Abstract

Phantom limb sensations may be linked to motor activities in the deafferented cortices of amputees, with artificial visual feedback of an amputated limb leading to enhanced phantom sensations. The present study was designed to verify if cortical motor activity related to an amputated limb can be triggered by visual input using an objective behavioral measure and with a neurophysiological correlate. Trauma amputees and normally limbed subjects showed superior performance in a mirror-drawing task when the mirror was placed sagittally (giving visual feedback of the amputated/inactive limb) compared with when it was placed frontally. Measurement of lateralized movement-related brain potentials showed that, under the lateral mirror condition, contralateral motor activity of the viewed hand was observed in both normal subjects and trauma amputees. In contrast, this activity was not observed in subjects with congenital limb absence. These findings suggest that, in traumatic amputees, motor enhancement due to visualization of the movements of the missing limb reflects the effectiveness of motor commands to the missing limb, strengthening the hypothesis of the functional survival of deafferented cortical motor areas.

INTRODUCTION

Almost immediately after the loss of a limb, more than 90% of patients experience vivid phantom sensations in the amputated limb (Ramachandran & Hirstein, 1998). This phantom limb phenomenon refers to the illusory impression that an amputated limb is still present: The specificity of normal somatosensory sensations associated with phantom limbs is preserved, as is the ability at times to move the missing limb. Although the neurological cause of this illusion is unclear, understanding of this perceptual phenomenon may provide valuable insight into the mechanisms underlying body representation in relation to its neurobiological substrate. Recent brain-mapping studies using fMRI, PET, ERP, and TMS indicate that the cortical representation of a missing limb may survive in the brains of amputees. Indeed, TMS (Mercier, Reilly, Vargas, Aballea, & Sirigu, 2006) and fMRI and PET scan studies (Roux et al., 2003) have shown that the cortical area previously devoted to a now-amputated hand could retain its original function of controlling hand (or phantom hand) movements, despite the absence of muscles normally targeted by the cortical outputs. Thus, despite dramatic reorganization at the neural level (Karl, Birbaumer, Lutzenberger, Cohen, & Flor, 2001; Flor et al., 1998; Pascual-Leone, Peris, Tormos, Pascual, & Catala, 1996; Ramachandran, Stewart, & Rogers-Ramachandran, 1992), body representation appears to be relatively maintained throughout life, even after drastic changes such as limb amputation.

Studies using a mirror box (Ramachandran & Rogers-Ramachandran, 1996) also provide evidence for maintenance of motor representation of a missing limb in the brains of amputees. In those studies, a mirror is oriented at the midsagittal plane, with the reflective surface facing the intact limb. Thus, the hand in the mirror occupies the place of the opposite hand in normally limbed subjects. For amputees, the reflection provides visual feedback of the missing limb. This illusion can have positive consequences in some amputees with painful paralyzed phantom limbs. Vision of the missing limb provides congruent motor (including proprioception) and visual information about that limb. In normally limbed subjects, studies focusing on the interactions between these two sensory inputs have shown that stimulation in one modality can have significant effects on the perception in a second modality (Berberovic & Mattingley, 2003; Michel, Pisella, et al., 2003; Michel, Rossetti, Rode, & Tilikete, 2003; Rossetti, Koga, & Mano, 1993; Hay, Pick, & Ikeda, 1965). Clinical data in amputees may reflect the persistence of this visuoproprioceptive integration related to the missing arm, as seen in the mirror. At a neural level, this could reflect activation of motor areas (including proprioceptive areas) related to the missing limb. This hypothesis was supported by the behavioral and physiological observation of patients with brachial plexus avulsion (Giraux & Sirigu, 2003). A visuomotor training program in which the patients were required to match “voluntary” movements of the paralyzed/phantom limb with movements of a virtual hand reduced phantom pain sensations. These changes were correlated with increased activity in the contralateral M1 region.

Brain mapping and clinical studies with amputees suggest that activation of sensorimotor cerebral regions may be triggered by visual feedback or mental imagery of the amputated limb and account for phantom limb sensations (Mercier et al., 2006; Roux et al., 2003; Ramachandran & Rogers-Ramachandran, 1996). In these studies, however, activation of cortical motor areas was examined solely in relation to sensed or imagined movements of the missing limb, with no information on behavioral performance in amputated patients. The aim of the present study was to determine (a) whether cortical motor areas can be activated by visual inputs from the corresponding limb, and therefore (b) whether cortical activations corresponding to the missing limb can affect motor performances of the remaining limb. On the basis of studies indicating that the image of the body and its neurobiological substrate develops only with experience (Montoya et al., 1998; Melzack, 1990), we also expected neurofunctional and motor differences between trauma amputees and subjects with congenital limb absence, in which cortical motor areas in relation to their missing limb are not or are less functional. Research on plasticity in somatosensory and motor areas has emphasized the role of critical periods during early developmental stages. Animal studies suggest that the formation of cortical and subcortical maps may be altered by peripheral damage of a limb during the prenatal period (Rhoades, Chiaia, Bennett-Clarke, Janas, & Fisher, 1994) as well as lesions or deprivation during development (Killackey, Rhoades, & Bennett-Clarke, 1995).

We have described a method to assess functional survival and activation of deafferented cortical motor areas using a visuomotor performance test based on visual feedback from the missing limb. Using a mirror, amputees received visual feedback of an intact hand while performing a motor task. The mirror was placed either in front of the subject, generating an inverted vision of the drawing hand, or in the midsagittal plane, giving the illusion that the missing hand was performing the motor task, as described in the mirror-box paradigm (Ramachandran & Rogers-Ramachandran, 1996).

