Immediate effect of alone and combined virtual reality, gait-like muscle vibration and transcranial direct current stimulation on neuropathic pain after spinal cord injury: a pilot study

Abstract

Study design

Quasi-experimental pilot study.

Objectives

Evaluate the immediate effect of virtual reality (VR), gait-like muscle vibration (MV) and transcranial direct current stimulation (tDCS) combined or alone on neuropathic pain in individuals with spinal cord injury (SCI).

Setting

Inpatient rehabilitation centre.

Methods

Four participants (two women and two men) with neuropathic pain after SCI participated in the pilot study. All participants received one session per week for four weeks. Each session started with a single-blind administration of active or sham tDCS (20 min) delivered in a pseudo-randomized order, followed by three interventions applied in a pseudo-randomized order (10 min each): gait-like muscle vibration only, watching a walking self-avatar in VR only and the combination of muscle vibration and VR. The intensity of pain was evaluated using a numeric rating scale (0–10, minimal clinically important difference: 2 points) before and after each stimulation.

Results

Participants reported significant reduction of pain (reduction of two points or more) in 4/7 stimulations where VR was associated with muscle vibration, in 1/8 for VR-alone stimulations and in 1/7 for MV-only stimulations. Significant change in pain was found in 1/8 sham tDCS, but not after active tDCS.

Conclusions

Our pilot study showed immediate pain relief when a walking-avatar VR stimulation was associated with gait-like muscle vibration. Even though previous studies supported tDCS for pain reduction, we did not observe any changes in pain after tDCS, likely due to its application once a week. Further research is needed to strengthen these promising results.

Introduction

Neuropathic pain is a very common, disabling condition after spinal cord injury (SCI), defined as “pain caused by a lesion or disease of the somatosensory nervous system” [1]. Symptoms include pain located in dermatomes at or below the level of injury, allodynia (pain due to a non-painful stimulus) and hyperalgesia (an abnormally increased sensitivity to pain) [2]. Usually described as burning, shooting, squeezing pain, neuropathic pain can be spontaneous or evoked [2].

Neuropathic pain is the consequence of maladaptive neuronal plasticity after SCI [3]. Various changes occur at the peripheral, spinal and supraspinal levels after SCI, with a phenomenon of sensitization, such that neuronal hyperexcitability contributes to the development and maintenance of neuropathic pain [2, 3].

Because underlying mechanisms remain unclear, neuropathic pain post-SCI is clinically challenging to treat [2]. To reverse neuropathic pain, alternative therapeutic approaches using virtual reality (VR) or transcranial direct current stimulation (tDCS) have been tested and have shown very promising results. Various VR treatments have been developed for pain relief, with protocols consisting of virtual walking through immersive VR, where a first-person view and full immersion seem to be more effective against pain [4, 5]. It is important to note that VR intervention can also increase engagement, motivation and have a short-term analgesic effect by distracting attention from neuropathic pain [6, 7]. But long-term pain relief could be a consequence of a reversal of the maladaptive neuronal plasticity through intense facilitation or activation of somatosensory cortex or could be explained by the activation of a “mirror system” with movement observation [5, 8]. Also, transcranial direct current stimulation (tDCS) is a safe and non-invasive technique to reduce neuropathic pain by targeting M1 activity which plays a central role in pain modulation [9, 10]. tDCS could regulate the dysfunctional brain activity by decreasing intracortical facilitation and increasing intracortical inhibition, which might be defective with neuropathic pain [11]. Finally, tDCS could also increase the endogenous opioid release, thereby contributing to decreasing pain [9].

Muscle vibration, and its strong proprioceptive activation, may be a powerful sensory complement to VR to reduce neuropathic pain [12]. Muscle vibration induces a perception of pain-free movement associated, such as walking, with an increased sensorimotor cortical activity [13, 14]. The perceived movement induced by muscle vibration corresponds to a lengthening of the vibrated muscle, e.g. if the quadriceps femoris is vibrated, the perception of movement is a flexion of the knee, in the absence of actual movement [15].

The combination of stimulation of several sensory modalities has shown a greater effect in reducing neuropathic pain. Because multisensory integration of vision, touch and muscle proprioception feedback plays an important role in motor control and body representation [16], interventions that combine sensory stimulations have been developed to improve perception of movement during stimulation, increase neural stimulation and further reduce neuropathic pain [4, 5]. Finally, perception of movement in VR can be improved and may have a stronger effect on neuropathic pain when visual information of movement is congruent with proprioceptive stimulation, as such a combination increased movement perception in participants who were not moving voluntarily [17, 18].

