Schwoebel (2001) pain and the body schema. evidence for peripheral effects on mental representations of movement

Brain (2001), 124, 2098–2104
Pain and the body schemaEvidence for peripheral effects on mental representations ofmovement John Schwoebel,1,2 Robert Friedman,3 Nanci Duda2 and H. Branch Coslett2,1 1Moss Rehabilitation Research Institute, 2Department of Correspondence to: H. Branch Coslett, University of Neurology, University of Pennsylvania School of Medicine Pennsylvania Medical Center, 3rd Floor Gates Bldg, and 3Department of Anesthesiology, Temple University 3400 Spruce St, Philadelphia, PA 19104-4283, USA School of Medicine, Philadelphia, Pennsylvania, USA E-mail hbc@mail.med.upenn.eduor Robert Friedman, Department of Anesthesiology,Temple University School of Medicine, 3401 N. Broad St.,Philadelphia, PA 19140, USAE-mail FRIEDMR@tuhs.temple.edu Summary
Some accounts of body representations postulate a real-
stimulus hand. We found that, as in previous investiga-
time representation of the body in space generated by
tions, participants’ response times (RTs) reflected the
proprioceptive, somatosensory, vestibular and other
degree of simulated movement as well as biomechanical
sensory inputs; this representation has often been termed
constraints of the arm. Importantly, a significant
the ‘body schema’. To examine whether the body schema
interaction between the magnitude of mental rotation and
is influenced by peripheral factors such as pain, we asked
limb was observed: RTs were longer for the painful arm
patients with chronic unilateral arm pain to determine
than for the unaffected arm for large-amplitude imagined
the laterality of pictured hands presented at different
movements; controls exhibited symmetrical RTs. These
orientations. Previous chronometric findings suggest that
findings suggest that the body schema is influenced by
performance on this task depends on the body schema,
pain and that this task may provide an objective measure
in that it appears to involve mentally rotating one’s hand
from its current position until it is aligned with the
Keywords: pain; body schema; plasticity; neglect; parietal
Abbreviations: RT ϭ response time, CRPS ϭ complex regional pain syndrome
Introduction
Take a moment to try the following experiment. With your
Classic neuropsychological observations led Head and eyes closed, place your left hand in front of you with the Holmes to postulate an on-line representation of body posture, palm up and fingers pointed straight ahead. Now rotate your or ‘body schema’, derived from multimodal sensory inputs hand until the fingers point first to your right and then to (including proprioceptive, vestibular, somatosensory and your left. For the purposes of this paper, there are two visual inputs) which interacted with motor systems and served noteworthy observations concerning your ability to perform to guide movements such as those in the above example the above tasks. First, the simple fact that such movements (Head and Holmes, 1911–1912). More recently, several lines are made effortlessly and without vision suggests that there of evidence have provided support for the postulated body must be an on-line mental representation of body posture. That schema and its role in the guidance of movement. For is, without real-time information concerning body position, it example, Cole and Paillard observed striking impairments in would be impossible to programme such efficient movements.
even routine movements, such as reaching towards an object Secondly, note that the movements to the right and left were or maintaining balance while sitting in a chair, for two not symmetrical: joint constraints on movement probably patients who were deprived of sensory input as a result of forced the movement to the left to be slower and more sensory neuropathy (Cole and Paillard, 1995). Interestingly, effortful than the movement to the right.
these patients were able to compensate partially for the lack of on-line information (e.g. proprioceptive) concerning body they involve lateral rather than medial mental rotations of posture by constant visual guidance of movements. Such the hand. These findings, in conjunction with those of Sirigu effortful compensation helps to highlight the normally and colleagues (Sirigu et al., 1995, 1996), suggest that both automatic and seamless interaction between the body schema actual and mentally simulated movements may depend on the body schema. Moreover, functional neuroimaging findings Neuropsychological evidence also suggests that the suggest that laterality judgements are associated with monitoring and updating of body position as well as the activation in motor and parietal areas which substantially ability to simulate body movements mentally may be impaired overlap with areas activated by actual movements (e.g.
in patients with parietal damage (Sirigu et al., 1995, 1996; Stephan et al., 1995; Porro et al., 1996; Parsons and Fox, Coslett, 1998; Wolpert et al., 1998; Schwoebel et al., 2001).