We anticipated that visual feedback of the missing hand through a midsagittally oriented mirror would activate proprioception of the missing/inactive hand in trauma amputees and normally limbed subjects, who should perform the task better than when receiving visual feedback of the performing hand when the mirror is oriented frontally. By contrast, we anticipated that, in subjects with congenital limb absence who have never experienced sight of the missing limb, a lateral mirror view of their drawing hands should not facilitate drawing performance when compared with a frontal mirror view.

In addition, and to further confirm the hypothesis of motor activation related to the seen limb, we also examined lateralized brain motor activity while subjects performed motor tasks using frontal or lateral visual feedback as described above. To determine whether vision influenced motor preparation, we analyzed the ERP component called the lateralized readiness potential (LRP), which reflects preparation of the responding hand. We anticipated that visual feedback from the responding hand would lead to cortical activity contralateral to the missing/inactive hand in normally limbed and trauma amputee subjects but not in congenital limb absence subjects.

EXPERIMENT 1

We focused on the objective measures of possible visuoproprioceptive persistence in the missing limb of trauma amputees, independent of any factor that could influence analysis such as phantom sensations, pain, medication, or presence of prosthesis. Hence, we measured motor performances of the remaining hand in a mirror-drawing task.

Methods

Participants

Four subjects with traumatic upper-limb loss and four subjects with congenital upper-limb absence were enrolled. Demographic and clinical data are summarized in Table 1. A third group of 26 normally limbed subjects was also enrolled (20 men and 6 women, mean age = 33.64 years). All participants were naive to the purpose of the study. Subjects gave informed consent before participation, in accordance with the guidelines of our local ethics committee, which approved the study.

Table 1. 

Clinical Features of the Amputee Groups for Experiments 1 and 2

Subjects, Sex
Age (Years)
Side of Amputation
Year of Amputation
Handedness before Amputation
Sensations
Pain
Prothesis
A. Main clinical features of the two amputee groups enrolled in the mirror-drawing task 
FV, M 29 Left hand 1998 Right-handed Yes Yes Aesthetic 
SS, M 50 Right arm 2003 Right-handed Yes Yes No 
EP, M 44 Right arm 1998 Right-handed Yes Sometimes No 
DT, M 50 Left arm 1997 Right-handed Sometimes Sometimes Myoelectric 
SR, F 30 Right arm Congenital – Never Never No 
LF, F 22 Left forearm Congenital – Never Never No 
NT, F 34 Left forearm Congenital – Never Never No 
TO, F 28 Right arm Congenital – Sometimes Never Aesthetic 
 
B. Main clinical features of trauma amputee and congenital limb absence groups enrolled in the electrophysiological experiment 
NB, F 37 Right hand 2000 Right-handed Yes Sometimes Aesthetic 
FV, M 29 Left hand 1998 Right-handed Yes Yes Aesthetic 
JM, M 46 Left arm 1978 Right-handed Yes Sometimes Aesthetic 
GS, M 59 Right arm 1968 Right-handed Yes Sometimes No 
LF, F 22 Left forearm Congenital – Never Never No 
MA, M 33 Right forearm Congenital – Sometimes Never No 
NT, F 34 Left forearm Congenital – Never Never No 
SR, F 30 Right arm Congenital – Never Never No 
Subjects, Sex
Age (Years)
Side of Amputation
Year of Amputation
Handedness before Amputation
Sensations
Pain
Prothesis
A. Main clinical features of the two amputee groups enrolled in the mirror-drawing task 
FV, M 29 Left hand 1998 Right-handed Yes Yes Aesthetic 
SS, M 50 Right arm 2003 Right-handed Yes Yes No 
EP, M 44 Right arm 1998 Right-handed Yes Sometimes No 
DT, M 50 Left arm 1997 Right-handed Sometimes Sometimes Myoelectric 
SR, F 30 Right arm Congenital – Never Never No 
LF, F 22 Left forearm Congenital – Never Never No 
NT, F 34 Left forearm Congenital – Never Never No 
TO, F 28 Right arm Congenital – Sometimes Never Aesthetic 
 
B. Main clinical features of trauma amputee and congenital limb absence groups enrolled in the electrophysiological experiment 
NB, F 37 Right hand 2000 Right-handed Yes Sometimes Aesthetic 
FV, M 29 Left hand 1998 Right-handed Yes Yes Aesthetic 
JM, M 46 Left arm 1978 Right-handed Yes Sometimes Aesthetic 
GS, M 59 Right arm 1968 Right-handed Yes Sometimes No 
LF, F 22 Left forearm Congenital – Never Never No 
MA, M 33 Right forearm Congenital – Sometimes Never No 
NT, F 34 Left forearm Congenital – Never Never No 
SR, F 30 Right arm Congenital – Never Never No 

Materials

The mirror-drawing apparatus was a wooden box with a vertical mirror facing a star pattern. Metallic plates were placed side by side on each of the 12 segments of the star (see Figure 1). Drawing was performed with a digital pen, which did not leave a visible trace, connected to a computer. A cover prevented direct vision of the drawing hand. The following data were collected: (a) the total time necessary to complete the star pattern (12 segments; segment width = 0.5 cm, segment length = 3.6 cm); (b) the number of errors, recorded each time the digital pen touched the metallic plates surrounding the star segments; and (c) the total time the digital pen was in contact with the metallic plates (another indicator of error).

Figure 1. 

The experimental setup for the mirror-drawing task. (A) The subject faced the mirror to obtain frontal visual feedback of the drawing hand. (B) The mirror was placed sagittally and provided a view of the nondrawing/missing hand. An opaque box positioned over the star pattern prevented the subject from having a direct view of the performing hand.