Our goal was to evaluate the immediate effects of VR, gait-like muscle vibration and tDCS combined or alone on neuropathic pain in individuals with SCI in a pilot study. Our hypothesis was that active tDCS combined with the association of virtual reality and muscle vibration would be more effective in reducing neuropathic pain after SCI.

Methods

To obtain preliminary data on the effect of multisensory virtual walking, and its combination with tDCS, we proposed a multisession, quasi-experimental study in individuals with neuropathic pain following SCI.

Participants

Four participants (2 women and 2 men) with neuropathic pain after spinal cord injury (SCI) were included in the study (Table 1). The inclusion criteria were: over 18 years old, complete or incomplete SCI (American Spinal Injury Association (ASIA) Impairment score (AIS) A to D), having ongoing neuropathic pain at or below the lesion (score ≥4 at the “Douleur Neuropathique 4” (DN4) questionnaire) and with stable pharmacological treatment. We excluded participants with ASIA E and with pain of other origins such as musculoskeletal pain, e.g., shoulder pain. The other exclusion criteria were related to tDCS contraindications: psychotic troubles, epilepsy, or pregnancy. The protocol was approved by the CRIR (Centre de Recherche Interdisciplinaire en Réadaptation du Montréal Métropolitain) research ethics committee and National Research Council Canada research ethics board. All participants gave their written informed consent.

Table 1.

Clinical characteristics of the participants.

Pt	Sex	Age (Years)	Cause of the Lesion	Years Since Injury	Neurological Level of Injury	AIS	Pain Location	Type of Pain	Initial DN4 Score
1	F	55	Car accident	32	C7	A	Lumbar region	Below the level of injury	7
2	F	36	Diving	22	C6	B	Lower limbs	Below the level of injury	6
3	M	32	Diving	2	C7	B	Lower limbs	Below the level of injury	6
4	M	60	Fall	6	C7	B	Sacrum, coccyx, lower limbs, pelvic region	Below the level of injury	5

Pt participant, F Female M Male, Neurological level of injury C Cervical.

AIS ASIA Impairment Scale, A Complete impairment with no motor or sensory function left in the sacral segments S4–S5, B: Sensory incomplete impairment with sensory function below the neurological level of injury (the first normal level above the level of injury) and at sacral segments S4-S5 and no motor function left more than 3 levels below the motor level on either side of the body. Score ≥4 at the DN4 indicates neuropathic pain.

Primary outcome measure

We measured pain intensity on an 11-point numerical rating scale (NRS), before and immediately after each intervention. NRS is a validated and easy tool to assess chronic pain after spinal cord injury [1]. NRS is a self-evaluation scale that ranges from 0 (no pain) to 10 (worst pain ever). Based on previous studies, we considered a reduction of 2 points as the minimal clinically important difference [19].

Study design

In this pilot study, each participant had one session per week for four weeks. Each session started with a single-blind administration of active tDCS or its sham for 20 minutes (m) in a pseudo-randomized order, so that at the end of the study, each participant received 2 sham and 2 active tDCS. After the initial tDCS stimulation, three interventions were applied in a pseudo-randomized order: muscle vibration (MV) only, virtual reality (VR) only and the combination of muscle vibration and virtual reality for 10 m each (See Fig. 1).

Fig. 1. Study design for one out of four sessions.

m : minutes, tDCS: transcranial direct current stimulation. * A single-blind administration of active tDCS or its sham in a pseudo-randomized order between sessions. **A 10-minute resting period without stimulation was offered between each intervention.

Interventions

All the interventions were delivered with the participants seated in their own wheelchair.

Transcranial direct current stimulation (tDCS)

To deliver the direct current stimulation, we used a saline-soaked pair of surface sponge electrodes (25 cm²) and a battery-driven constant current stimulator (Model 1300 A; Soterix Medical, New York, NY, USA). Based on the electroencephalogram 10/20 system, the anode was placed over C3-C4 for the stimulation of primary motor cortex (M1) and the cathode over the contralateral suborbital area [10]. In case of symmetrical neuropathic pain, we targeted M1 of the dominant side and for asymmetrical pain, the contralateral M1 was targeted [10]. When direct current was applied, participants felt tingling sensations for a few seconds then nothing. For active tDCS, we delivered a 2 mA current for 20 m. For sham stimulation, a 2 mA current was delivered for 30 s to mimic the active condition, before being turned off automatically.

