Article Text

Download PDFPDF

Pain perception in humans: use of intraepidermal electrical stimulation
  1. Koji Inui,
  2. Ryusuke Kakigi
  1. Integrative Physiology, National Institute for Physiological Sciences, Okazaki, Aichi, Japan
  1. Correspondence to Dr K Inui, Integrative Physiology, National Institute for Physiological Sciences, Okazaki, Aichi, Japan; inui{at}


The choice of a system specific stimulus is difficult when investigating the human nociceptive system, in contrast with the tactile, auditory and visual systems, because it should be noxious but not actually damage the tissue. The discomfort accompanying system specific stimulation must be kept to a minimum for ethical reasons. In this review, recent progress made in the study of human pain perception using intraepidermal electrical stimulation (IES) is described. Also, whether IES is a viable alternative to laser stimulation is discussed. IES selectively activates Aδ nociceptors, elicits a sharp pricking sensation with minimal discomfort and evokes cortical responses almost identical to those produced by laser stimulation. As IES does not require expensive equipment, and is easy to control, it would seem useful for pain research as well as clinical tests.

Statistics from

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.


Pain, particularly its emotional component, is essential for survival. However, excessive pain is distressful. Therefore, pain research in humans is important for uncovering the underlying mechanisms of this essential function as well as for establishing treatment for pain relief. The recent development of non-invasive techniques has enabled us to examine directly the human brain, and the number of reports on pain perception using functional brain imaging techniques has progressively increased in the past 20 years. In general, studies using non-invasive techniques, such as electroencephalography, magnetoencephalography (MEG), positron emission tomography and functional MRI (fMRI) have found that noxious stimuli activate several areas of the brain, including the thalamus, basal ganglia, primary (S1) and secondary (S2) somatosensory cortex, insula and cingulate cortex (figure 1A).

Figure 1

Use of intraepidermal electrical stimulation (IES) for studies on human pain perception. (A) Cortical activity detected by magnetoencephalography (MEG) in the primary somatosensory cortex (S1), secondary somatosensory cortex (S2), insula and cingulate cortex following IES to the left hand. (B) Photograph of the concentric bipolar needle electrode for IES. (C) The current passing through the electrode is spatially restricted to the superficial part of the skin where nociceptive free nerve endings are located. (D) Comparison of evoked potentials following stimulation of the hand between IES (blue) and laser (red) stimulation. Note the similar waveform and a 40 ms delay for laser stimulation. (E) Estimation of conduction velocity (CV) in the spinal cord. Very similar waveforms are evoked by IES to the back midline at the C7 and Th10 levels. (F) Primary responses to IES in S1 recorded by MEG. Note the triphasic waveform of the early S1 activity with polarity reversals at a 10 ms interval. (G) Estimation of CV of C fibres of the lower limb. The calculated CV was 1∼1.1 m/s for N2, P2 and RT (reaction time). (H) Activation of Aβ, Aδ and C receptors by one electrode. By using different parameters, different receptors can be stimulated at the same site.

The choice of an appropriate stimulus is another important aspect of pain research in humans because research into the human nociceptive system is limited by ethical constrains because of possible tissue damage and the discomfort evoked by a noxious stimulus. There are various ways to activate the nociceptive system, including chemical, thermal, electrical and mechanical stimulation. Each method has its own advantages and disadvantages but, ideally, the stimulation should be safe, reproducible and quantifiable.1 In addition, it should stimulate Aδ or C nociceptors selectively if one wants to specifically investigate activation of the nociceptive system. For research or clinical testing that requires precise information of latency, such as evoked potentials, a steep rise in the intensity of the stimulus is also important. From an ethical point of view, the discomfort accompanying system specific stimulation should be as weak as possible.

Electrical stimuli fulfil many of these requirements but lack selectivity. Because mechanoreceptors have a lower electrical threshold than nociceptors, electrical stimuli always coactivate mechanoreceptors of the tactile system at a noxious intensity. Mechanical stimuli, such as pinpricks, which are often used for clinical tests, lack selectivity as well as the steepness. For a similar reason, the usefulness of chemicals for pain research is limited.2

Laser stimuli delivered as a brief pulse with a steep rise in intensity can activate cutaneous nociceptors without the concomitant activation of mechanoreceptors.3 Therefore, laser stimulation is the best means of activating the human nociceptive system at present. In fact, lasers are used in research as well as clinical testing.4 5 One problem with laser stimulation however is that the equipment needed is expensive.

Here we review studies using intraepidermal electrical stimulation (IES) developed for the selective activation of cutaneous nociceptors. An electrical method that can selectively stimulate nociceptors would clearly be useful for pain research or clinical tests.