Figure 1. 

The experimental setup for the mirror-drawing task. (A) The subject faced the mirror to obtain frontal visual feedback of the drawing hand. (B) The mirror was placed sagittally and provided a view of the nondrawing/missing hand. An opaque box positioned over the star pattern prevented the subject from having a direct view of the performing hand.

Procedure

Subjects participated in two successive sessions, each lasting approximately 5 minutes. Participants were seated at a table in front of the mirror-drawing apparatus. In the first session, the mirror was oriented in the midsagittal plane (lateral feedback). In the second session, the mirror was oriented in the frontal plane (frontal feedback). While viewing the image through the mirror, subjects were instructed to trace the six-pointed star pattern as rapidly and as accurately as possible with their intact/dominant hand. Half the participants within each group completed the first session under lateral feedback conditions, whereas the other half completed the first session under frontal feedback conditions.

Results

Normally Limbed Subjects

A greater number of normally limbed subjects (compared with trauma amputees and congenital limb absence subjects) were recruited to test a possible order effect between the two experimental conditions. Previous research has shown that, in a mirror-drawing task with the mirror in the frontal plane, scores improve from the first to the second trial (Lajoie et al., 1992). Although in this study the two successive trials were not performed under equivalent conditions (i.e., lateral vs. frontal mirror placement), the training effects remain an issue and, more importantly, the training effects may differ according to which mirror position is used first. Results from normally limbed subjects showed that mirror order presentation had no effect on drawing time, t(24) = 1.51, p = .14, or error frequency, t(24) = −0.82, p = .42.

Time and Error Indices

Drawing performances were evaluated in terms of time and errors. The total number of errors was weighted for error duration for each subject. Hence, a high error score may reflect either a large number of brief touches or a small number of long touches of the star borders.

Data Transformation

The difference between the total time needed to complete the drawing under frontal and lateral feedback conditions (Δt) was calculated; a positive Δt value indicated that the time to complete the task was longer with frontal feedback than with lateral feedback. Error index was calculated similarly, yielding a single Δe value for each subject. These performance evolution indices were preferred to raw data for two major reasons. First, the sample sizes of normally limbed and amputee subjects differed. Variance homogeneity and normality could, therefore, not be tested, and nonparametric tests appeared to be more appropriate for evaluating the significance of performance differences. However, interactions cannot be evaluated using nonparametrical tests, only between- or within-group effects. Second, time performances in the mirror-drawing task were highly variable within each group. Hence, transforming two scores into a single signed score resulted in a lower degree of freedom and a greater statistical power.

The mean Δt values for each group are shown in Figure 2A. Trauma amputees and normally limbed subjects showed positive Δt values, indicating that the task took longer under frontal rather than lateral feedback conditions. We also found that Δt values differed among the three groups [H(2,34) = 8.47, p = .014, Kruskal–Wallis ANOVA]. Paired comparisons (Mann–Whitney test) showed that Δt values differed significantly between normally limbed and congenital limb absence subjects (z = 2.56, p = .01) and between trauma amputees and congenital limb absence subjects (z = −2.31, p = .02). This latter probability is close to the adjusted threshold of 0.017 for three comparisons (Bonferroni correction). In contrast, Δt values were similar for normally limbed and trauma amputee groups. The robustness of the feedback condition effect was confirmed by analysis of individual data. All four trauma amputees had shorter task completion times with lateral compared with frontal feedback. Conversely, all four congenital limb absence subjects had shorter task completion times with frontal compared with lateral feedback. Eighteen of the 26 normally limbed subjects had shorter task completion times with lateral compared with frontal feedback (p = .02, binomial test).

Figure 2. 

Performances. (A) Mean drawing time differences (Δt) under frontal and lateral feedback conditions (raw data are given for each condition). Positive (negative) differences indicate longer (shorter) completion times under frontal feedback than under lateral feedback. (B) Mean differences between error indices under frontal and lateral feedback conditions (Δe = error frequency × error duration).

Figure 2. 

Performances. (A) Mean drawing time differences (Δt) under frontal and lateral feedback conditions (raw data are given for each condition). Positive (negative) differences indicate longer (shorter) completion times under frontal feedback than under lateral feedback. (B) Mean differences between error indices under frontal and lateral feedback conditions (Δe = error frequency × error duration).

Mean Δe values for each group are shown in Figure 2B. Mean Δe values differed between the three groups, H(2) = 6.02, p = .048, Kruskal–Wallis ANOVA. Although paired comparisons (Mann–Whitney test) suggested different Δe values between normally limbed and congenital limb absence subjects (z = 1.89, p = .058) and between congenital limb absence subjects and trauma amputees (z = −1.73, p = .083), Δe values did not differ between normally limbed subjects and trauma amputees. Individual Δe values showed the same distribution as individual Δt values.

