Alterations in electrical pain thresholds by use of acupuncture-like transcutaneous electrical nerve stimulation in pain-free subjects.

This study was designed to investigate the effect of acupuncture-like transcutaneous electrical nerve stimulation (ALTENS) on pain thresholds in the index fingers of healthy human subjects. Acupuncture-like transcutaneous electrical nerve stimulation utilizes a high intensity, low frequency stimulus delivered via surface electrodes to acupuncture points. Rates between 1 and 10 Hz have been used in ALTENS and electroacupuncture to mirror rates of stimulation produced by manual twirling of acupuncture needles.[1-3] In order to simulate acupuncture, the stimulus intensity must be sufficient to elicit strong muscle contractions just below the level of frank pain.[2,4] Endogenous opiate release has been implicated in the analgesia produced by ALTENS and electroacupuncture.[5-9]

Aside from ALTENS, there are four other categories of transcutaneous electrical nerve stimulation (TENS). The most common type is "conventional TENS," which is based on the gate control theory.[10] Low intensity, high frequency (50-100 Hz) stimulation is applied to a major nerve, thus activating A[keyable] neurons, which "gate" ascending pain transmission in the spinal cord.[7,11] In another category, termed "brief, intense TENS," C fibers are recruited using very high intensity (painful), high frequency stimulation for brief periods of time, probably acting via diffuse noxious inhibitory controls (DNIC).[12] In addition, a category called "pulse-trains TENS" uses high frequency bursts at low frequency intervals.[13] The remaining category, called "modulated TENS," uses variable pulse widths and heights to overcome habituation of the nervous system.[13] Recently, an ALTENS device has been developed to overcome habituation by a new approach (see "Method" section).

Unlike the analgesia produced by ALTENS, which has a delayed onset and can last for hours or days after cessation of stimulation,[2,7,11] the analgesia observed with both conventional TENS and brief, intense TENS is usually coincident with the onset of stimulation and usually lasts only the duration of the treatment.[11,14] Unfortunately, many researchers[15-18] have used conventional TENS approaches (eg, low intensity, high frequency stimulation applied to a mixed nerve) while looking for ALTENS outcomes (eg, prolonged pain relief, naloxone antagonism, nonsegmental effects) with less than favorable results.

Some TENS research can be questioned because of the lack of proper controls. Some investigators[19,20] have failed to include a placebo treatment in their experimental designs, and some researchers[21] have used bed rest as a control therapy. The most viable control to be used on subjects who are receiving ALTENS treatment is one that utilizes the same experimental setup while delivering only sufficient current to elicit a faint tapping (ie, light touch) sensation. This level of stimulation has been shown not to produce analgesia, yet it can provide a convincing placebo effect.[4,22]

One unforeseen outcome of TENS treatments reported in the literature has been a correlation between initial pain threshold and change in pain threshold produced by TENS. Golding and coworkers[15] reported that those subjects with high pretreatment electrical pain thresholds in their index fingers experienced a decrease in pain threshold during one-half hour of conventional TENS applied to the median nerve. Conversely, those with low initial pain thresholds reported a rise in their pain threshold (analgesia) during TENS. When the results were pooled, no change in pain thresholds was observed. Other researchers[19,23] had previously noted the same correlation between initial electrical pain thresholds and change in pain thresholds.

This study was designed to address these three issues facing TENS research. By comparing very weak (ie, placebo) stimulation with strong ALTENS stimulation, the overall efficacy of ALTENS can be evaluated. By monitoring the time course of change in acute pain thresholds before, during, and after ALTENS, the appearance of a "gatelike," "DNIC-like," or "opiatelike" time course can be determined. Finally, by correlating initial pain thresholds to change in pain thresholds, individual differences can be accounted for and the potential benefits of treatment can be assessed.

Method

Subjects

Ten pain-free subjects (2 women, 8 men), ranging in age from 21 to 30 years, participated in the study. None of the subjects had any history of neurological dysfunction, and none were taking medications. All were pain-free prior to the study. The procedures were explained in writing, and each subject signed an ethics committee-approved consent form and was free to withdraw from the study at any time. Subjects were compensated a nominal $5 (Canadian) per session.