Intraepidermal electrical stimulation


This method is based on the fact that nociceptive fibre terminals are located in the epidermis and superficial layer of the dermis, while other fibres end deep in the dermis. When the superficial layer of the skin is electrically stimulated, the localised current is expected to selectively activate nociceptors. For this purpose, we made a pushpin-like electrode with a stainless steel needle, 0.2 mm in length.6 Although it successfully stimulated cutaneous Aδ nociceptors,6 7 the range of current at which there was no concomitant activation of Aβ mechanoreceptors was narrow—that is, as the current increased in intensity, it reached far enough to activate mechanoreceptors located deeper than nociceptors. Then we improved the method by employing a concentric bipolar configuration (figure 1B). The cathode used was an outer ring, 1.2 mm in diameter, and the anode was an inner needle that protruded 0.1 mm from the outer ring.8 The effective range for the selective activation of nociceptors widened because less current spread to undesired skin layers (figure 1C). We confirmed the effectiveness of the concentric configuration at reducing undesired loop current in rats.9


When the electrode is gently pressed against the skin, the needle tip is inserted adjacent to the nerve endings of the thin myelinated fibres in the epidermis and superficial part of the dermis. As there is no blood in the epidermis, the IES electrode cannot cause bleeding. Although we have never had an infection due to insertion of the electrode, the skin is first disinfected with alcohol and the electrode is for single use only. Unlike with laser stimulation, there is no undesired skin effect, such as heat burn or erythema. The electric stimulus can be a conventional square wave pulse of 0.5∼1.0 ms but a slowly rising pulse, such as a triangular wave,10 is better. Double pulses with a 10∼25 ms interval are usually used to obtain clear responses but a single pulse is also used when a precise response latency is necessary (eg, see Inui et al11). The current is of an intensity that produces a definite sensation of pain, 2∼6 on the visual analogue scale (0∼10). IES can be applied to any area of the body. To augment the response, two or three electrodes, 10 mm apart, are used. In recent studies, a triple electrode type (NM-980W; Nihon Kohden, Tokyo, Japan) has been used.


When a weak current, of approximately 0.1∼0.5 mA, is applied by IES, a sharp pricking sensation, an indication of Aδ nociceptor activation, is elicited without any other sensations. The magnitude of the pricking sensations increases with an increase in stimulus intensity, number of pulses and pulse duration. The intensity of the painful sensation increases slightly with the use of multiple electrodes. These results suggest that for painful sensations, the contribution of temporal summation is greater than that of spatial summation. The pricking sensation is abolished by the local application of lidocaine.12

Cortical responses to IES and conduction velocity

Cortical responses to IES are recorded using an evoked potential (EP) as large vertex potentials consisting of a negativity (N2) and a positivity (P2),6 sometimes preceded by an earlier positivity (P1).13 Figure 1D compares N2/P2 following IES of the hand and following stimulation of Aδ nociceptors by a CO2 laser at the same site. The waveform is very similar but with a latency difference of 40 ms due to the temperature conduction time in the skin for laser stimulation.3 Therefore, IES and laser evoked potentials are almost the same. As the latency of the somatosensory vertex potentials depends on the time taken for the signals to reach the brain, the type of peripheral nerve can be roughly estimated based on N2/P2 latency. For example, in the hand area, stimulation of Aβ fibres results in N2 peaking at 140 ms while N2 for Aδ and C nociceptor stimulation peaks at about 200 ms and 800∼900 ms, respectively.

Responses in the somatosensory cortex are also very similar between IES and laser stimulation—that is, in studies using MEG, S1 in the hemisphere contralateral to the stimulation and S2 of both hemispheres are activated.11 14 15 When the time delay for laser beams due to temperature conduction is taken into consideration, the response latency of each cortical activity is almost the same. In addition, the temporal profile of the IES induced cortical response is similar to that evoked by high intensity electrical stimulation in a patient who has no Aβ fibres due to sensory neuropathy.16

As the latency of EP components is longer following IES of a distal rather than a proximal site due to the distance travelled, peripheral conduction velocity (CV) can be calculated by dividing the difference in latency between the EP components by the distance between the two sites. With this method, mean CV was 15.1 m/s using EPs (hand and upper arm)6 and 15.6 m/s using MEG (hand and elbow).7 Both values are within the range for Aδ fibres (4∼30 m/s), as measured by microneurographic studies.17 18