EXPERIMENT 2

The findings shown above indicate that a view of the missing or inactive hand can improve mirror-drawing task performances in trauma amputees and healthy controls. We hypothesized that this enhancement may be due to activation of the motor cortex of the seen hand. Thus, in both groups, ipsilateral motor-cortical activity of the active/intact hand would be anticipated because movement is seen in a sagittally positioned mirror. In the same manner, individuals with congenital limb absence did not show enhanced performance when the mirror was placed in the sagittal plane. Thus, we hypothesized that, in these subjects, vision of the missing hand would not generate activity of the motor cortex ipsilateral of the performing hand. To test these hypotheses, we examined LRP, a measurement reflecting contralateral hand preparation. LRP was chosen because neuroanatomical evidence from surface and depth electrode recordings has shown that this component is generated, at least in part, in the primary motor cortex (Gemba, Sasaki, & Tsujimoto, 1990; Okada, Williamson, & Kaufman, 1982; Vaughan, Costa, & Ritter, 1968). LRP corresponds to an increasing negative asymmetric scalp potential over the hemisphere contralateral to the movement side. This premovement lateralized negativity reflects differential involvement of the left and right motor cortices in preparing to execute unimanual motor acts. Because motor asymmetry may be overlapped by a variety of asymmetries related to other functional or structural differences between hemispheres, motor asymmetry has been defined as that part of the total potential recorded between opposite scalp locations above the motor cortex that reverses when the response hand is reversed and everything else is kept constant. Thus, asymmetry of motor-cortical activation can be obtained by summing total activation asymmetry of the two cerebral hemispheres obtained for right- and left-hand responses: (C3 − C4) right-hand response + (C4 − C3) left-hand response. This is equivalent to subtracting the opposite asymmetry obtained for the left hand from the asymmetry obtained for the right hand: (C3 − C4) right-hand response − (C3 − C4) left-hand response, with C3 and C4 corresponding to the electrode site above the hand area of the left and right motor cortices, respectively.

Although LRP is usually computed from tasks involving both hands, we computed LRP from a task involving only one hand with a direct visual feedback of that hand and a visual feedback given by a sagittally placed mirror, which gave the impression of seeing the opposite/missing hand. Although traditional LRP was computed as the difference between right- and left-hand cortical activities, we computed LRP by comparing right/intact hand cortical activities while subjects had a direct view of their right/intact hand with right/intact hand cortical activity and while subjects had a visual feedback of the hand in the sagittally placed mirror (giving an image of the left/missing hand). Although the same hand is performing under both conditions, if vision of the performing hand influences motor activity, a lateralized cortical motor activity should emerge from this comparison. Conversely, if vision does not influence motor activity, the two waveforms should be identical because the motor task is identical and no LRP should emerge.

The LRP component reflects response preparation, thus commencing before muscle contraction begins, and can occur in the absence of an overt response (Minelli, Marzi, & Girelli, 2007; Galdo-Alvarez & Carrillo-de-la-Pena, 2004; Miller & Hackley, 1992; de Jong, Wierda, Mulder, & Mulder, 1988). This component is, therefore, applicable to amputees. If cortical motor representation survives in their brains, an LRP should be recorded, even in the absence of muscles targeted by the cortical outputs, and motor performances of the responding hand, as shown in Experiment 1, would be enhanced under lateral visual feedback conditions by cortical motor activity of the inactive/amputated hand. That is, we anticipated lateralized motor preparation initiated by a mirror-reflected view of the nonresponding/missing hand.

Methods

Participants

Four subjects with traumatic upper-limb loss and four subjects with congenital upper-limb absence were enrolled. Demographic and clinical data are summarized in Table 1. A third group of eight normally limbed subjects (four men, mean age = 31.5 years) was also enrolled. All normally limbed subjects except one were right-handed according to self-reporting. Subjects gave informed consent before participation, in accordance with the guidelines of our local ethics committee, which approved the study.

Materials and Procedure

Subjects performed a forced-choice task unimanually under three different conditions defined by different views of the responding hand. Subjects were seated at a table and, using the intact/right hand, were asked to press as quickly as possible one of two response keys placed beside two light-emitting diodes that were switched on randomly for 200 msec. The two light-emitting diodes were placed 3 cm apart on the table (Figure 3). Conditions were identical except for the type of visual feedback of the responding hand.

  • (1) 

    In the “direct view” condition, participants performed the task with a direct view of the performing hand (intact/right hand).

  • (2) 

    In the “sagittal mirror view” condition, subjects performed the task with a mirror placed sagittally, which gave the impression of a direct view of the missing/inactive hand.

  • (3) 

    In the “frontal mirror view” condition, the mirror was placed frontally giving visual frontal feedback of the performing hand.

Figure 3. 

Experimental setup for the mirror task in Experiment 2. (A) Lateral mirror condition: The mirror placed sagittally provided a view of the nonmoving/missing hand, hidden behind the mirror. (B) Frontal mirror condition: The mirror provided a frontal visual feedback of the performing hand. An opaque cover (not shown in the figure) positioned over the performing hand prevented the subject from having a direct view.

Figure 3. 

Experimental setup for the mirror task in Experiment 2. (A) Lateral mirror condition: The mirror placed sagittally provided a view of the nonmoving/missing hand, hidden behind the mirror. (B) Frontal mirror condition: The mirror provided a frontal visual feedback of the performing hand. An opaque cover (not shown in the figure) positioned over the performing hand prevented the subject from having a direct view.

All normally limbed subjects performed the task with the right hand under the direct view, the sagittal mirror view, and the frontal mirror view conditions. In addition, to compute a reference LRP (i.e., obtained with both hands), normally limbed subjects performed the task under a fourth condition—they performed the motor task with the left hand with a direct view (direct left view condition).

An opaque cover was used to prevent direct vision of the performing hand in all conditions involving the mirror (sagittal and frontal mirror-view conditions). Two hundred trials randomly spaced from 2600 to 3000 msec were performed under each condition.

Electrophysiological Recording

EEG were recorded using Ag/AgCl electrodes mounted in an elastic cap. The electrodes were placed at C3 and C4 sites according to the 10/20 system (Klem, Luders, Jasper, & Elger, 1999), above the hand area of the left and right motor cortices, and sampled at a rate of 500 Hz (band-pass filter = 0.01–500 Hz, with off-line digital smoothing, 10 Hz cutoff). Impedances were maintained at less than 5 kΩ.