Transcutaneous Electrical Nerve Stimulation Characteristics

Either ALTENS or placebo TENS was delivered to the six acupuncture points on the hands[7]: LI4 (first dorsal interosseous muscle), TW3 (third dorsal interosseous muscle), and Ht7 (ulnar nerve at the wrist), bilaterally, with the ground electrode on the left heel (Fig. 1). We used self-adhesive pad electrodes (Codetron Pads(*1)) that had a carbon-impregnated base and a surface of conductive karaya gel and that had been cut into 2-cm circles. Stimulation was provided by the Codetron model 05(*1)--a constant-current device designed to produce a randomized switching of output among the six electrode pads. There was a 10-second "on" time and a 1-second pause before rotation to another pad. This type of stimulus rotation has been shown to keep cortical arousal high and to overcome habituation attributable to repetitive stimulation.[24] The frequency of stimulation during each 10-second period was set at 4 Hz, each biphasic square wave pulse being 1 millisecond in duration. For ALTENS treatment, the intensity was increased until it just became painful and then decreased slightly just below pain threshold so that a dull, achy, radiating sensation was produced (known, in Chinese, as "de qi") and strong muscle contractions were elicited. Subjects were told that the sensation may be similar to the ache produced when a "tight" muscle is pressed on, but it should not be painful. A number of researchers[24,25] have reported the importance of de qi. Yet rarely do investigators report on the quality of the sensation they elicit with ALTENS. A sharp or superficial pain may indicate A[unkeyable] fiber activation from the skin, whereas type II and III muscle afferent activation has been reported to be necessary for optimal production of acupuncture analgesia.[7,26,27]

For placebo TENS, the intensity was set just above sensation threshold so that a very faint, but distinct, tapping was perceived. This level of stimulation has previously been shown to neither produce analgesia nor release endorphins and has been used by other investigators as a placebo.[4,22] Aside from intensity level, all other stimulation characteristics were the same for placebo TENS and ALTENS.

Volunteers were informed that the purpose of the study was to compare the analgesic effects of two forms of TENS already in use. When questioned after the study, none of the subjects were aware that the lower intensity was being utilized as a placebo control.

Sensory Threshold Stimulation

Sensory thresholds were determined with constant-current electrical stimulation delivered to the pad of each index finger. The current was generated by a Nicolet Pathfinder I[unkeyable] and was delivered via 2X4-cm self-adhesive pad electrodes (described previously) secured with tape, with the negative lead on the distal phalanx and positive lead on the proximal phalanx (Fig. 1). The machine was set to deliver a maximum of 15 mA, which was controlled by a hand-held diode with a sliding scale. The current was increased in 0.75-mA steps. The frequency of stimulation was 2 Hz, with each pulse lasting 250 microseconds.

First detectable sensation thresholds as well as pain thresholds were determined via the methods-of-limits technique. The current was increased from zero at a constant rate (approximately 1.5 mA/s) until a light tapping (sensation threshold) was perceived. The stimulus was then immediately turned off, and the procedure was repeated three times. The average of the three current levels was taken as the true sensory threshold for that finger. For pain thresholds, the same process was repeated, but the intensity was increased until the first pain (pain threshold) was perceived.

Procedure

Each subject participated in three sessions. The first was an acclimatizing session in which the electrodes were attached to the subject and the subject was familiarized with the procedure and the sensations produced by the stimulation. In the second session, the subject was randomly given either ALTENS or placebo TENS. In the third session, the remaining stimulation type was delivered in a cross-over design. The two experimental sessions were spaced at least 48 hours apart to prevent any carryover effect of the treatments. In addition, to control for circadian rhythm in pain thresholds,[28] the experiments were performed at the same time of day for each subject.

All leads were affixed and the subject placed in a comfortable reclining position before testing began. Sensory thresholds were determined every 15 minutes--twice before, once during, and three times after treatment (t=-15, 0, +15, +30, +45, and +60, respectively). Thresholds were determined on the left side first for each subject to ensure consistency. The ALTENS or placebo TENS treatment began immediately after the second entire set of threshold tests were completed, and the treatment lasted 30 minutes. For the sensory reading taken during placebo TENS or ALTENS, threshold measurements were taken only while the contralateral hand was actively stimulated. Immediately upon cessation of ALTENS or placebo TENS treatments, thresholds were again determined. Subsequent threshold measurements were taken 15 minutes and 30 minutes posttreatment, between which the subjects rested comfortably.