Visual virtual reality (VR)

Visual virtual walking consisted of an avatar seen through a head-mounted display (Oculus rift^®), in first-person view, walking along a forested path. The simulation used a customized version of the Stroll scene of the bWell platform (run on Unity^®) with the attentional task disabled [20]. The avatar represented a man or a woman, according to the preferences of the participant. Skin colour could also be customized to increase the sense of embodiment. Noise from the environment and steps on the ground were provided in addition to the visual scene. Furthermore, the patient could control the avatar’s upper-limb movements by means of hand-held controllers.

Proprioceptive stimulation/muscle vibration (MV)

To create a perception of gait movement at the lower limbs, we used proprioceptive stimulation through multiple muscle vibration (MV) [13]. We followed a pattern of stimulation based on normal gait kinematics already described in a previous study (see Fig. 2). Briefly, it consists of 12 vibrators (VB115, techno concept; Mane, France) placed on the flexor and extensor muscle groups of the lower limbs (hips, knees, ankles), activated according to the sequence of muscle lengthening extrapolated from actual lower-limb movements during gait. Muscle vibration is indeed known to reproduce proprioceptive activity of the neuromuscular spindles associated with muscle lengthening [15]. When activated, the target vibration frequency was set at 80 Hz. Vibrators were controlled using a custom-made interface programmed under Labview (National Instruments, Austin, Texas, USA). Participants did not voluntarily move their lower limbs during the stimulation.

Fig. 2.

Schematic vibration pattern based on normal gait kinematics.

For both visual and proprioceptive stimulations, we proposed two consecutive periods of gait of 5 m each. The same visual and proprioceptive patterns were used but scaled to represent gait cadence of 60 or 120 steps per minute.

Combination of visual and proprioceptive stimulation

The third condition combined the visual (VR) and proprioceptive (MV) conditions described above. This combination is known to enhance the perception of gait movement compared to VR alone [18]. Stimulations were synchronized through a trigger signal from Unity^® to Labview^® through an Arduino Nano^® board.

Data analysis

The minimal clinically important difference on pain (at least 2-point reduction on the NRS) was used to determine immediate, clinically significant effect individually, in the 4 participants of this pilot study. A change of 2 points and more on the NRS were reported for each condition (VR, MV or combined VR + MV) and each participant, separately for active or sham tDCS, to determine which condition or their combination produced a clinically meaningful immediate effect in a larger proportion in the trial.

Results

Effects of association of virtual reality with muscle vibration versus virtual reality and muscle vibration alone

Every participant except one received 4 stimulations of the association (VR + MV). Indeed, participant 4 received only 3 stimulations due to an increase in pain intensity during MV. This participant did not receive VR + MV at the last session. Pain decreased 2 points or more for 8 out of 15 VR + MV stimulations. Participant 1 showed a reduction of pain for 1 out of 4 stimulations. Participant 2 had a significant reduction of pain for 3 out of 4 stimulations. Participant 3 showed the greatest results with changes of 2 points and more for each stimulation. We had no changes for participant 4 (See individual scores in electronic Supplementary materials).

Every participant received 4 stimulations of VR. Two out of 16 stimulations induced a significant pain reduction. We observed no changes for participants 1 and 2 but for participants 3 and 4, pain reduction was noted for 1 out of 4 stimulations each.

All participants received MV for 4 sessions each, except participant 4, who felt a + 2-point pain during MV at the third session. Two out of 15 stimulations induced a significant change of pain. The improvement was observed only for participant 3.

Overall, VR + MV showed a significant reduction of pain in 53% of the stimulations compared with VR or MV alone that showed a significant reduction in 12.5% or 13.3% of the stimulations respectively. Moreover, 3 out of 4 participants experiencing neuropathic pain had significant improvement with VR + VM, 2 out of 4 with VR alone and 1 out of 4 with MV alone.

Effects of tDCS active and tDCS sham alone and effects on the other interventions

No clinically significant changes were found after active tDCS (Fig. 3), but patient 4 indicated a significant reduction of pain after the sham tDCS (Fig. 4) at the first session. No immediate differences were found between the sessions starting with active tDCS and the sessions starting with sham tDCS.