Vertical, horizontal EOG and muscular potentials were recorded from bipolar derivations using Ag/AgCl electrodes (band-pass filter = 0.01–200 Hz) to monitor ocular artifacts and EMG activity from the muscles of the intact hand in amputees and of the right and left hands in normally limbed subjects.

The artifact rejection and the duration of averaging epoch were from 200 msec before stimulus onset to 500 msec afterward. After rejection of invalid trials (contaminated by ocular, muscular, and cerebral artifacts), data analyses were performed on a mean of 186.4 trials per condition and subject.

EEG Analysis

To calculate response-specific lateralization according to the viewed hand, ERPs were computed separately for each of the four conditions defined by the visual feedback of the performing hand, with all potentials subjected to a two-step subtraction procedure. This double subtraction method accounted for unilateral motor activity evoked by a voluntary movement. First, the difference in waveforms between the contralateral and the ipsilateral central electrodes with respect to the performing hand was calculated for each trial and condition. Second, from the average waveform for the direct view condition trials, we subtracted the average waveform for the sagittal mirror view, the frontal mirror view, and the direct left view condition trials. For a classic LRP, the formula for normally limbed subjects is
formula

This formula computes the difference of the contralateral-minus-ipsilateral difference for right-hand response and ipsilateral-minus-contralateral difference for left-hand response. This corresponds to the total potential recorded between opposite scalp locations above the motor cortex that reverses when the response hand is reversed, eliminating any overall differences between the left and right hemispheres (independent of hand) and any overall differences between the left and right hands (independent of hemisphere). All that remains is the extent to which the response is generally larger over the hemisphere contralateral to the hand being prepared, regardless of which hand is being prepared or which hemisphere is being recorded.

Using this analytical design, a comparison of the direct view and direct left view conditions should be indicative of the classic LRP in normally limbed subjects. A comparison between the direct view and the sagittal mirror view conditions should be indicative of an LRP due to hand vision. Finally, no LRP was expected when the waveforms from the direct view and the frontal mirror view conditions were compared because both conditions involved the same lateralized view of the performing hand. Thus, subtracting a negative waveform from a similar negative waveform should result in a baseline waveform, which was termed the control potential.

Results

ERP amplitudes were analyzed for 200 msec from LRP onset. Onset of LRP was estimated for each subject using a segmented regression method, based on average ERP amplitudes (Mordkoff & Gianaros, 2000). This method defined the LRP onset as the point of intersection between two straight lines that were fitted to the LRP waveform. One line was fitted to the putative preonset segment of the LRP, whereas the other line was fitted to the segment that rose to the peak.

In the normally limbed group, right and left hands under direct view yielded an LRP (direct view vs. direct left view conditions). In addition, an LRP in the normally limbed group was observed when electrophysiological activity under lateral mirror view conditions was subtracted from activity under direct view conditions (direct view vs. sagittal mirror view conditions). The averaged LRPs for the normally limbed group are shown in Figure 4A. These results indicate that a right-hand response was associated with activation of the left-hand primary motor cortex (right hemisphere). When we compared the mean amplitudes of the two LRP waveforms with the control potential, we found that the mean amplitudes for both the direct and lateral mirror-view LRPs differed from the control potential (z = 2.52, p = .012 and z = 2.38, p = .017, respectively, Wilcoxon test). There was also a difference between direct and lateral mirror view LRPs (z = 2.52, p = .012). This difference in amplitude may reflect the incongruence between the motor preparation of the viewed hand and the motor preparation of the responding hand in the LRP induced by the lateral mirror. Although vision of the reflected hand may lead to motor preparation in the hemisphere contralateral to the seen hand (Kilner, Vargas, Duval, Blakemore, & Sirigu, 2004), the actual moving hand will also generate motor activation in the opposite hemisphere. Thus, the response potential will be less lateralized and will yield a smaller global negativity (LRP). Even when motor preparation was observed in the hemisphere contralateral to the viewed hand, no actual movements were executed by that hand (opposite hand placed behind the mirror), as shown by EMG records (Figure 4D).

Figure 4. 

Electrophysiological and EMG activity. (A) Grand average waveforms depicting the voltage differences between the scalp electrodes (C3 and C4) for the three subtraction procedures in control subjects. (B and C) Grand average waveforms for the two subtraction procedures in trauma amputees and congenital limb absence subjects, respectively. (D) EMG activity from the muscles of the performing hand for all three groups and of the nonmoving left hand in normally limbed subjects for the “sagittal mirror view” condition.

Figure 4. 

Electrophysiological and EMG activity. (A) Grand average waveforms depicting the voltage differences between the scalp electrodes (C3 and C4) for the three subtraction procedures in control subjects. (B and C) Grand average waveforms for the two subtraction procedures in trauma amputees and congenital limb absence subjects, respectively. (D) EMG activity from the muscles of the performing hand for all three groups and of the nonmoving left hand in normally limbed subjects for the “sagittal mirror view” condition.

In the trauma amputee group, an LRP was also observed when electrophysiological activity under lateral mirror view conditions was subtracted from activity under direct view conditions (direct view vs. sagittal mirror view conditions; Figure 4B). As observed in normally limbed subjects, the frontal mirror condition did not induce any LRP in the trauma amputee group (control potential). The mean LRP amplitude induced by the lateral mirror differed significantly from the control potential (z = 1.82, p = .06, Wilcoxon test). LRP amplitudes induced by the lateral mirror relative to those of the control potential were similar for both normally limbed subjects and trauma amputees (z = 1.02, p = .31).

