboard reversal with a flashing frequency of 8 Hz was displayed. During the resting state, the subject was instructed to keep his or her eyes closed.

Image Postprocessing and Data-Analysis

For image processing and data analysis, Paravision/Xtip, a Bruker (Karlsruhe, Germany) system software and the statistical parametric mapping software BrainVoyager (R. Goebel, Brain Innovation B.V.) were used. The standard BrainVoyager three-dimensional motion-correction was performed. Due to strong motion artifacts, two data sets of group II were excluded, thereby reducing the number of subjects in group II to six.

For the groups with additional visual stimulation (I and II), functional maps for each visual activation cycle were calculated by correlation analysis with a correlation coefficient (cc) larger than 0.4. Time courses averaged over significantly activated voxels were obtained. The number of activated voxels and the percentage signal change in each visual stimulation cycle averaged over all volunteers were calculated to characterize the activation effects. For direct activation effects due to acupuncture, the data was tested as follows: For the effect of needle insertion, pixel intensities in 10 images before and after needle insertion, as well as 20 images before and after needle insertion, were compared. The effect of needle twisting was examined by comparing four images before vs. four images after needle manipulation during those twisting episodes falling outside of a visual stimulation cycle (Fig. 1).

The functional maps were transformed into the stereotactic coordinate system based on Talairach et al (17), using an affine transformation.

For analysis of the small but distinct direct activation effects of acupuncture, the cc threshold of 0.4 used for the functional maps of each visual stimulation cycle was reduced to 0.15. In order to evaluate the recurrence of activated voxels in single subjects, summary maps were obtained. The maps of the subjects with and without additional visual stimulation were added up separately. The summary maps were spatially gaussian filtered with full width half maximum (FWHM) 5 mm. For visualization, these maps were overlaid onto a single subject brain transformed into talairach space. The frequency of activation within the groups of six subjects was color-coded. A threshold was set to display only voxels activated in at least 50% of the subjects. The center of gravity (given in talairach coordinates) was determined for all clusters of activated voxels, containing voxels that were activated at least in four out of six subjects.


For the groups of subjects with visual stimulation (I and II), successful activation of the visual cortex in each of the activation cycles was demonstrated. Figures 2 and 3 show the percentage signal changes in the visual stimulation cycles (S1–S8). The data demonstrate that BOLD-signal changes of the visual cortex in response to optical stimulation were not significantly modulated by acupuncture. Effects with intensified acupuncture (II) did not significantly differ from effects with soft acupuncture (I). For group III, no direct effects of acupuncture in the visual cortex as described previously (10, 11) could be found.

Figure 2.

Percentage signal change and number of activated pixels in each visual-stimulation cycle (except of cycle S3, when the needle was inserted) averaged over all volunteers. At an individual level, this refers to an averaging of all voxels with a cc > 0.4 within the visual cortex (see Materials And Methods, Image Postprocessing and Data-Analysis). The stated standard deviations represent the interindividual variations of the BOLD-effect.

Figure 3.

Graphical representation of the percentage signal change in each visual-stimulation-cycle averaged over all volunteers. The dotted line represents the results obtained for group II, and the solid line represents the results for group I.

Group analysis for volunteers with additional visual stimulation, as well as for the volunteers without visual stimulation, revealed no statistically significant activation due to insertion of the acupuncture needles.

In response to needle twisting, slight but distinct BOLD-signal changes were observed for both groups with intensified acupuncture (groups II and III). The transformed functional maps of the six volunteers with and without visual stimulation showed 14 (group II) and 10 (group III) clusters of activated voxels with significant activation in at least four out of six subjects in the insular cortex, parietal operculum, parieto-temporal cortex, inferior parietal lobule, superior colliculi, cuneus, middle occipital gyrus, and cingulate gyrus (Table 1). Four of these clusters are presented as an overlay onto a high-resolution MRI of an individual brain (Fig. 4).

Table 1. Talairach Coordinates of Clusters With Significantly Activated Voxels
 BAAcupuncture with visual stimulation (group II)Acupuncture without visual stimulation (group III)
Right hemisphereLeft hemisphereRight hemisphereLeft hemisphere
  1. BA stands for Brodmann area (18). The clusters printed in bold letters are demonstrated in Fig. 4.

Insular cortex1−38−81341−6−3
Parietal operculum405−49−1213443−1212
Parieto-temporal cortex399−53−4371049−57116−50−5613749−6016
Inferior parietal lobule408−50−4531953−2844
Superior colliculi110−33−2110−33−2
Middle occipital gyrus371440−708
Cingulate gyrus104−1733
Figure 4.

BOLD-response in the parieto-temporal cortex of the group without (III) (clusters 6 and 7, left) and with visual stimulation (II) (cluster 10, right). The BOLD-response in the superior colliculi for group II (cluster 11) is shown below. For cluster numbers, see Table 1 (the demonstrated clusters are printed in bold letters).

Figure 5 shows the time course for needle twisting averaged over subjects of group III with activated voxels in clusters 6 and 7, and over all twisting periods. The standard deviation is represented by the dotted lines.

Figure 5.

Time course averaged over subjects of group III, with voxels in clusters 6 and 7 (dotted lines: standard deviation). Since the displayed time course is averaged over all twisting periods, the center (t = 0 seconds) represents the moment of needle twisting.