1998). Parsons and colleagues state ‘In summary, motor For example, Sirigu and colleagues observed strong imagery appears generally to involve the same movement correlations between the time to imagine and execute a series representation used by the executive motor processes— of finger movements for both normal subjects and patients a unitary representation of movements as they occur, in accordance with the physical laws underlying motor control movements of the affected limb were both slowed for patients and implementing all physiological and pathophysiological with motor cortex damage (Sirigu et al., 1995). However, constraints.’ (Parsons and Fox, 1998, p. 586).
imagined and executed movement times were poorly These and other (e.g. Lackner, 1988) lines of evidence correlated for patients with parietal damage (Sirigu et al., suggest that the body schema represents on-line information 1996). Taken together, these data suggest that the parietal concerning body posture and that it subserves both real and area is an integral component of the neural substrate for the imagined movements. Furthermore, the body schema appears body schema as it appears to be involved in monitoring to be sensitive to central insults that affect motor performance, the sensory and motor information necessary for accurate such as motor cortex lesions and basal ganglia dysfunction (Dominey et al., 1995; Sirigu et al., 1995). However, to Parsons further argues that the body schema underlies the our knowledge, few previous investigations have examined performance of normal participants on a task that requires whether there are peripheral factors that influence the them to judge the laterality of pictured hands (Parsons, body schema (for a discussion of neural plasticity in 1987a, b, 1994). On the basis of several lines of evidence, Ramachandran and Hirstein, 1998). The present experiment judgements by imagining their left hand moving into the was designed to determine if pain influences the body schema.
orientation of left-hand stimuli and their right hand moving More specifically, we examined whether performance on a into the orientation of right-hand stimuli. Furthermore, such modified version of the hand laterality judgement task developed by Parsons (Parsons, 1987a) would differ when representations of the contralateral hand. Consistent with the judgements involved mental rotations of affected and these suggestions, Parsons and colleagues (Parsons et al., unaffected limbs in patients suffering from chronic arm pain.
1998) found that the accuracy of laterality judgements wasunimpaired in split-brain patients when the stimulus handwas contralateral to the perceiving hemisphere (e.g. a left-hand stimulus presented in the left visual hemifield), but that performance was not above chance when the stimulus hand Participants
was ipsilateral to the perceiving hemisphere. Furthermore, Participants included 13 (six with right arm pain, seven with Parsons observed that the time required for such laterality left arm pain) patients with arm pain of at least 3 months judgements in normal participants increased as the stimulus duration. Patients were all diagnosed as suffering from hand was presented at orientations further from that of the complex regional pain syndrome (CRPS) and were referred participant’s hand: the time required to judge the laterality from a pain control centre, where they were undergoing of a palm-up stimulus hand was modified by whether a participant’s own hand was palm-up or palm-down and by characteristics, pain severity and treatment. The medications the degree of angular disparity between the stimulus hand prescribed for these patients had a broad range of effects, and the participant’s hand (Parsons, 1994). Strong correlations from the pharmacological relief of pain to the anti-seizure were also observed between the time required for hand effects of gabapentin and the antidepressant effects of laterality judgements and the time required for participants amitriptyline, but we were interested primarily in within- to actually align their hand with a stimulus. Importantly, subject comparisons of response times to left- and right-hand laterality judgement times were also found to reflect stimuli. Thus, the effects of medications are not likely to biomechanical constraints on movement. Thus, just as lateral account for any differences in response times to left- and movements away from the body’s midline are more effortful right-hand stimuli within a given patient. Eighteen right- and time-consuming than medial movements towards the handed, age-matched (mean age 47 years, SD 11 years) midline, hand laterality judgement times are also longer when participants served as controls. Testing was approved by Table 1 Clinical details of patients with complex regional pain syndrome
Morphine, amitriptyline clonazepam, tramadol Codeine, gabapentin,rofecoxib, paracetamol VAS ϭ visual analogue scale: pain severity rated from 0 ϭ no pain to 10 ϭ worst pain experienced. F ϭ female; M ϭ male.
Fig. 1 Examples of left-hand stimulus in the palm-down view at orientations of 0°, 90° medial, 90° lateral and 180°.
the Internal Review Board of Temple University and the Design and procedure
For each patient, the 16 different stimuli (8 conditions ϫ 2 hands) were presented eight times to give a total of 128trials. Controls viewed each stimulus four times in a total of64 trials.
