An investigation into enlarging and reducing the size of mirror reflections of the hand on experimentally-induced cold-pressor pain in healthy human participants

1 Introduction

Pain is a complex sensory, emotional and cognitive phenomenon that is influenced by a variety of biopsychosocial factors including fear, anxiety, attention and expectation. Painful conditions including phantom limb pain and complex regional pain syndrome are known to distort the sense of body image [1, 2, 3]. In complex regional pain syndrome the affected limb may be perceived as large, swollen, heavy or stuck in one position and this may lead to neglect and/or learned non-use of the limb [4,5]. Pain perception can be modulated by observing a mirror reflection of a non-painful limb whilst a painful limb is hidden behind the mirror (i.e. out of view), termed mirror visual feedback (mirror therapy). Mirror visual feedback using normal size reflections of non-painful limbs has been found to reduce clinical and experimentally-induced pain of hands and feet [6, 7, 8, 9, 10].

The findings of studies on patients in pain suggest that reducing the visual appearance of the size of the painful body part reduces pain. Moseley et al. [11] used binoculars to change the visual appearance of chronically painful hands and found that enlarging the view of the hand increased pain and swelling evoked by movement and reducing the view of the hand decreased pain and swelling evoked by movement. Ramachandran et al. [12] used mirror visual feedback to reduce the size of a reflected hand and found that this reduced phantom limb pain. In contrast, a study using healthy human participants by Mancini et al. [13] found that enlarging a reflected view of the hand reduced experimentally-induced contact heat pain (i.e. increased pain threshold) and reducing the size of the reflected view of the hand pain increased pain. One possible reason for the difference in findings was that there were no cues of the presence of injury or of an impending noxious threat in the study by Mancini et al. [13] because the Peltier-type contact thermode used to elicit experimental heat pain was visually inert.

Studies using experimentally induced pain afford a greater degree of control over the environment reducing the impact of confounding variables and maximizing the internal validity of the research [14]. Experimentally induced cold-pressor pain involves immersing an extremity into iced water to produce a deep aching pain. Cold-pressor pain has excellent test-retest stability to assess pain threshold and pain tolerance in student populations [15] and generates higher pain intensity ratings than contact thermode-delivered cold stimuli [16]. Most individuals expect exposure to ice to generate pain and therefore cold-pressor pain is likely to be perceived as a more authentic noxious stimulus than that delivered by a contact thermode. The aim of our study was to compare the effect of enlarging and reducing the visual appearance of a hand using mirror visual feedback on pain threshold, intensity and tolerance in healthy human participants exposed to cold-pressor pain.

2 Methods

2.3 Experimental procedure

Each participant took part in one experiment that measured pain threshold, intensity and tolerance response in a hand immersed in ice-water slurry under four conditions (Fig. 1 ):

• Whilst viewing the painful (immersed) hand (i.e. no reflection control)

• Whilst viewing a normal size reflection of a hand aligned with the painful (immersed) hand (i.e. normal mirror image)

• Whilst viewing an enlarged size reflection of a hand aligned with the painful (immersed) hand (i.e. enlarged size mirror image)

• Whilst viewing a reduced size reflection of a hand aligned with the painful (immersed) hand (i.e. reduced size mirror image)

Block randomization was used to sequence the order of presentation of the four experimental conditions between participants (operationalized using computerised random numbers and sealed enveloped) and a washout period between conditions of 5 min was used to minimise contamination of findings from potential carryover effects and a learning effect from repetitive exposure to cold pressor pain.

Fig. 1

Experimental set up. (a) No reflection control, (b) normal mirror image, (c) enlarged size mirror image, (d) reduced size mirror image.