Fig. 3. Pain intensity after active tDCS, VR and MV alone and combined when the sessions were started with active tDCS.

Pain changes rated on the NRS for pain between 0 “no pain” and 10 “worst pain ever”. The number of stimulations is represented by a greyscale coded according to pain changes, with a maximum reduction (darker greys at the bottom of the histogram) of pain reported of 3 points and with a maximum increase (lighter greys at the top of the histogram) of pain of 2 points. We considered a reduction of 2 points as the minimal clinically important difference. tDCS: Active transcranial direct current stimulation, Active tDCS: tDCS stimulation at 2 mA for 20 min after 30 s intensity ramp-up, VR: Visual virtual reality alone, MV: Gait-like muscle vibration alone, VR + MV: Visual virtual reality and gait-like muscle vibration combined.

Fig. 4. Pain intensity after sham tDCS, VR and MV alone and combined when the sessions were started with sham tDCS.

Pain changes rated on the NRS for pain between 0 “no pain” and 10 “worst pain ever”. The number of stimulations is represented by a greyscale coded according to pain changes immediately after the stimulation, with a maximum reduction (darker greys at the bottom of the histogram) of pain reported of 4 points and with a maximum increase (lighter greys at the top of the histogram) of pain of 2 points. We considered a reduction of 2 points as the minimal clinically important difference. tDCS: Sham transcranial direct current stimulation, sham tDCS: Inactive tDCS after a 30 s intensity ramp-up, VR: Visual virtual reality alone, MV: Gait-like muscle vibration alone, VR + MV: Visual virtual reality and gait-like muscle vibration combined.

Perceptions and adverse effects

Because it is a pilot study, we explored the spontaneous feedback given by the participants to orient future studies. Participant 1 had fewer peaks of pain and less spasticity in the week after the sessions starting with active tDCS. Participant 2 felt less pain during VR and the stimulation felt like she was “floating/sliding”. The gait-like sensation was even stronger when VR and MV were associated, and it was pain-free during all VR + VM interventions. With VM alone, participant 2 felt gait-like movements and when he tried to visualize gait, he felt a “muscle spasm”. After the second session, participant 2 felt sensation that could be related to cyber sickness [21]. Participant 3 twice had the feeling of being in a standing position and feeling relieved, with a sensation of wellness. He added that it could be better in a standing position during the session. During some sessions, an increase of pain score was noted for all the participants. But overall, at the end of each session, the pain score was the same or lower than it was at the beginning of each session. Thus, we did not increase pain except for participant 4 with MV because of the vibrators located at the hip flexors muscles which was a painful area for this participant.

Discussion

The results of this pilot study showed that multisensory virtual walking using VR and MV induced greater reductions of neuropathic pain when these stimulations were combined. But adding tDCS did not show any pain relief alone or combined with VR or MV.

These positive effects could be explained by a better sense of virtual embodiment during the combined proprioceptive-visual VR. Embodiment requires different features such as agency, ownership and self-location [22], and the increase of any of these features could be correlated with greater pain relief [4, 5]. Moreover, we know that proprioception and vision are important for embodiment in real life when they are congruent, so that MV could increase virtual embodiment during VR even more when the stimulations are congruent [4, 23]. In addition, perceived movement during MV could be improved when associated with VR by adding consistent visual feedback for sitting or standing participants [18]. Better virtual embodiment means a more immersive and multisensory processing that demands more attention from participants and lead to better distraction from pain [5, 7]. Beyond distraction, a greater virtual embodiment could lead to cortical activation and contribute to long-term pain relief [5, 8, 24]. This warrants further research on the long-term effects of combined visual-proprioception stimulation on neuropathic pain.

Our study did not show any changes immediately after active tDCS, with no difference between the sessions that started with active tDCS or with sham tDCS. A recent review showed that more evidence is needed to support tDCS for pain relief [25]. In addition, longstanding neuropathic pain after SCI, as in the case of our four participants, could be more refractory to tDCS [26]. Moreover, a greater number of sessions seems to be important to obtain pain relief, as studies with actual pain relief after tDCS used at least 5 consecutive sessions [25], and effects could be immediate or delayed [11, 27]. Finally, the neuromodulating effects of tDCS could enhance the effects of other therapies such as VR or mirror therapy when associated and should be considered for further studies on neuropathic pain despite the results obtained in this study [28, 29].