Stimuli
Participants sat with their hands resting palm-down on the Digitized pictures of a left or right hand were presented on table in front of them with fingers resting on the response a computer monitor in palm-up and palm-down views at 0° keys. On each trial, a single stimulus hand appeared centred on (facing up), 90° medial (facing towards the participant’s the computer screen and remained there until the participant midsagittal plane), 90° lateral (facing away from participant’s indicated the laterality of the hand by pressing a key. For midsagittal plane) and 180° (facing down) orientations patients, responses were made by pressing a left or right key with the index or middle finger of their unaffected limb.
Thus, for both right and left hands there were a total of Controls responded with the index and middle fingers of eight different stimuli. All stimuli were created by digitally their right hand. All participants were instructed to respond manipulating one picture of a palm-up view and one picture as quickly and accurately as possible. Psyscope software of a palm-down view of the same hand in order to ensure (Cohen et al., 1993) was used to generate a random order that each stimulus was identical except for the change in of stimulus presentation for each participant and to record response time (RT) and accuracy data.
Fig. 2 Mean response times for laterality judgements involving
Fig. 3 Mean response times for laterality judgements involving
the affected and unaffected limbs of patients for palm-down the affected and unaffected limbs of patients for palm-up stimuli stimuli in the four orientation conditions.
[F(1,12) ϭ 7.31, P Ͻ 0.03], indicating slower RTs for mental A 2 (limb: affected and unaffected arms) ϫ 2 (view: palm rotations of the affected than the unaffected limb. This effect up and palm down) ϫ 4 (orientation: 0°, 90° medial, 90° appears to be driven by slower RTs for the affected limb in lateral and 180°) repeated measures analysis of variance was the 180° condition, as indicated by a significant interaction used to analyse RT and accuracy data separately for patients between orientation and limb [F(3,36) ϭ 8.12, P Ͻ 0.001].
and controls (for controls, limb refers to the left and right Planned comparisons yielded significant differences between arms). Analyses of RT included only data for correct RTs involving the affected and unaffected limbs for the 180° responses. RTs Ͼ2 SD above each participant’s grand mean conditions [palm down, F(1, 12) ϭ 9.58, P Ͻ 0.009; palm were also excluded from analyses (Ratcliff, 1993), resulting up, F(1,12) ϭ 7.35, P Ͻ 0.02].
in the loss of 5% of trials for both patients and controls. For This effect was also consistent across patients. Twelve of both patients and controls, RT outliers were distributed the 13 patients exhibited slower RTs for the affected limb equally across responses involving motor imagery of the left compared with the unaffected limb for the 180° condition and right (affected and unaffected) hands, but were more (sign test, P Ͻ 0.002). On average, mental rotations of the likely to occur when the disparity between the stimulus hand affected limb were 1123 ms (SD ϭ 1136 ms) slower than and the participant’s hand was greatest (i.e. the 180° palm- those of the unaffected limb in the 180° condition.
Accuracy data for patients and controls are presented in Table 2. Consistent with the RT data, analyses of accuracyrevealed a significant main effect of orientation [F(3,36) ϭ Patients
19.29, P Ͻ 0.001] and a significant interaction between As indicated in Figs 2 and 3, there was a significant main orientation and view [F(3,36) ϭ 7.07, P Ͻ 0.001], indicating effect of orientation [F(3,36) ϭ 22.01, P Ͻ 0.001], such that that accuracy reflected the disparities between stimulus and RTs increased as the difference between the orientation of participant hand postures as well as the different constraints the participant’s hand and the stimulus hand increased. There on palm-down and palm-up rotations of the hand. There was was also a significant interaction between orientation and no significant main effect of limb [F(1,12) Ͻ 1], suggesting view [F(3,36) ϭ 5.44, P Ͻ 0.005], indicating that RTs an absence of speed–accuracy trade-offs. Overall accuracy reflected the disparities between stimulus hands and the for the affected and unaffected limbs was 84 and 83%, participant’s own hand postures (i.e. 0° orientation and palm down) as well as the different movement constraints forpalm-up and palm-down rotations of the hand. Furthermore,consistent with biomechanical constraints on medial and Controls
lateral movements, RTs were significantly longer for palm- As indicated in Figs 4 and 5, there was a significant main up views of 90° lateral compared with 90° medial stimuli effect of orientation [F(3,51) ϭ 19.95, P Ͻ 0.001] and view [palm down, F(1,12) ϭ 1.92, P Ͻ 0.19; palm up, F(1,12) ϭ [F(1,17) ϭ 22.73, P Ͻ 0.001] and an interaction between orientation and view [F(3,51) ϭ 8.76, P Ͻ 0.001], suggesting Of greatest import was the significant main effect of limb that, consistent with the patient data, RTs reflected the Fig. 4 Mean response times for laterality judgements involving
Fig. 5 Mean response times for laterality judgements involving
the left and right limbs of controls for palm-down stimuli in the the left and right limbs of controls for palm-up stimuli in the four Table 2 Mean accuracy (proportion correct) for patients
no significant effect of limb. Eleven of 18 controls exhibited slower RTs for the left than for the right limb for the 180° condition (sign test, P Ͻ 0.12). RTs for the left limb were,on average, 74 ms slower (SD ϭ 781) than those for the Consistent with the RT data, analyses of accuracy indicated a significant main effect of orientation [F(3,51) ϭ 10.13, Ͻ 0.001] and a significant interaction between orientation and view [F(3,51) ϭ 7.29, P Ͻ 0.001], indicating that accuracy reflected the disparities between stimulus and participant hand positions and the different movement constraints on palm-up and palm-down rotations of the hand.