2.3.1 Cold pressor pain

During each cold-pressor pain test the participant sat on a seat with both arms resting on a desk and flexed at the elbows. They then immersed their non-dominant hand (19 left hand, 1 right hand) in warm water maintained at 37°C for 3min to neutralize hand temperature with core body temperature [19,20]. One minute prior to the removal of their non-dominant hand from the warm water participants aligned their dominant forearm so that they could observe a reflection of the hand in a mirror that was fixed to the side of a container containing a thick slurry of crushed ice and water maintained at 1–2 °C as measured by a digital thermometer. Participants then transferred their non-dominant hand into the ice-water slurry and aligned it with the reflection of the dominant hand. To achieve congruence between the non-dominant hand in ice and the reflected hand, participants had to position their head so that the dominant hand remained out of view. Participants were instructed to observe the reflection of their dominant hand in a mirror whilst concentrating on any sensations occurring in the non-dominant hand in the ice-water. Participants stated ‘Pain’ when they experienced the first sensation of pain anywhere in the non-dominant hand (Fig. 2). The hand remained in the ice-water slurry and a verbal rating of pain intensity was taken 10 s after pain threshold prompted by a visual analogue scale (VAS) where 0 cm = no pain and 10 cm = worst pain imaginable. The hand remained in the ice-water slurry until the participant could no longer tolerate pain at which point they removed their hand. It was planned that participants would be asked to remove their hand from the ice-water slurry 5 min after ‘Pain’ if they had not done so themselves, although no participants reached this cut point. Participants dried their hand, completed a Longo questionnaire [6] and silently waited until the beginning of the next experimental cycle by resting both arms on the table. Pain threshold was measured as the time from immersion of the non-dominant hand in ice-water slurry to ‘Pain’. Pain tolerance was measured as the time from pain threshold until removal of the hand from the ice-water slurry.

Fig. 2

Cold-pressor pain: one measurement cycle.

2.3.2 Longo questionnaire

The Longo questionnaire was used to determine whether the mirror prompted an illusion that the reflected hand was like viewing a real hand directly and consists of 3 items:

• It felt like I was directly looking at my hand rather than a mirror image

• It felt like the hand I was looking at was my hand

• Did it feel like the hand you saw was a right hand or a left hand?

The questionnaire was delivered in an identical manner to a study by Longo et al. in 2009 [6]. Items 1 and 2 were rated according to agreement or disagreement on a 7-point Likert scale (+3 = ‘strongly agree’ to –3 = ‘strongly disagree’). Item 3 was rated using a scale from 0 to 100 (+100 = strong feeling of viewing the right [dominant non-immersed] hand to –100 = strong feeling of viewing the left [non-dominant immersed] hand) with the scale reversed for the one participant whose left hand was dominant.

2.3.3 Normal, enlarged and reduced size mirror image conditions

Standard (normal size), convex (enlarged size) and concave (reduced size) circular mirrors (diameter = 200 mm) were attached to the side of the ice-water slurry container to produce mirror images of the dominant (non-immersed) hand. Participants were seated at a table with their dominant forearm aligned parallel with the mirror so that the middle finger was at a distance of 20 cm perpendicular from the mirror. This enabled participants to view the mirror image of their dominant (non-immersed) hand so that, with minor adjustments it gave an illusion that the hand was attached to the partially visible forearm of the non-dominant hand which was out of view and immersed in the ice-water slurry. A distance of 20 cm perpendicular from the mirror enabled the reflected image created by the normal mirror to be the same size as the real hand. Reflections for the enlarged and reduced size conditions were by a factor of ×2 normal size. A white neutral background was used so that there were no additional images in the background that could distract the participant from viewing the reflected hand.

2.3.4 No reflection control

In this condition participants were seated at a table so that the middle finger of their dominant forearm was aligned parallel and at a distance of 20 cm to the wall of a container containing the thick slurry of crushed ice. The container was transparent which allowed participants to view their non-dominant immersed hand whilst it was immersed in the slurry of crushed ice.

3 Results

Twenty healthy volunteers expressed interest in the study and all started and completed the experiment (mean ± SD age = 23.55 ±4.01 years, maximum = 36 years, mini-mum=18 years; mean±SD weight = 71.9±15.57kg; mean±SD height = 168.39 ± 9.53 cm; female n = 10; right handed n = 19).

Summary data is shown in Table 1. It was noteworthy that SDs for pain threshold and pain tolerance were large and inspection of raw data revealed that values from two participants were large and outliers. A sensitivity analysis was conducted by removing these data points and it was found that this brought the SD values into range but did not change the findings of the statistical analysis. Therefore, the statistical analysis that included data from these two participants is presented below.