In summary, IES elicits a sharp pricking sensation without other sensations via peripheral signals ascending through Aδ fibres, and produces cortical responses that are almost the same as those evoked by laser stimulation. Both the painful pricking sensation and evoked brain responses (Aδ fibre latency) are abolished by local application of lidocaine. These findings suggest that IES selectively activates Aδ nociceptors. Recently, a European group verified this by showing that: (1) after the selective denervation of capsaicin sensitive nociceptors by 72 h application of a capsaicin cream, IES evoked cortical responses were almost abolished, and the threshold for detecting IES increased markedly (0.09 vs 0.6 mA) compared with controls; (2) when the conduction of myelinated nerve fibres was selectively blocked by compression, the time course of the blockade of responses to IES followed closely the time course of the blockade of Aδ fibre responses to laser stimulation; and (3) IES with a high current (2.5 mA) coactivated Aβ-fibres.19 In some studies, a similar concentric bipolar electrode without a needle was used to elicit a painful sensation.20 Stimulation with this electrode is easy to control and less invasive than IES but it activates Aβ fibres.

Merits of IES

IES has several advantages over other noxious stimuli and should contribute to progress in pain studies. We next present some of the studies taking advantage of IES.

Selective activation of Aδ nociceptors without the need for expensive equipment

The stimulus is easy to control and requires no specialised skills. In certain clinical patients, the assessment of small fibre function is important. However, because few hospitals have a laser stimulator, mechanical stimuli are often used even though they lack nociceptive selectivity. As IES is very simple, it seems useful for clinical tests. A new portable stimulator weighing just 290 g and specialised for IES (PNS-7000; Nihon Kohden) should enhance its use in clinics. When starting IES with a current of 0.01 mA and increasing the stimulus in steps of 0.01 mA, a weak pricking sensation occurs at the threshold. As there is no other sensation below the threshold, the threshold for Aδ nociceptor activation can be easily assessed. Therefore, IES is expected to detect functional changes in peripheral small fibres. To test this possibility, we examined the effects of lidocaine tape on pain threshold and EPs.12 As expected, local application of lidocaine significantly elevated the pain threshold, and almost abolished EPs while effects on tactile sensation and tactile EPs were very small.

One possible use of IES is for so-called small fibre neuropathy. The diagnosis of small fibre neuropathy is often difficult because small fibres are invisible in routine nerve conduction studies. For example, in diabetic patients, disturbances begin in small fibres in the distal part of the limb. Our recent study21 confirmed the usefulness of IES for evaluating small fibre function in diabetic patients. As another example, Obayashi et al22 recently reported a case of domino liver transplantation induced amyloid neuropathy. The patient, due to sclerosing cholangitis, underwent a domino liver transplantation reusing a resected liver from a patient with familial amyloid polyneuropathy. When thermohypesthesia and hypoalgesia appeared 7 years after the transplantation, results of neurological examinations, including tests of tendon reflexes, vibration sense, proprioception and nerve conduction, were all normal but the Aδ nociceptor pain threshold by IES was elevated. This report suggests that a follow-up examination of small fibre function is important for such patients and IES can serve this purpose.

IES can be applied to any part of the body

With most MEG or fMRI machines, applying laser beams to areas other than the limbs is difficult. Therefore, if we want to stimulate these cutaneous sites and record brain responses, IES is useful. A recent study showed that cortical magnetic responses are clearly recorded following IES to various areas, including the neck, face and back (Omori and Isose, unpublished data). The use of IES would stimulate pain studies using fMRI. Yoshino et al23 reported that IES could be used safely in an fMRI room and evoked clear brain activity detectable as haemodynamic changes.

By recording EPs following stimulation of two cutaneous sites, we can measure the CV of the periphery as well as in the spinal cord. This information may be useful for certain clinical cases such as demyelinating diseases. The CV in the spinothalamic tract can be estimated by stimulating two different levels of the back midline. As the peripheral conduction distance is short and similar between two sites, the latency difference is due to the conduction time difference in the spinal cord. However, because N2/P2 is an endogenous EP component, its latency and amplitude are affected by the subject's internal state. To reduce this undesired effect, random stimulation of the two sites is useful. Figure 1E shows an EP recording following stimulation of the back at the C7 and Th10 vertebral spinous process levels. In this case, very similar waveforms were evoked although that for the distal site had a longer response latency. The calculated CV is 13.5 m/s for P1 and 13.0 m/s for N2. The results indicate that the CV is similar between the periphery and spinal cord, which is consistent with the results of studies using laser stimulation.24–26 Although N2/P2 is an endogenous component common to all sensory modalities,14 27 it is easily recorded and useful for estimating CV or task related EP components.28