Acupuncture of the vision-related acupoint GB 37 revealed no significant BOLD-effect of the visual cortex contrary to the observation by Cho et al (10, 11). Apart from not leading to direct activation (group III), acupuncture also did not lead to a modulation of the BOLD-signal change from visual stimulation (groups I and II). The corresponding t-test showed that any modulation of the BOLD-effect (Figs. 2 and 3) larger than 0.2% would have been detected in our study with P < 0.05.

Using an intensified acupuncture stimulus similar to that described previously (12, 13), a significant BOLD-response in different cortical areas could be observed during periodic needle twisting (groups II and III). The areas reproducibly showing BOLD-contrast in both groups of subjects (insular cortex, parietal operculum regarded as secondary somatosensory area, parieto-temporal cortex) have been shown to be involved in the cortical processing of painful and somatosensory stimuli (19). Both stimuli can be expected to occur as side effects of acupuncture. Activation in the anterior cingulate (not within the focus of our study) due to painful stimuli was reported by other authors (20, 21). In our sample, reproducible caudal cingulate activation, as described previously (22), was demonstrated in the group without visual stimulation (group III), but it was also present in two subjects of the visual-stimulated subjects (group II). This is in accordance with previous functional imaging studies on acupuncture (12) and cerebral processing of painful stimuli (21, 23). The physiological significance of clusters 11–14 that occur in the group of subjects with visual stimulation (group II) is unclear (Fig. 4). Hemodynamic crosstalk between the effect of visual stimulation and that of needle twisting is ruled out due to the evaluation design (Fig. 1), which regarded only twisting episodes, which were separated from the visual stimulation periods by more than 15 seconds. The superior colliculus receives retinotopic input from the retina, as well as from the visual cortex, and gives output to structures of the midbrain and brain stem involved in saccadic eye movement generation. Some of the cells cannot only be driven by visual, but also by somatosensory and auditory stimuli. As a structure of visual and sensorymotor integration, the superior colliculus may be a link for the action of acupuncture of GB 37 on diseases of the visual system.

In respect to further studies conducted in this field, we want to emphasize the importance of accurate description of how acupuncture was carried out, which points were needled, and what kind of stimulation was applied. In order to design reproducible experiments, it might be favorable to use electrical stimulation. If the setting does not allow that, it is important to standardize the manual manipulation as far as possible. The intensity of stimulation has to be considered carefully, especially if the experimental setting is rather long where the De Qi-phenomenon might increase and cause volunteers to drop out early. In this respect, we do think that the acupuncture point selection is important. The selection ot the points used should not only be based upon the fact that it is an important and often-used point in the treatment of specific conditions, but also on its location. Some points, e.g., on the dorsal aspect of the foot are simply more painful when needled than others (e.g., on the leg), so that a reproducible stimulation over a certain period of time might not be possible. In respect to pain perception, it might also be of significance whether the volunteers are of Asian descent or have had acupuncture treatment before. We have found it very different to produce a De Qi-phenomenon in non-Asians compared to Asians, and volunteers who had acupuncture experience before can not only differentiate between needle insertion and De Qi-phenomenon much better, but also seem to tolerate acupuncture stimulation better. Thus, it might also be important to obtain certain psycho-physiological parameters such as heart rate, end-tidal CO2, and the type and intensity of sensation experienced by the volunteers during acupuncture. The next step would be to include proper control experiments, such as minimal acupuncture (13) or superficial tactile stimulation (12) in the experiment. The use of acupuncture control points with known influence only on local structures and not on brain function will be helpful in identifying acupuncture-specific cortical activation.


Although the more remarkable findings reported previously (10, 11) could not be reproduced by our study, our results reveal some interesting activation effects that have been previously unobserved. Even though our results are quite encouraging for further elucidation of acupuncture by fMRI, it should be emphasized that the setup of an fMRI-experiment is somewhat alien to acupuncture. The highly artificial environment of the MR scanner and the significant scanner noise are opposed to the tranquility and relaxation that is required for successful therapeutic application of acupuncture. At the current exploratory stage, we find it prudent to pursue whole brain studies in order to derive hypotheses, which will then have to be tested with experimental setups that are designed to more closely resemble that of a therapeutic session. This might necessitate the use of silent fMRI-sequences, which have now become available, although at the cost of reduced time resolution and/or volume coverage. It should be pointed out that our study does not allow one to distinguish between true acupuncture effects and mere somatosensory effects that invariably accompany needle insertion. This requires careful double-blinded experiments, including some sort of placebo acupuncture, which are currently under way in our department.

In our report, we included only results conforming to established statistical tests. Anecdotal findings on single subjects have been omitted, as well as areas of activation below our cc thresholds or not reproducible in at least four out of six subjects. It should be noted that acupuncture is especially prone to produce stimulus correlated to motion effects in areas with strong susceptibility gradients. Especially for observations in pertinent areas, carefully designed control experiments seem to be mandatory. The experimental design used in our study appears to be well suited to deal with at least some of these confounds, since the activation effect caused by visual stimulation can be used as a reference for other acupuncture-related effects. It should be noted that these considerations not only apply to fMRI of acupuncture, but also for many other fMRI studies, which often lack appropriate control experiments and solely rely on the comparison of stimulation episodes with some control condition.