LRPs were not observed in congenital limb absence subjects, neither under lateral mirror nor frontal mirror conditions (Figure 4C).

A posteriori examination of LRP onsets revealed significant differences between conditions. In the normally limbed group, the LRP began at a mean of 123 msec following stimulus onset. This latency period was statistically shorter than for the lateral mirror view (mean = 186 msec, z = 2.38, p = .02, Wilcoxon paired test). Differences in stimulus-locked LRP onset are caused by premotor processes (Mordkoff & Gianaros, 2000) and may be explained by the same reasons as those proposed for amplitude differences: Subjects had to inhibit movement of the real left hand. This inhibition process, probably taking place in the motor cortex, may explain the observed differences in onset latency.

In the trauma amputee group, the mean onset latency estimated using the segmented regression procedure was 124 msec. This latency did not differ from the LRP involving the two hands in normally limbed subjects (123 msec). Thus, in trauma amputees, the LRP induced by the lateral mirror view of the hand had the same temporal characteristics as LRPs in normally limbed subjects performing the task with both hands. Unlike normally limbed subjects, amputees do not have to inhibit an actual movement, even if the motor command is sent, because the muscles that are targeted are absent, diminishing any incongruence between what is seen (missing hand) and what has to be done (movement with the intact hand).

DISCUSSION

We have investigated whether activation of deafferented cortical motor areas could be revealed by visuomotor performances when visual feedback of the amputated limb was provided. Using performance and physiological measures, we found that a view of one hand could generate cortical motor preparation in the nonresponding hand. Our performance test results indicate that normally limbed subjects and trauma amputees deal with incongruent visuoproprioceptive information in a similar manner. Both of these groups had more difficulties with the task when the mirror was placed in the frontal plane rather than in the lateral plane, whereas the opposite was true for congenital upper-limb loss subjects. For trauma amputees, the image of the missing hand provided by a lateral mirror facilitated motor actions of the remaining/drawing hand; a finding also observed in normally limbed subjects. Conversely, congenital loss subjects who had never experienced the missing hand either visually or proprioceptively showed reduced motor performances for the remaining hand when visual feedback was given in the position of the missing hand compared with frontal visual feedback. These findings indicate that, under lateral feedback conditions, the discordance between vision and proprioception is reduced in normal and trauma amputee subjects but is enhanced in subjects with congenital upper-limb loss. Thus, trauma amputees likely conserve proprioceptive inputs for their missing limbs, and this proprioceptive information is used to perform visuomotor activities.

Our results are consistent with clinical rehabilitation studies using the “mirror-box illusion” to restore appropriate visual feedback to proprioceptive sensations in upper-limb amputees (Ramachandran & Rogers-Ramachandran, 1996). This study found that proprioceptive sensations could emerge from the missing limb when the remaining limb was viewed through a mirror, even several years after amputation. Generally, research on phantom limbs has been based on subjective measurement, that is, descriptions of sensations in the patients' phantom limbs (Ramachandran & Hirstein, 1998; Melzack, 1990). In the present study, cortical motor activity in M1 permitted an objective measure of kinesthetics, supplementing the performance data.

Our performance test results were consistent with the results of physiological studies using fMRI or TMS showing that motor commands to the missing limb remain effective and that cortical activity was present in motor areas previously devoted to the missing limb when performing “virtual” movements with the amputated limb (Mercier et al., 2006; Roux et al., 2003). Although those studies referred only to proprioception (kinesthetic aspect), functional integration of visual and kinesthetic information is likely preserved. In healthy subjects, integrative areas of converging visual and proprioceptive information have been localized essentially to the premotor cortex (Balslev, Nielsen, Paulson, & Law, 2005), and these cerebral regions may still receive converging inputs from both systems in trauma amputees.

Our electrophysiological results correlate with the performance results. A view of the missing hand in trauma amputees and the inactive hand in normally limbed subjects was found to enhance motor performance in a difficult task such as mirror drawing. We hypothesized that better performances correspond to motor activation that is related to the viewed hand (amputated/inactive) owing to the integration of visual feedback and a proprioceptively possible position of the hand. To determine whether vision influences motor preparation in an RT task, we analyzed LRP, which reflects preparation of the responding hand. As expected, a reliable LRP was observed when the seen hand was not the responding hand, indicating that motor preparation depends, at least partially, on the available visual information. This LRP was observed in both trauma amputees and normally limbed individuals, but not in the congenital group, in accordance with the performance findings. Our LRP data, thus, provide an objective physiological measure of visuoproprioception persistence in the missing limb of trauma amputees. Moreover, studying proprioception through vision using motor performances circumvents the problem of virtual movements by amputees recorded in fMRI (Roux et al., 2003). Factors that can influence the final analysis (such as speed of phantom movement and amplitude and type of movement) can only be determined based on patient descriptions of phantom-limb movement, an extremely subjective measure (Roux et al., 2003).

The recording of an LRP in the lateral mirror condition supports the hypothesis that the motor cortices of trauma amputees still respond to the missing hand and that the visual and the motor systems in humans are tightly related. These results are consistent with previous electrophysiological studies in healthy subjects, showing that observation of a movement leads to a cortical motor activation, even without an actual movement (Kilner et al., 2004). The findings shown here also indicate that this motor activation is lateralized, depending on the visual feedback. These findings are also in accordance with physiological studies showing that observing an action is sufficient to induce changes in the sensory-motor cortex. Indeed, lateralized activity is observed in fMRI images on the contralateral side to the observed movement (Oouchida et al., 2004). Moreover, that cortical activity depends on hand orientation, with a significant lateralized activation of the precentral gyrus when the observed movement is in the first-person perspective (Jackson, Meltzoff, & Decety, 2006).