Steep rise in stimulus intensity

As IES is an electrical method, it provides a good time locked stimulus, which is important when analysing responses in the order of milliseconds. Several studies have taken advantage of this. In an MEG study investigating early responses to IES in S1,11 the early S1 activity was a sharp transient of approximately 80∼100 ms following IES and reversed its polarity once or twice after a 10 ms interval similar to the 20/30 ms component evoked by tactile stimulation of the hand (figure 1F), a common feature of the primary response among sensory modalities (for visual and auditory systems, see Inui et al29 30). Because of the polarity reversing nature of primary cortical responses with a 10 ms interval, a small latency jittering of 10 ms of peripheral activation among each trial is enough to cancel out the response. In fact, no studies using laser stimulation have detected early cortical activation except one by Wang et al31 in which jittering of the response was corrected for each trial. As another example, there are two studies using a pair stimulation paradigm.8 32 When one wants to deliver two different stimuli at various conditioning test intervals, the timing of the onset of peripheral activation is particularly important. In an MEG study by Inui et al,8 cortical responses to paired noxious (IES) and innocuous (conventional transcutaneous electrical stimulation) stimuli applied to the back at 11 conditioning test intervals of −500∼500 ms were recorded to reveal cortical mechanisms underlying pain relief by tactile inputs. Results showed that IES induced responses were markedly inhibited when transcutaneous electrical stimulation was applied 20∼60 ms later and 0∼500 ms earlier than IES. Based on the time taken for each signal to reach the spinal cord and the cortex, we concluded that cortical responses to IES can be inhibited by innocuous tactile stimuli at the cortex without a contribution at the spinal level.

Minimal discomfort and possible use in animal studies

IES evokes clear cortical responses at a weak current around twice the threshold. However, at this intensity, some subjects report that the stimulus is not painful at all (visual analogue scale 0∼1). This means that the IES evoked sensation is a pure noxious sensation with minimal discomfort—that is, pain. Indeed, IES evoked sensations are very weak compared with the uncomfortable feeling caused by conventional transcutaneous electrical stimulation at a painful intensity. Unless the aim of the study is discomfort, a less uncomfortable stimulus is better with respect to the ethical restrictions of studies on the nociceptive system. In an MEG study by Wang et al33 investigating the effects of sleep on IES induced cortical responses, subjects were rarely awakened by IES at an intensity high enough to obtain clear cortical responses before sleep.

In animal studies, the use of mechanical stimuli, such as pinching the tail, is common. However, such a stimulus inevitably coactivates Aβ mechanoreceptors. Therefore, if possible, selective stimulation is desirable. In spite of its usefulness for pain studies in humans, laser stimulation is rare in animal studies. One reason is the cost of laser machines but another may be the difficulty of applying laser beams to freely moving animals. In addition, adjusting the laser energy to an appropriate strength is difficult. IES can be applied at any time without immobilising the animal once the electrode is attached to the skin.

For research into the animal nociceptive system, one Australian group used IES. In awake dogs, van Oostrom et al34 recorded EPs following IES to the hind paw. Results showed that: (1) the amplitude of the N2/P2 components increased with an increase in stimulus intensity (0.2∼1.0 mA); (2) CV was 5∼20 m/s; and (3) when the stimulus intensity was increased, there were mild behavioural reactions, withdrawal of the stimulated hind paw and lip licking. It is worth noting that a clear EP recording is possible in awake animals and, in addition, the behavioural response is mild when IES evokes clear cortical responses.

Weak points of IES

As nociceptive free nerve endings are located in the epidermis while the other thicker fibres run more deeply in the dermis, the current passing through the electrode should be spatially restricted to the superficial layer of the skin. In other words, IES activates tactile mechanoreceptors in the dermis when the current is too strong. In fact, results of a study by Mouraux et al19 using a nerve conduction blockade showed that IES at 2.5 mA activates Aβ mechanoreceptors in addition to Aδ nociceptors. Therefore, one cannot use a strong current even when intense sensations of pain are necessary. Usually, the threshold for stimulation of Aδ nociceptors by IES with double pulses is below 0.1 mA, and 2∼3 times the threshold is enough to obtain clear cortical responses. At around this intensity, IES selectively activates Aδ nociceptors. However, one would consider the painful sensation to be too weak at this intensity. For a stronger sensation, spatial summation by use of multiple electrodes or temporal summation by a long duration pulse or pulse train should be considered instead of an increase in intensity.