There was no significant main effect of limb [F(1,17) suggesting the absence of speed–accuracy trade-offs. Overall accuracy for the left and right limbs was 86 and 87%, Discussion
Consistent with previous investigations (Parsons, 1987a, b,
disparities between stimulus and participant hand positions 1994; Parsons and Fox, 1998; Parsons et al., 1998), analyses and the different movement constraints on palm-up and palm- for both patients and controls demonstrated that RTs and down rotations of the hand. Furthermore, there were longer accuracy were significantly influenced by the degree of RTs for palm-up views of lateral than medial stimuli [palm imagined movement necessary to align participants’ hands down, F(1,17) ϭ 2.19, P Ͻ 0.16; palm up, F(1,17) ϭ 37.03, with stimuli. Furthermore, RTs were consistent with P Ͻ 0.001], indicating that RTs reflected the biomechanical previously observed biomechanical constraints on movement (e.g. Parsons, 1994). These findings support the suggestion that the body schema underlies performance on the hand significance [F(1,17) ϭ 3.83, P Ͻ 0.10], reflecting a small laterality task. The major, and novel, finding of the present but consistent advantage for responses involving the right investigation is that a brain representation of the body is (i.e. dominant) hand, there was no significant interaction influenced by pain. Patients, but not controls, exhibited between orientation and limb [F(3,51) Ͻ 1]. Thus, in contrast slowed RTs when responses required large-amplitude mental to the consistently slower RTs observed for the affected rotations of their affected relative to their unaffected arm.
relative to the unaffected limbs in the 180° condition for the Before we discuss the theoretical implications of these patient data (i.e. 12 out of 13), control participants exhibited data, it is important to emphasize that these findings cannot be attributed to pain inhibition of movement or ‘guarding’ severity of pain at the time of testing. This hypothesis is for several reasons. First, patients did not, in fact, move the painful arm during the experimental task. Nor did patients The sensory and motor remappings reported in patients report pain in the course of the imagined movements.
and animals following amputation or deafferentation may Secondly, the slowing of RTs for the painful arm was also have an anatomical underpinning that differs from the observed only in the 180° condition; if guarding were elicited alteration induced in the body schema by pain. Investigations automatically by any stimulus depicting the painful extremity, in animals (Merzenich et al., 1984; Pons et al., 1991) as one would have expected the slowing of RTs to be observed well as magnetoencephalogram and transcranial magnetic for stimuli in all four orientations. We postulate that slowing stimulation experiments in humans (Cohen et al., 1991;Ramachandran, 1993; Yang et al., 1994; Pascual-Leone et al., in the 180° condition occurred because, unlike the other 1996) suggest that the remapping after amputation occurs at conditions, the 180° condition required large-amplitude the level of the primary sensory and motor cortices. In simulated movements at both distal and proximal joints and contrast, PET investigations of what we term the ‘body was thus more likely to involve painful regions of the arm schema’ in humans have suggested that this representation that tended to include both the elbow and the shoulder.