Data Analysis

The vast majority of studies do not give criteria for what level of analgesia achieved will be deemed useful or successful. Thus, it is difficult to compare or design studies based on the literature. An exception to this trend has been the naloxone studies.[5,29] Researchers have defined what they call a "useful" level of analgesia in order to then test naloxone reversal. For acute pain, two different groups used a 20% rise in current needed to elicit the sensation of pain as the minimum acceptable change to be called a successful analgesia.[5,29] Because we also chose to use a model of acute pain, we decided to adopt the 20% guideline.

For each subject, pain thresholds (in milliamperes) for the left and right index fingers were monitored before, during, and after treatment. The initial two readings were averaged, and this average was taken as the baseline value for that hand (Tab. 1). All readings were then converted to a percentage of change from the baseline value.

The percentage of change in pain thresholds during placebo TENS was then subtracted from the percentage of change in pain threshold during ALTENS for each hand. The effects on the left and right hands of each subject were averaged so that each subject had one value for each point in time during the experiment. This represented the efficacy of ALTENS over placebo TENS at altering acute pain thresholds. Overall group mean changes in pain thresholds were calculated along with the standard deviation and 95% confidence interval (CI) for each point in time (Tab. 2).

Sample-size calculations were performed using the standard deviation of our data to ensure accuracy and protect against a type II statistical error (ie, saying there is no change when there might have been one that was missed because of a small sample size). In Table 2, the calculation for n is the number of subjects needed to ensure detection of a 20% change in pain threshold, if one existed. We would need six subjects (ie, 5.7) to detect a 20% change during treatment (t=+15), as the standard error (SE) was 14.69, and only three subjects (ie, 2.5) to detect a 20% change immediately after treatment (t=+30), as the SE was 9.83. Thus, in our experiment, we had more than enough subjects (n=10) to detect a 20% analgesia, should one exist.

Results

Figure 2 illustrates the group overall effect of ALTENS. Each data point represents the group mean change in pain threshold from baseline when placebo TENS has been factored out. The two pretreatment readings used to calculate the baseline values are included to show the variability of the baseline data. The bars indicate a CI of 95%. This graph reveals that ALTENS did not produce a significant change in acute pain thresholds in the fingertips at any time during or after treatment. None of the points reach the 20% change needed to be deemed successful analgesia. The largest change was 4.2%, which occurred at +30 minutes, immediately after treatment.

To determine whether the placebo TENS and the ALTENS treatments both had an effect, but the difference between the two treatments was small, the group mean effects were calculated for the ALTENS and placebo TENS treatments separately. The percentage of change in pain thresholds during the experiment for the ALTENS and placebo TENS treatments is shown in Figure 3. Data points show group mean changes, and bars show the CIs. From these graphs, it is clear that neither ALTENS alone nor placebo TENS alone produced a satisfactory analgesia at any point in time because the data points and the CIs never reach above 20% change.

Individual subjects' responses to the ALTENS and placebo TENS treatments (Tab. 3) indicated that very few subjects reached satisfactory analgesia at any point in time. Out of 40 possible data points (4 readings for each of 10 subjects), the ALTENS treatment produced important analgesia at only 4 points and the placebo TENS treatment produced analgesia at only 2 points. Thus, there is no "hidden" analgesia masked by the group data, as individuals rarely obtained analgesia. In order to ensure that readings from the dominant and nondominant hands could legitimately be pooled for each subject, the two hands were treated separately and there was no effect of handedness.