Mauchley’s test showed that sphericity of pain threshold data could not be assumed so Greenhouse-Geisser corrections were used. There were no significant effects for Condition (F(2.06, 37.02) = 0.79, p = 0.464), Sex (F(1,18) = 1.15, p = 0.298) or Condition × Sex interaction (F(2.06, 37.02) = 1.22, p = 0.308). Mauchley’s test showed that sphericity of pain intensity data could be assumed. There were no significant effects for Condition (F(3, 54) = 0.58, p = 0.62), Sex (F(1,18) = 1.15, p = 0.298) or Condition × Sex interaction (F(3, 54) = 1.58, p = 0.20). Mauchley’s test showed that sphericity of pain tolerance data could not be assumed so Greenhouse-Geisser corrections were used. There were no significant effects for Condition (F(1.93, 34.7) = 1.88, p = 0.169), Sex (F(1,18) = 2.19, p = 0.157) or Condition × Sex interaction (F(1.93, 34.7) = 2.08, p = 0.141).

The no reflection condition produced the strongest illusion of viewing one’s own left hand when viewing the right hand when compared with all reflection conditions (1 ). Mauchley’s test showed that sphericity of each item of the Longo questionnaire could not be assumed so Greenhouse-Geisser corrections were used. For item 1 there were significant effects for Condition (F(2.45, 44.11) = 11.715, p < 0.001) but no significant effects for Sex (F(1, 18) = 0.028, p = 0.868) or Condition × Sex interaction (F(2.45, 44.11) = 0.855, p = 0.452). Pairwise comparisons revealed that participants agreed more strongly with the statement “It felt like I was looking directly at my hand rather than at a mirror image” for the no reflection condition compared with the 3 mirror reflection conditions. There were no significant differences between the 3 mirror reflection conditions.

For item 2 there were significant effects for Condition (F(2.05, 45.01) = 8.09, p< 0.001) but no significant effects for Sex (F(1, 18) = 1.53, p = 0.232) or Condition × Sex interaction (F(2.05, 45.01) = 1.76, p = 0.176). Pairwise comparisons revealed that participants agreed more strongly with the statement “It felt like the hand I was looking at was my hand.” for the no reflection condition compared with the 3 mirror reflection conditions. There were no significant differences between the 3 mirror reflection conditions.

For item 3 there were significant effects for Condition (F(2.80, 50.24) = 6.75, p = 0.001) but no significant effects for Sex (F(1, 18) = 0.013, p = 0.912) or Condition × Sex interaction (F(2.80, 50.24) = 0.677, p = 0.56). Pairwise comparisons revealed that participants agreed more strongly with the statement “Did it seem like the hand you saw was a right hand or a left hand?” for the no reflection condition compared with the 3 mirror reflection conditions. There were no significant differences between the 3 mirror reflection conditions.

4 Discussion

Mirror visual feedback was originally used to relieve post amputation phantom limb pain [21] and nowadays it is has been found to be beneficial for the management of chronic pain of hands and feet [22], especially complex regional pain syndrome [23] and phantom limb pain [24]. There is also evidence that mirror visual feedback is beneficial for rehabilitation of limb(s) following, stroke, traumatic injury or surgery [25]. Evidence suggests that there is a relationship between pain reduction during mirror visual feedback and a reversal of dysfunctional cortical reorganization in the primary somatosensory cortex and a decrease of activity in the inferior parietal cortex [26]. Mirror visual feedback also increases neural activity in brain regions associated with motor control, attention and cognitions accompanying motor action control [10]. Most practitioners use normal size reflected images of the non-painful limb during mirror visual feedback training.

Our study found that enlarging and reducing the size of a hand using mirror visual feedback did not affect pain threshold, intensity and tolerance in healthy human participants exposed to cold pressor pain. Mirror visual feedback, graded motor imagery and sensory discrimination retraining are techniques that have been developed to normalise cortical remapping of limbs in some chronic pain sates including phantom limb pain and complex regional pain syndrome. Mirror visual feedback may have less influence on experimentally-induced pain in healthy participants because of the absence of cortical remapping. Nevertheless, previous studies using healthy human participants have detected changes. Our findings differed from Mancini et al. [13] who found that enlarging the view of a reflected hand reduced pain (increased pain threshold) of a hand exposed to contact heat delivered using a Peltier-type thermode 13 mm in diameter and held on the skin by a mechanical arm. Mancini et al. [13] also found that reducing the size of the reflected hand increased pain (decreased pain threshold). Recently, Romano and Maravita [27] used a lens to magnify the view of a hand and found that enlarging the appearance of a hand reduced the intensity of experimental pain induced by pressing the blunt end of a needle on the surface of the skin. They suggested that magnifying the appearance of body parts results in cognitive and anticipatory reactions, mediated by expectancies about the intensity of painful stimuli, to prepare the individual for contact with the noxious stimulus [27]. We selected cold-pressor pain because individuals are familiar with the association between exposure to ice and a resultant pain sensation [28] and this would be stronger than expectancies about Peltier-type contact thermode attached to the skin [13] which provides no visual threat, at least on first exposure, because the individuals have no prior association with a Peltier-type contact ther mode and pain. Our study design did not enable estimation of the influence of expectancy of the intensity of the painful stimulus and outcome. However, Romano and Maravita [27] did detect differences in experimental pain to different sized reflected limbs to stimuli that would be expected to produce pain (i.e. a blunt end of a needle pressed on the surface of the skin). This suggests that our inability to detect differences was not predominantly due to expectancy.