is supported by posterior parietal and dorsolateral frontal cortices (Bonda et al., 1996; Parsons and Fox, 1998). As a demonstrations that pathological conditions may alter the PET study using the hand laterality task reported in the body schema. As previously noted, Sirigu and colleagues present paper demonstrated activation in parietal cortex reported data from patients with parietal lesions, demonstrat- (Parsons and Fox, 1998), we believe that the pain-induced ing that central lesions might disrupt the body schema (Sirigu alteration in the body schema was likely to be mediated by et al., 1996). Furthermore, using a task similar to that reported higher-level sensory cortices of the posterior superior parietal here, Coslett demonstrated that patients with right-hemisphere lobes rather than the primary sensory cortex.
lesions resulting in left neglect, but not other patients with Working with patients suffering from CRPS also called right hemisphere lesions, exhibited an impaired ability to reflex sympathetic dystrophy, Galer and colleagues (Galer identify pictures of left compared with right hands (Coslett, et al., 1995; Galer and Jensen, 1999; for a similar account, 1998). In the light of previous evidence suggesting that the see Rommel et al., 1999) have also suggested that the identification of pictured hands depends on the body schema frequently observed reduction in movement associated with (Parsons, 1987a, b, 1994), this asymmetrical performance this syndrome may be attributable to a central neglect- suggests that at least some features of the neglect syndrome like disorder. Their conclusions, however, were based on may be attributable to disruption of the body schema.
observations of movement and questionnaire data, raising thepossibility that the findings were attributable to guarding. As The claim that a central representation of the body such noted previously, the findings from the present experiment, as the body schema may also be altered by ‘peripheral’ in which participants did not move or report pain, cannot factors is not without precedent. This phenomenon has been readily be attributed to this factor.
investigated most extensively in patients with phantom limbs.
In addition to the theoretical implications of the present As noted by Ramachandran and Hirstein in a recent review, findings, we note that the hand laterality task, after further several lines of evidence suggest that, in both animals examination and modification, may also be of clinical value.
and humans, primary sensory and motor cortices may be As a blind and objective measure, it may be that performance ‘remapped’ after amputation or deafferentation of a body on the hand laterality task could provide a more reliable and part (Ramachandran and Hirstein, 1998). However, the valid measure of pain than the currently used self-reported alteration in the body schema exhibited by our patients may ratings of pain. Furthermore, response time on the hand differ from that exhibited by patients with phantom limbs.
laterality task may prove to be a more sensitive measure of Whereas amputation or deafferentation may be expected to changes in pain than subjective ratings. Finally, we note that induce long-standing or even permanent changes (but see the use of the hand laterality task as an assessment tool need Ramachandran, 1993), the changes in the body schema not be limited to patients suffering from CRPS, but can be associated with chronic pain may reflect the current state of used as an assessment or screening tool for diverse patient nociceptive (and other sensory) feedback. In this sense, the populations suffering from motor impairments.
alteration in the body schema exhibited by our patients mayapproximate more closely the distortions of body representa- Acknowledgements
tions observed when inconsistencies are induced between We wish to thank Jena Friedman for her valued contribution multiple sensory inputs (Ramachandran and Hirstein, 1998) as a research assistant. This work was supported by NIH and tactile or muscle-stretch inputs (Lackner, 1988). If, as grant R01 NS37920-02 awarded to H.B.C.
we have argued elsewhere (Coslett, 1998; Buxbaum andCoslett, 2001; Schwoebel et al., 2001), the body schema is References
an on-line, real-time representation of the position and Bonda E, Frey S, Petrides M. Evidence for a dorso-medial parietal possible actions of the body, one might expect the changes we have observed to be influenced significantly by the Buxbaum LJ, Coslett HB. Specialised structural descriptions for lateralized mental representations of hand shape and movement.
human body parts: evidence from autotopagnosia. Cogn Neuro- Pascual-Leone A, Peris M, Tormos JM, Pascual AP, Catala MD.
Cohen JO, MacWhinney B, Flatt M, Provost J. Psyscope: A Reorganization of human cortical motor output maps following new graphic interacting environment for designing psychology traumatic forearm amputation. Neuroreport 1996; 7: 2068–70.
experiments. Behav Res Meth Instr Comput 1993; 25: 257–71.
Pons TP, Garraghty PE, Ommaya AK, Kaas JH, Taub E, Mishkin M.