As previous authors had reported a relationship between initial pain threshold and TENS analgesia, it was necessary to determine whether an effect was masked by pooling data from subjects with low and high initial pain thresholds. In order to explore this relationship, the variables were plotted in Figure 4. This figure compares the initial pain threshold (in milliamperes) with the percentage of change in pain threshold produced by ALTENS and placebo TENS at the first posttreatment measurement (t=+30). This point in time was chosen because acupuncture analgesia should be peaking.[2,7] The correlation was nonsignificant (ALTENS, R=.288, P>.05; placebo TENS, R=.314, P>.05). Other authors used absolute changes in pain threshold to study this relationship. When we calculated this correlation, still no effect was observed either immediately after treatment (t=+30) or during treatment (t=+15). We chose t=+15 because it was in the middle of the treatment. Therefore, the level of analgesia produced does not appear to be related to the subject's initial pain threshold. In addition to pain thresholds, sensation thresholds did not change significantly for any of the stimulation characteristics tested.

[TABULAR DATA OMITTED]
[TABULAR DATA OMITTED]

Discussion

Under these experimental conditions, ALTENS did not alter electrical pain thresholds in the fingertips when compared with placebo TENS. When analyzed alone, neither ALTENS nor placebo TENS produced analgesia. There was no significant difference attributable to handedness of the subjects. Finally, no correlation was found between the initial pain threshold and the change in pain threshold of each subject.

Our findings on ALTENS' failure to shift acute electrical pain thresholds is in agreement with the findings of previous researchers who also attempted to induce finger analgesia[18,30] and researchers who attempted to induce wrist analgesia.[31,32] Our findings, however, are in contrast to the findings of one study that showed a significant increase in electrical pain threshold in the wrists of healthy subjects after ALTENS.[33] The results of that study may be questioned because the placebo treatment (ie, bed rest) was inadequate, the duration of treatment (ie, 4 minutes) was extremely low, only one baseline sensory threshold measurement was taken, no acclimatization session was given, and the study did not use a cross-over design. Hence, the changes seen may have been due to alterations in the subjects' attention, expectation, and arousal.[28]

There are several possible explanations for our finding of the absence of analgesia. First, the site of pain testing may have been inappropriate for an acute pain analgesia model. Chapman and colleagues[34] have commented that it is difficult to produce acupuncture analgesia in the periphery (eg, arms and legs), whereas good analgesia is obtained close to the midline, for example, with tooth pulp pain. Chapman and colleagues reported that in China, where electroacupuncture was used, "acupuncture analgesia is most often successful for surgery of the head and trunk while effects are less reliable on the extremities."[34]

A second possible explanation for the lack of analgesia is that the type of pain stimulation used in this study (electrically evoked pain) may be difficult to block with acupuncture or ALTENS.[32] It has been argued that cutaneous electrical stimulation produces an unnatural synchronous volley of information to the central nervous system, whereas natural pain stimulation (eg, pinch) activates skin receptors in a spatial and temporal distribution that is asynchronous.[32] The perception of this naturally induced pain can be altered by changing slightly a single component of the afferent barrage (ie, changing the pain pattern).[32] Moreover, asynchronous inputs do not produce as much spatial summation and thus may be easier to block.[32] The perception of pain produced by synchronous activation of afferents by electrical currents may be more difficult to reduce.

A third related explanation for our results is that acute and chronic pain experiences are different and may be blocked by different agents with varying degrees of efficacy. Mannheimer[35] has noted that electrical stimulation applied to the fingertips produces a "first" or "superficial" pain that is sharp, well localized, and short lasting (typical of A fiber activation). In contrast, chronic pain is dull, achy, poorly localized, and long lasting (characteristic of C fiber pain). In addition, he points out that behavioral and psychological components of acute and chronic pain experiences are quite different.[35] Analgesic agents can affect these pain modalities differently. Morphine has been shown to have differential effects on A and C fiber activity. Recording from rat spinal cord, Jurna and Heinz[36] found that it took four times the amount of intravenous morphine to significantly depress evoked A fiber activity than it took to depress evoked C fiber activity (ie, 2 mg/kg versus 0.5 mg/kg). In addition, spontaneous A fiber discharge was not influenced by morphine injection, even at high dosages, whereas spontaneous C fiber discharge could be significantly reduced by a relatively low dosage of morphine.[36] Apparently, acute (A fiber-type) pain would be harder to block than chronic (C fiber-type) pain with any therapeutic modality that involved opioids (ALTENS, for example). Clear support for this concept comes from Strassburg and colleagues.[32] They found that, using TENS, they could produce sufficient analgesia in 29 of 30 patients to extract teeth and in 7 of 9 subjects to perform muscle and peripheral nerve biopsies of the forearm. Yet, with the same stimulation, they could not alter acute electrical pain thresholds in the wrists of 10 subjects.[32] Thus, the ALTENS used in our experiment may not be effective at shifting the acute pain threshold in the fingers, even though it is highly effective at producing analgesia in patients who have chronic pain.[37,38] [TABULAR DATA OMITTED]