The present results are also in agreement with previous studies, highlighting perceptual differences between traumatic amputees and individuals who have born without a limb (Flor et al., 1998; Ramachandran & Hirstein, 1998). Phantom sensations are rare in congenital amputees (Flor et al., 1998; Montoya et al., 1998). In the present study, the absence of cortical motor activity in congenital subjects compared with traumatic amputees and normal individuals reflects a functional difference, consistent with the assumption that congenital subjects have not developed somatosensory and motor maps due to limb deprivation during critical periods of development. However, there are some reports indicating that congenital subjects may have such sensations (Brugger et al., 2000; Ramachandran & Hirstein, 1998; Simmel, 1962). The present data do not allow us to interpret this paradox. However, these cases have often been cited as evidence that the body schema has an innate component (Melzack, 1990) as well as a possible role related to action observation (Ramachandran & Hirstein, 1998).

Thus, the present experimental findings indicate the stability of body representation in trauma amputees and suggest that the organization of visuomotor behavior depends on body representation established through prior experience. On the basis of studies showing that LRP is an on-line measure of response preparation (Coles, 1989), the presence of this component in the hemisphere contralateral to the amputation indicates neural activity in the primary motor cortex with respect to the amputated hand. Using visual feedback of the missing limb via a sagittally placed mirror, our findings demonstrate a model in which the LRP and the performance indices can be used to reliably measure cortical motor activity related to an amputated/inactive hand.

Reprint requests should be sent to Pascale Touzalin-Chretien, Laboratoire d'Imagerie et de Neurosciences Cognitives, UMR 7191, CNRS-ULP, 21 rue Becquerel, 67087 Strasbourg, France, or via e-mail: pascale.touzalin@linc.u-strasbg.fr.