Cohen LG, Bandinelli S, Findlay TW, Hallett M. Motor Massive cortical reorganization after sensory deafferentation in adult reorganization after upper limb amputation in man. Brain 1991; macaques. Science 1991; 252: 1857–60.
Porro CA, Francescato MP, Cettolo V, Diamond ME, Baraldi P, Cole J, Paillard J. Living without touch and peripheral information Zuiani C, et al. Primary motor and sensory cortex activation during about body position and movement: studies with deafferented motor performance and motor imagery: a functional magnetic subjects. In: Bermudez JL, Marcel A, Eilan N, editors. The body resonance imaging study. J Neurosci 1996; 16: 7688–98.
and the self. Cambridge (MA): MIT Press; 1995. p. 245–66.
Ramachandran VS. Behavioral and magnetoencephalographic Coslett HB. Evidence for a disturbance of the body schema in correlates of plasticity in the adult human brain. [Review]. Proc neglect. Brain Cogn 1998; 37: 527–44.
Natl Acad Sci USA 1993; 90: 10413–20.
Dominey P, Decety J, Broussolle E, Chazot G, Jeannerod M. Motor Ramachandran VS, Hirstein W. The perception of phantom limbs: imagery of a lateralized sequential task is asymmetrically slowed the D. O. Hebb lecture. [Review]. Brain 1998; 121: 1603–30.
in hemi-Parkinson’s patients. Neuropsychologia 1995; 33: 727–41.
Ratcliff R. Methods for dealing with reaction time outliers. Psychol Galer BS Jensen M. Neglect-like symptoms in complex regional pain syndrome: Results of a self-administered survey. J Pain Rommel O, Gehling M, Dertwinkel R, Witscher K, Zenz M, Malin J-P, et al. Hemisensory impairment in patients with complex regional Galer BS, Butler S, Jensen MP. Case reports and hypothesis: a pain syndrome. Pain 1999; 80: 95–101.
neglect-like syndrome may be responsible for the motor disturbance Schwoebel J, Coslett HB, Buxbaum LJ. Compensatory coding of in reflex sympathetic dystrophy (complex regional pain syndrome-1).
body part location in autotopagnosia: evidence for extrinsic [Review]. J Pain Symptom Manage 1995; 10: 385–91.
egocentric coding. Cogn Neuropsychol 2001; 18: 363–81.
Head H, Holmes G. Sensory disturbances from cerebral lesions.
Sirigu A, Cohen L, Duhamel JR, Pillon B, Dubois B, Agid Y.
Congruent unilateral impairments for real and imagined hand Lackner JR. Some proprioceptive influences on the perceptual movements. Neuroreport 1995; 6: 997–1001.
representation of body shape and orientation. Brain 1988; 111: Sirigu A, Duhamel JR, Cohen L, Pillon B, Dubois B, Agid Y. The mental representation of hand movements after parietal cortex damage. Science 1996; 273: 1564–8.
Schoppmann A, Zook JM. Somatosensory cortical map changes Stephan KM, Fink GR, Passingham RE, Silbersweig D, Ceballos- following digit amputation in adult monkeys. J Comp Neurol 1984; Baumann AO, Frith CD, et al. Functional anatomy of the mental representation of upper extremity movements in healthy subjects.
Parsons LM. Imagined spatial transformations of one’s hands and feet. Cogn Psychol 1987a; 19: 178–241.
Wolpert DM, Goodbody SJ, Husain M. Maintaining internal Parsons LM. Imagined spatial transformation of one’s body. J Exp representations: the role of the human superior parietal lobe. Nat Parsons LM. Temporal and kinematic properties of motor behavior Yang TT, Gallen CC, Ramachandran VS, Cobb S, Schwartz BJ, reflected in mentally simulated action. J Exp Psychol Hum Percept Bloom FE. Noninvasive detection of cerebral plasticity in adult human somatosensory cortex. Neuroreport 1994; 5: 701–4.
Parsons LM, Fox PT. The neural basis of implicit movements usedin recognising hand shape. Cogn Neuropsychol 1998; 15: 583–615.
Received January 25, 2001. Revised May 17, 2001. Parsons LM, Gabrieli JD, Phelps EA, Gazzaniga MS. Cerebrally

Source: http://wexler.free.fr/library/files/schwoebel%20(2001)%20pain%20and%20the%20body%20schema.%20evidence%20for%20peripheral%20effects%20on%20mental%20representations%20of%20movement.pdf