The ALTENS analgesia produced in the teeth is a notable exception to this evidence. Electrical pain thresholds can be altered with ALTENS (though naloxone reversal is not consistently shown).[29,39] Perhaps there is something unique about the neurophysiology of the afferent connections in the tooth. The tooth pulp is almost solely innervated by A[unkeyable] and C fibers with limited non-nociceptive afferents.[20] This situation is quite different from that of the skin and muscle, and perhaps the sensations from tooth pulp are more readily influenced as a result.

Another possible explanation for the lack of analgesia to cutaneous pain with ALTENS comes from recent research indicating that deep tissue, but not skin, nociceptive input is modulated via descending opiate controls. Yu and colleagues[40] recorded from rat dorsal horn neurons with convergent input from cutaneous and deep tissue nociceptors. They found that tonic descending inhibition to deep tissue nociceptive input, but not skin nociceptive input, could be blocked by naloxone injections into the third ventricle. Their results indicate that the descending antinociceptive system to deep tissue nociception is endorphinergic in nature, whereas that to skin is not.[40] Thus, our observations seem to support their results (ie, ALTENS that operates primarily via the endogenous opiate system did not alter nociceptive input from the skin, yet has previously been shown to be effective in producing analgesia from chronic, deep tissue pain).[37]

Our observation that pain thresholds were symmetrical on the two body sides in healthy subjects is in keeping with previous literature.[23,41] The results of our study simply add further support to the previous findings.

The lack of correlation in our data between initial pain thresholds and change in pain thresholds disagrees with past studies. Golding and co-workers[15] reported a negative correlation between these measures: Subjects with high initial thresholds required less current to produce the same pain sensation during TENS (ie, these subjects had hyperalgesia); those with low initial thresholds required relatively more current to elicit pain during TENS (ie, these subjects exhibited analgesia). The research team's results, however, also show that a similar significant correlation existed in subjects given sham TENS as well as on the control side of subjects given real TENS. The authors do not comment on these observations in their discussion, despite claiming to have shown a bidirectional effect of TENS analgesia.[15] Zoppi and colleagues[23] also reported a differential change in pain thresholds dependent upon initial thresholds in healthy subjects. Without a control, their data are difficult to interpret. The same group later replicated their results in an evoked potential paradigm. Again, no controls were used.[19] In none of these three studies did subjects undergo an acclimatizing session (as did our subjects). Therefore, their results may be due to changes in arousal, as mentioned previously.

The methods used in this study can be altered to develop a useful model for studying ALTENS analgesia. The type of pain stimuli could be changed from electrical to natural stimuli. The location of pain testing could be moved from the fingertips to a point closer to the midline, especially on the face, in order to maximize the chance of producing good analgesia. Ideally, pain studies should be conducted using patients who have chronic pain, because the underlying physiology of acute and chronic pain seems to be quite different.

The results presented demonstrate that the experimental paradigm used in this study provides a poor model for studying analgesia. The implication of this observation also carries into the field of somatosensory evoked potentials. Researchers are using acute electrical pain models in hands and wrists to elicit cortical responses and are investigating the effect of TENS and electroacupuncture.[15,19,28] Our results show that pain thresholds from fingertip stimulation are not easily changed by electrotherapy; hence, we do not expect to see electrotherapy effects on cortical potential evoked by this method.

Conclusion

Based on the results of this study, we question the apporpriateness of similar acute pain models for studying the therapeutic effects of ALTENS.