REFERENCES

Balslev
,
D.
,
Nielsen
,
F. A.
,
Paulson
,
O. B.
, &
Law
,
I.
(
2005
).
Right temporoparietal cortex activation during visuo-proprioceptive conflict.
Cerebral Cortex
,
15
,
166
169
.
Berberovic
,
N.
, &
Mattingley
,
J. B.
(
2003
).
Effects of prismatic adaptation on judgements of spatial extent in peripersonal and extrapersonal space.
Neuropsychologia
,
41
,
493
503
.
Brugger
,
P.
,
Kollias
,
S. S.
,
Muri
,
R. M.
,
Crelier
,
G.
,
Hepp-Reymond
,
M. C.
, &
Regard
,
M.
(
2000
).
Beyond re-membering: Phantom sensations of congenitally absent limbs.
Proceedings of the National Academy of Sciences, U.S.A.
,
97
,
6167
6172
.
Coles
,
M. G.
(
1989
).
Modern mind-brain reading: Psychophysiology, physiology, and cognition.
Psychophysiology
,
26
,
251
269
.
de Jong
,
R.
,
Wierda
,
M.
,
Mulder
,
G.
, &
Mulder
,
L. J.
(
1988
).
Use of partial stimulus information in response processing.
Journal of Experimental Psychology: Human Perception and Performance
,
14
,
682
692
.
Flor
,
H.
,
Elbert
,
T.
,
Muhlnickel
,
W.
,
Pantev
,
C.
,
Wienbruch
,
C.
, &
Taub
,
E.
(
1998
).
Cortical reorganization and phantom phenomena in congenital and traumatic upper-extremity amputees.
Experimental Brain Research
,
119
,
205
212
.
Galdo-Alvarez
,
S.
, &
Carrillo-de-la-Pena
,
M. T.
(
2004
).
ERP evidence of MI activation without motor response execution.
NeuroReport
,
15
,
2067
2070
.
Gemba
,
H.
,
Sasaki
,
K.
, &
Tsujimoto
,
T.
(
1990
).
Cortical field potentials associated with hand movements triggered by warning and imperative stimuli in the monkey.
Neuroscience Letters
,
113
,
275
280
.
Giraux
,
P.
, &
Sirigu
,
A.
(
2003
).
Illusory movements of the paralyzed limb restore motor cortex activity.
Neuroimage
,
20(Suppl. 1)
,
S107
S111
.
Hay
,
J. C.
,
Pick
,
H. L.
, Jr., &
Ikeda
,
K.
(
1965
).
Visual capture produces by prism spectacles.
Psychonomic Science
,
2
,
215
216
.
Jackson
,
P. L.
,
Meltzoff
,
A. N.
, &
Decety
,
J.
(
2006
).
Neural circuits involved in imitation and perspective-taking.
Neuroimage
,
31
,
429
439
.
Karl
,
A.
,
Birbaumer
,
N.
,
Lutzenberger
,
W.
,
Cohen
,
L. G.
, &
Flor
,
H.
(
2001
).
Reorganization of motor and somatosensory cortex in upper extremity amputees with phantom limb pain.
Journal of Neuroscience
,
21
,
3609
3618
.
Killackey
,
H. P.
,
Rhoades
,
R. W.
, &
Bennett-Clarke
,
C. A.
(
1995
).
The formation of a cortical somatotopic map.
Trends in Neurosciences
,
18
,
402
407
.
Kilner
,
M. J.
,
Vargas
,
C.
,
Duval
,
S.
,
Blakemore
,
S. J.
, &
Sirigu
,
A.
(
2004
).
Motor activation prior to observation of a predicted movement.
Nature Neuroscience
,
7
,
1299
1301
.
Klem
,
G. H.
,
Luders
,
H. O.
,
Jasper
,
H. H.
, &
Elger
,
C.
(
1999
).
The ten-twenty electrode system of the International Federation. The International Federation of Clinical Neurophysiology.
Electroencephalography and Clinical Neurophysiology
,
52(Suppl.)
,
3
6
.
Lajoie
,
Y.
,
Paillard
,
J.
,
Teasdale
,
N.
,
Bard
,
C.
,
Fleury
,
M.
,
Forget
,
R.
,
et al
(
1992
).
Mirror drawing in a deafferented patient and normal subjects: Visuoproprioceptive conflict.
Neurology
,
42
,
1104
1106
.
Melzack
,
R.
(
1990
).
Phantom limbs and the concept of a neuromatrix.
Trends in Neurosciences
,
13
,
88
92
.
Mercier
,
C.
,
Reilly
,
K. T.
,
Vargas
,
C. D.
,
Aballea
,
A.
, &
Sirigu
,
A.
(
2006
).
Mapping phantom movement representations in the motor cortex of amputees.
Brain
,
129
,
2202
2210
.
Michel
,
C.
,
Pisella
,
L.
,
Halligan
,
P. W.
,
Luaute
,
J.
,
Rode
,
G.
,
Boisson
,
D.
,
et al
(
2003
).
Simulating unilateral neglect in normals using prism adaptation: Implications for theory.
Neuropsychologia
,
41
,
25
39
.
Michel
,
C.
,
Rossetti
,
Y.
,
Rode
,
G.
, &
Tilikete
,
C.
(
2003
).
After-effects of visuo-manual adaptation to prisms on body posture in normal subjects.
Experimental Brain Research
,
148
,
219
226
.
Miller
,
J.
, &
Hackley
,
S. A.
(
1992
).
Electrophysiological evidence for temporal overlap among contingent mental processes.
Journal of Experimental Psychology: General
,
121
,
195
209
.
Minelli
,
A.
,
Marzi
,
C. A.
, &
Girelli
,
M.
(
2007
).
Lateralized readiness potential elicited by undetected visual stimuli.
Experimental Brain Research
,
179
,
683
690
.
Montoya
,
P.
,
Ritter
,
K.
,
Huse
,
E.
,
Larbig
,
W.
,
Braun
,
C.
,
Topfner
,
S.
,
et al
(
1998
).
The cortical somatotopic map and phantom phenomena in subjects with congenital limb atrophy and traumatic amputees with phantom limb pain.
European Journal of Neuroscience
,
10
,
1095
1102
.
Mordkoff
,
J. T.
, &
Gianaros
,
P. J.
(
2000
).
Detecting the onset of the lateralized readiness potential: A comparison of available methods and procedures.
Psychophysiology
,
37
,
347
360
.
Okada
,
Y. C.
,
Williamson
,
S. J.
, &
Kaufman
,
L.
(
1982
).
Magnetic field of the human sensorimotor cortex.
International Journal of Neuroscience
,
17
,
33
38
.
Oouchida
,
Y.
,
Okada
,
T.
,
Nakashima
,
T.
,
Matsumura
,
M.
,
Sadato
,
N.
, &
Naito
,
E.
(
2004
).
Your hand movements in my somatosensory cortex: A visuo-kinesthetic function in human area 2.
NeuroReport
,
15
,
2019
2023
.
Pascual-Leone
,
A.
,
Peris
,
M.
,
Tormos
,
J. M.
,
Pascual
,
A. P.
, &
Catala
,
M. D.
(
1996
).
Reorganization of human cortical motor output maps following traumatic forearm amputation.
NeuroReport
,
7
,
2068
2070
.
Ramachandran
,
V. S.
, &
Hirstein
,
W.
(
1998
).
The perception of phantom limbs. The D. O. Hebb lecture.
Brain
,
121
,
1603
1630
.
Ramachandran
,
V. S.
, &
Rogers-Ramachandran
,
D.
(
1996
).
Synaesthesia in phantom limbs induced with mirrors.
Proceedings of the Royal Society of London, Series B, Biological Sciences
,
263
,
377
386
.
Ramachandran
,
V. S.
,
Stewart
,
M.
, &
Rogers-Ramachandran
,
D. C.
(
1992
).
Perceptual correlates of massive cortical reorganization.
NeuroReport
,
3
,
583
586
.
Rhoades
,
R. W.
,
Chiaia
,
N. L.
,
Bennett-Clarke
,
C. A.
,
Janas
,
G. J.
, &
Fisher
,
C. M.
(
1994
).
Alterations in brainstem and cortical organization of rats sustaining prenatal vibrissa follicle lesions.
Somatosensory & Motor Research
,
11
,
1
17
.
Rossetti
,
Y.
,
Koga
,
K.
, &
Mano
,
T.
(
1993
).
Prismatic displacement of vision induces transient changes in the timing of eye-hand coordination.
Perception & Psychophysics
,
54
,
355
364
.
Roux
,
F. E.
,
Lotterie
,
J. A.
,
Cassol
,
E.
,
Lazorthes
,
Y.
,
Sol
,
J. C.
, &
Berry
,
I.
(
2003
).
Cortical areas involved in virtual movement of phantom limbs: Comparison with normal subjects.
Neurosurgery
,
53
,
1342
1352
; discussion
1352
1343
.
Simmel
,
M. L.
(
1962
).
Phantom experiences following amputation in childhood.
Journal of Neurology, Neurosurgery and Psychiatry
,
25
,
69
78
.
Vaughan
,
H. G.
, Jr.,
Costa
,
L. D.
, &
Ritter
,
W.
(
1968
).
Topography of the human motor potential.
Electroencephalography and Clinical Neurophysiology
,
25
,
1
10
.