Stimulation of C fibre by IES

Now we shall describe our recent attempt to selectively stimulate C nociceptors by IES. Because of the very high electrical threshold, it is difficult to stimulate C nociceptors selectively by conventional transcutaneous electrical stimulation. Although in isolated nerves of animals a specific method such as anodal blocking35 can be used for this purpose, such an invasive technique cannot be applied to humans. Based on differences in the distribution of C and Aδ nociceptors (Aδ <C), Plaghki's group reported the successful stimulation of C nociceptors by laser beams for the first time.36 They stimulated a tiny area of the skin with a laser beam that is expected to hit C nociceptors exclusively, and the results supported this. The difference in the threshold of the response to thermal stimuli between C (40°C) and Aδ (46°C) nociceptors is also useful for the selective activation of C nociceptors by laser beams. Tran et al37 successfully stimulated C nociceptors by employing a low energy laser beam to a tiny skin area.

As for electrical stimulation, the selective activation of C nociceptors seems impossible because of the high electrical threshold. However, the higher density of C nociceptors might be an advantage if the current passing through the concentric electrode is limited to a very small area. In addition, there are several reports of factors that are effective at activating C fibres: a pulse of long duration, a slowly rising pulse, temporal and spatial summation, and anodal stimulation (see Otsuru et al10). Based on these reports, we tested IES for the selective stimulation of C nociceptors under the following conditions: (1) the anode was the inner needle and the cathode was the outer ring; (2) the electric pulse was a triangular wave with a rise and fall time of 1 ms; (3) the stimulus was a train of three pulses at 50 Hz; and (4) three electrodes 10 mm apart were used.10


IES elicits weak sensations with a reaction time of approximately 1 s following stimulation of the hand area. The sensory threshold is approximately 0.04 mA. The sensation evoked varies among subjects or among penetrations but is usually the feeling of a light touch and sometimes pricking or slight burning, which is similar to the sensations evoked by a low intensity laser beam applied to a tiny area of the skin.37 38 A warm or itchy sensation is very rare. Under these conditions, there is no clear axonal reflex flare reaction mediated by mechanoinsensitive C nociceptors.39 We consider IES to activate polymodal C nociceptors.40 Therefore, the electrical threshold for the stimulation of C nociceptors is not as high as generally considered.

Conduction velocity

Similar to Aδ nociceptors, CV can be measured by recording EPs. Because of the long reaction time (RT) for C nociceptor stimulation, RT can be also used to estimate CV. Figure 1G shows an example of EPs following stimulation of the foot and knee. Note the very late N2/P2 component compared with Aδ stimulation. The distance between the two stimulated sites was 43 cm and the latency difference was 392 ms for N2, 404 ms for P2 and 440 ms for RT, which yielded a conduction velocity of 1.0∼1.1 m/s. The mean CV following stimulation of the hand and forearm among eight subjects was 1.5 m/s.10

Stimulation of Aβ, Aδ and C fibres through the same electrode

As IES stimulates Aδ nociceptors when the inner needle is the cathode, Aδ and C nociceptors can be stimulated with one electrode by switching the polarity. Figure 1H shows an example of hand stimulation. First, we recorded C responses with anodal stimulation. The peak latency of N2 was about 1 s. Then the polarity was switched. The cathodal stimulation now evoked Aδ responses with N2 peaking at 200 ms.

As the outer ring of the IES electrode is attached to the skin, Aβ mechanoreceptors can be activated by standard monopolar stimulation through the outer ring. The waveform in response to Aβ stimulation in figure 1H has a well known N2 component peaking at 140 ms. In this example, each type of stimulation evoked very similar potentials, consisting of P1, N2 and P2 components. If we want to stimulate different nerve fibres at the same cutaneous site, this method seems useful and easy.

Future perspectives

We have reviewed studies using IES. We believe that it is useful for basic research as well as clinical tests, and will help us to better understand the physiology of the nociceptive system, pathology of pain related disorders or mechanisms of the analgesic effects of drugs. To validate the usefulness of IES and to establish normative data, however, studies using a large group of normal subjects are necessary. For clinical testing, IES can be used for pain disorders such as fibromyalgia as well as small fibre neuropathies such as diabetic neuropathy. IES seems particularly suitable as a screening test because it can be used easily in a clinical setting. As for the stimulation of C nociceptors by IES, however, we feel that there is still room for improvement to obtain responses good enough for clinical testing.



  • Funding This study was supported by the Takeda Science Foundation, grant number tokutei II 2008.

  • Competing interests None.

  • Ethics approval Ethics approval was provided by the National Institute for Physiological Sciences.

  • Provenance and peer review Commissioned; externally peer reviewed.