(*1)Electronic Health Machines Inc, 20 Eglinton Ave W, Ste 1007, Toronto, Ontario, Canada M4R 1K8.

[unkeyable]Nicolet Instruments, 1200 Aerowood Dr, Unit 1, Mississauga, Ontario, Canada L4W 2S7.

References

1 Melzack R, Vetere P, Finch L. Transcutaneous electrical nerve stimulation for low back pain: a comparison of TENS and massage for pain and range of motion. Phys Ther. 1983;63:489-493.

2 Andersson S, Ericson T, Holmgren E, Lindqvist G. Electro-acupuncture: effect on pain threshold measured with electrical stimulation of teeth. Brain Res. 1973;63:393-396.

3 Oliveri AC, Clelland JA, Jackson J, Knowles C. Effects of auricular transcutaneous electrical nerve stimulation on experimental pain threshold. Phys Ther. 1986;66:12-16.

4 Andersson S, Holmgren E, Roos A. Analgesic effects of peripheral conditioning stimulation, II: importance of certain stimulation parameters. Acupunct Electrother Res. 1977;2:237-246.

5 Mayer D, Price D, Rafii A. Antagonism of acupuncture analgesia in man by the narcotic antagonist naloxone. Brain Res. 1977;121:368-372.

6 Sjolund B, Terenius L, Eriksson M. Increased cerebralspinal fluid levels of endorphins after electro-acupuncture. Acta Physiol Scand. 1977;100:382-384.

7 Pomeranz B. Scientific basis of acupuncture. In: Stux G, Pomeranz B, eds. Acupuncture: Textbook and Atlas. New York, NY: SpringerVerlag New York Inc; 1987.

8 Sjolund B, Eriksson M. Electro-acupuncture and endogenous morphines. Lancet. November 1976:1085.

9 Clement-Jones V, Tomlin S, Rees L, et al. Increased [unkeyable]-endorphin but not Met-enkephalin levels in human cerebral spinal fluid after acupuncture for recurrent pain. Lancet. November 1980:946.

10 Melzack R, Wall P. Pain mechanisms: a new theory. Science. 1965;150:971-979.

11 Mannheimer JS, Lampe G. Clinical Transcutaneous Nerve Stimulation. Philadelphia, Pa: FA Davis Co; 1984.

12 LeBars D, Dickenson A, Besson J. Diffuse noxious inhibitory controls (DNIC), I: effects on dorsal horn convergent neurons in the rat. Pain. 1979;6:305-327.

13 Nelson R, Currier DP. Clinical Electrotherapy. East Norwalk, Conn: Appleton & Lange; 1987.

14 Melzack R. Prolonged relief of pain by brief, intense transcutaneous somatic stimulation. Pain. 1975;1:357-373.

15 Golding J, Ashton H, Marsh R, Thompson J. Transcutaneous electrical nerve stimulation produced variable changes in somatosensory evoked potentials, sensory perception and pain thresholds: clinical implications for pain relief. J Neurol Neurosurg Psychiatry. 1986;49:1397-1406.

16 Salar G, Iob I, Mingrino S. Cortical evoked responses and transcutaneous electrotherapy. Neurology. 1980;30:663-664.

17 Ashton H, Ebenezer I, Golding J, Thompson J. Effects of acupuncture and transcutaneous electrical nerve stimulation on cold-induced pain in normal subjects. J Psychosom Res. 1984;28:301-308.

18 Barr JO, Nielsen DH, Soderberg GL. Transcutaneous electrical nerve stimulation characteristics for altering pain perception. Phys Ther. 1986;66:1515-1521.

19 Francini F, Maresca M, Procucci P, Zoppi M. Relationships between somasensory evoked potential components and cutaneous pain thresholds: effects of transcutaneous electrical nerve stimulation. In: Courjon J, Mauguiere F, Revol M, eds. Clinical Applications of Evoked Potentials in Neurology. New York, NY: Raven Press; 1982.

20 Chapman R, Chen C, Bonica J. Effects of intrasegmental electrical acupuncture on dental pain: evaluation by threshold estimation and sensory decision theory. Pain. 1977;3:213-227.