Stimulation used alone had little instantaneous effect on neuropathic pain intensity. Unexpected results such as pain reduction after sham tDCS were observed for one session for one participant. It could be explained by the endogenous opioid release, i.e. the same as with active tDCS, or due to a placebo response previously observed [9, 30]. VR is commonly used in rehabilitation for pain management. However, our results showed pain reduction in only one session in two participants. In comparison, previous studies had more or/and longer VR sessions than ours [5]. In addition, long-lasting SCI could be refractory to the illusion of leg ownership [4]. Also, MV showed pain relief for 3 sessions. Perception of pain-free movement through MV could be less vivid after SCI and explain the low rate of pain reduction in our study [15].

Our main limit is that the sample size was small. Moreover, all our participants had almost the same clinical characteristics (neurological level of injury, AIS, pain location: see Table 1) hence they are not representative of the population with SCI and we cannot generalize our findings. Our results will guide further research but are not enough to support clinical use of any of the protocols tested here. In addition, some effects observed here may be due to the additive effect of the consecutive applications of VR and MV alone or combined, as changes often appeared at the third modality, i.e., the one applied last in the session.

Only short-term effects could be observed with our protocol. However, each modality could have had a long-term effect because each of them could play a role in neural plasticity. In addition, we used only a unidimensional scale (NRS) to evaluate chronic neuropathic pain that is a multidimensional phenomenon, with different characteristics (paroxysmal pain, continuous pain, allodynia, burning, electrical shocks…), with different underlying mechanisms and different responses to treatments [2]. Moreover, the pain score was not the same before each modality and between participants. At times, we had a very low pain score (0–2/10), making it difficult to observe any reduction of pain. In addition, the effect of the stimulation may vary according to pre-stimulation pain level. Besides, the location of vibrators for MV has proven problematic in cases of allodynia or hyperalgesia. Adaptation of MV delivery should thus be evaluated to enable the combined application of VR + MV in all participants.

Including individuals with AIS A and B lesions could be questionable. However, these are all considered to have a complete motor deficit, that prevents them to voluntarily produce gait movement, and the associated somatosensory afferents. In addition, all the participants are tetraplegic, which increases the risk in developing neuropathic pain [31]. Moreover, they all presented below-level neuropathic pain, which is mostly due to supra-spinal changes [32]. Finally, the AIS classification is a clinical evaluation with some limitations. It is indeed possible to have a cortical activation after somatosensory stimulation on body regions below the level of an injury considered sensory-complete clinically (AIS A) [33]. We thus believe that including one individual with an AIS A lesion fits the pilot nature of the present study.

Conclusion

In conclusion, our results offer pilot data for greater pain relief after VR associated with MV, with limited effect of tDCS, likely because of the duration of the protocol. tDCS should thus still be considered to enhance the effects of another modality in the longer term. Indeed, our goal is to find a combination with lasting pain relief. This warrants further research on innovative modalities, such as VR, MV and tDCS, and their combination, as they also produce fewer side effects than medication. Although we know that virtual walking showed better pain relief than virtual wheeling [5], it could be interesting to evaluate the psychological impact of virtual walking more specifically in the early stages to be sure we do not create psychological issues or interfere with the grief process linked with SCI [34].

Author contributions

All co-authors have significantly contributed to the study, discussed, and interpreted data, critically reviewed the manuscript, and approved the final version. ND, PM, ML, MTL, JH, DB, CP and CD. planned and designed the study. ND, PM, MS, CP and CD acquired the data. PS, CD analysed the data, produced figures, and prepared the draft of the manuscript. DB, CP, and CD supervised the study.

Funding

Funding was obtained from the New Initiatives Program from the Centre for Interdisciplinary Research in Rehabilitation (CRIR) of Greater Montreal, funded by Fonds de recherche du Québec—Santé et —Société et culture.

Data availability

The database generated and analysed in this study is available in the Supplementary file online and from the corresponding author.

Competing interests

The authors declare no competing interests.

Ethical approval

Ethical approval for the protocol has been granted by the research ethics committee of the CRIR and National Research Council Canada research ethics board. All methods were performed in accordance with the relevant guidelines and regulations.

Supplementary information

The online version contains supplementary material available at 10.1038/s41394-024-00696-5.