The fact that we did not detect a difference in cold-pressor pain between reflected and non-reflected conditions is of concern and may be explained in part by the strength of the mirror visual feedback illusion. The no reflection condition which involved participants observing their hand directly, was reported to produce a stronger sense of viewing one’s own hand compared with all of the reflection conditions. This may suggest that mirror visual feedback was not strong. Alternatively, it is possible that participants interpreted the questions literally, for example, “I was directly looking at my hand rather than a mirror image” instead of “It felt I was directly looking at my hand rather than a mirror image”. This is supported by the fact that the hand was not entirely visible in the no reflection condition as it was submerged in ice and observed through a semi-opaque container wall. Furthermore, the reflected hand did not appear to be sub-merged in ice-water slurry and this might have compromised the feeling that they were looking at a real hand rather than a mirror image. We plan to modify the experimental set up in future studies by generating a mirror reflection that generates an illusion that the reflected hand is sub-merged in ice-water slurry.

The function of mirror visual feedback is to create a sense that the reflected limb belongs to oneself, that is, the reflection is perceptually embodied within the body schema. We did not measure perceptual embodiment in the current study, and it would be interesting to determine whether there were differences in the degree of perceptual embodiment achieved using the different sized mirror reflections. In future, investigators using mirror visual feedback may consider measuring perceptual embodiment in their studies using tools that capture subjective experience (e.g. the three item embodiment scale adapted from Mussap and Salton [29]) or physiological correlates (e.g. proprioceptive drift [30] or skin conductance response [31, 32, 33]). Moreover, it may also be important to include an adaptive phase to ensure the mirror reflection has been perceptually embodied in the clinical setting.

In contrast to studies using experimental pain the findings of studies using individuals with pre-existing painful conditions suggest that enlarging painful limbs increases pain. Moseley et al. [11] found that magnifying the view of a chronically painful and dysfunctional arm using binoculars increased pain and swelling evoked by movement. Ramachandran et al. [12] found that reducing the size of a reflected limb using mirror visual feedback reduced phantom limb pain. Preston and Newport [34] used real-time video capture techniques to manipulate a first person view of their hand so that it appeared stretched and shrunk in 20 individuals with painful osteoarthritic hands. They found that 85% of participants reported benefit, but only when the painful part of the hand was manipulated. There was no reported benefit when the appearance of the whole hand, which had painful and non-painful areas, was enlarged or reduced. Interestingly, some participants found that pain was reduced when the painful area was stretched and shrunk, whereas others found that pain reduction was greater when the painful area was stretched, or when the painful area was shrunk. The authors suggested that participants may vary in their cortical representations causing variations in matching of cortical representations and reduction of apparent swelling. Inter and intra individual variation in response to visual distortion of the size of painful limbs was large and they speculated that their findings may also represent experimental ‘noise’ and or placebo effects.

5 Conclusion

In conclusion, enlarging and reducing the visual appearance of the size of a hand using mirror visual feedback did not alter pain threshold, intensity and tolerance in healthy human participants exposed to cold-pressor pain. The different sizes of hands generated by mirror visual feedback created an illusion of looking at their own hand but this was not as strong as looking directly at the hand. We hope that these findings catalyse further research, including investigations to determine relationships between perceptual embodiment and mirror reflections of different sized body parts. In this regard, investigators and clinicians using mirror visual feedback may consider including an adaptive phase to ensure the reflection has been perceptually embodied.

Highlights

• Mirror visual feedback produces relief of pain.

• We assessed different sized hand reflections on cold-pressor pain.

• Changing the size of mirror reflections of the hand did not affect pain variables.

• Studies on relationships between embodiment, reflections and pain are needed.