21 Krause AW, Clelland JA, Knowles CJ, Jackson JR. Effects of unilateral and bilateral auricular transcutaneous electrical nerve stimulation on cutaneous pain threshold. Phys Ther. 1987;67:507-511.

22 Facchinetti F, Sandrini G, Petraglia F, et al. Concomitant increase in nociceptive flexion reflex threshold and plasma opioids following transcutaneous electrical nerve stimulation. Pain. 1984;19:295-303.

23 Zoppi M, Francini F, Maresca M, Procucci P. Changes of cutaneous sensory thresholds induced by non-painful transcutaneous electrical nerve stimulation in normal subjects and in subjects with chronic pain. J Neurol Neurosurg Psychiatry. 1981;44:708-717.

24 Pomeranz B, Niznik G. Codetron: a new electrotherapy device overcomes the habituation problems of conventional TENS devices. Am J Electromed. 1987;1:22-26.

25 Mao W, Ghia J, Scott D, et al. High versus low intensity acupuncture analgesia for treatment of chronic pain: effects on platelet serotonin. Pain. 1980;8:331-341.

26 Wang K, Yao S, Zian Y, Hou Z. A study on the receptive fields of acupoints and the relationship between characteristics of needle sensation and groups of afferent fibres. Sci Sin. 1985;28:963-971.

27 Chiang C, Chang C, Chu H, Yang L. Peripheral afferent pathways for acupuncture analgesia. Sci Sin. 1973;16:210-217.

28 Bromm B. Evoked cerebral potential and pain. In: Fields H, ed. Advances in Pain Research and Therapy, Vol 9. New York, NY: Raven Press; 1985.

29 Chapman R, Benedetti C, Colpitts Y, Gerlach R. Naloxone fails to reverse pain thresholds elevated by acupuncture: acupuncture analgesia reconsidered. Pain. 1983;16:13-31.

30 Jette DU. Effect of different forms of transcutaneous electrical nerve stimulation on experimental pain. Phys Ther. 1986;66:187-190.

31 O'Brien WJ, Rutan FM, Sanborn C, Omer GE. Effect of transcutaneous electrical nerve stimulation on human blood [unkeyable]-endorphin levels. Phys Ther. 1984;64:1367-1374.

32 Strassburg H, Krainick J, Thoden U. Influence of transcutaneous electrical nerve stimulation (TNS) on acute pain. J Neurol. 1977;217:1-10.

33 Lein DH Jr, Clelland JA, Knowles CJ, Jackson JR. Comparison of effects of transcutaneous electrical nerve stimulation of auricular, somatic, and the combination of auricular and somatic acupuncture points on experimental pain threshold. Phys Ther. 1989;69:671-678.

34 Chapman R, Gehrig J, Wilson M. Acupuncture, pain and signal detection theory. Science. 1975;189:65.

35 Mannheimer JS. Commentary. Phys Ther. 1986;66:191-192.

36 Jurna I, Heinz G. Differential effects of morphine and opioid analgesics on A and C fibre evoked activity in ascending axons of the rat spinal cord. Brain Res. 1979;171:573-576.

37 Cheng R, Pomeranz B. Electrotherapy of chronic musculoskeletal pain: comparison of electroacupuncture and acupuncture-like transcutaneous electrical nerve stimulation. Clin J Pain. 1987;2:143-149.

38 Fargas-Babjak A, Rooney P, Gerecz E. Randomized trial of Codetron for pain control in osteoarthritis of the hip/knee. Clin J Pain. 1989;5:137-141.

39 Olausson B, Eriksson L, Rydenhag B, et al. Effects of naloxone on dental pain threshold following muscle exercise and low frequency transcutaneous nerve stimulation: a comparative study in man. Acta Physiol Scand. 1986;126:229-305.

40 Yu H, Hua M, Mense S. The effects of intra-cerebroventricular injection of naloxone, phentolamine and methysergide on the transmission of nociceptive signals in rat dorsal horn neurones with convergent cutaneous-deep input. Neuroscience. In press.

41 Laitinen L, Eriksson T. Electrical stimulation in the measurement of cutaneous sensibility. Pain. 1985;22:139-150.