European Journal of Neuroscience, Vol. 40, pp. 2581–2588, 2014

doi:10.1111/ejn.12615

COGNITIVE NEUROSCIENCE

Training based on mirror visual feedback influences transcallosal communication Laura Avanzino,1 Alessia Raffo,2 Elisa Pelosin,2 Carla Ogliastro,2 Roberta Marchese,2 Piero Ruggeri1 and Giovanni Abbruzzese2 1

Department of Experimental Medicine, Section of Human Physiology and Centro Polifunzionale di Scienze Motorie, University of Genoa, Genoa, Italy 2 Department of Neuroscience, Rehabilitation, Ophthalmology, Genetics and Maternal Child Health, University of Genoa, Genoa, Italy Keywords: interhemispheric inhibition, mirror visual feedback, training, transcranial magnetic stimulation

Abstract Mirror visual feedback (MVF) therapy has been demonstrated to be successful in neurorehabilitation, probably inducing neuroplasticity changes in the primary motor cortex (M1). However, it is not known whether MVF training influences the hemispheric balance between the M1s. This topic is of extreme relevance when MVF training is applied to stroke rehabilitation, as the competitive interaction between the two hemispheres induces abnormal interhemispheric inhibition (IHI) that weakens motor function in stroke patients. In the present study, we evaluated, in a group of healthy subjects, the effect of motor training and MVF training on the excitability of the two M1s and the IHI between M1s. The IHI from the ‘active’ M1 to the opposite M1 (where ‘active’ means the M1 contralateral to the moving hand in the motor training and the M1 of the seen hand in the MVF training) increased, after training, in both the experimental conditions. Only after motor training did we observe an increase in the excitability of the active M1. Our findings show that training based on MVF may influence the excitability of the transcallosal pathway and support its use in disorders where abnormal IHI is a potential target, such as stroke, where an imbalance between the affected and unaffected M1s has been documented.

Introduction One of the most important goals of motor rehabilitation is to restore voluntary movement control. Functional motor recovery, e.g. after stroke, has shown to be mainly dependent on therapy intensity and active motor repetition (H€omberg, 2013). However, in addition to active motor training, different strategies may allow patients to train motor function when voluntary movement is impaired (Langhorne et al., 2011). Among them, mirror visual feedback (MVF) therapy has been increasingly used in recent years (Rothgangel et al., 2011; Shiri et al., 2012). MVF therapy is aimed at supplying visual feedback from the affected arm in a very peculiar way, as the person is performing the action with the opposite limb (thus receiving motor training and proprioception from that limb), but receives visual feedback from the opposite limb (Ramachandran & Altschuler, 2009). A pivotal component of normal learning and recovery after neural injury is represented by the ability of sensory and motor cortices to dynamically re-organize, i.e. cortical plasticity. The data available in the literature suggest that MVF therapy may be able to induce plasticity modifications in the primary motor cortex (M1). TouzalinChretien et al. (2010) showed that, in a motor task, viewing one hand in a sagittal mirror (giving the impression of seeing the

Correspondence: L. Avanzino, as above. E-mail: [email protected] Received 12 February 2014, revised 7 April 2014, accepted 9 April 2014

opposite hand) generates electroencephalogrpahic activity in the motor cortex of the seen hand (i.e. of the non-moving hand). Also, a transcranial magnetic stimulation (TMS) study revealed that MVF could enhance excitatory functions in M1 (Nojima et al., 2012). However, in this framework, the question of the influence of MVF training on the hemispheric balance between the M1 contralateral and ipsilateral to the moving hand remains open. Indeed, balanced interhemispheric interactions between the contralateral and ipsilateral M1, via the corpus callosum, are required for the generation of proper voluntary movements (Ferbert et al., 1992). This topic is also of great relevance when MVF training is applied to stroke rehabilitation (Hummel & Cohen, 2006). The competitive interaction between the two hemispheres induces abnormal interhemispheric inhibition (IHI) that weakens motor function in stroke patients (Takeuchi et al., 2012). In particular, an abnormally high interhemispheric inhibitory drive from the intact to the lesioned M1 was observed in patients with chronic subcortical stroke, and was more prominent in cases with greater motor impairment (Murase et al., 2004), suggesting that interhemispheric balance influences functional recovery. The aim of this study was to directly investigate whether training based on MVF of movement is able to influence the hemispheric balance between the two M1s. To this end, we investigated, by means of TMS, both the excitability of the two M1s and the transcallosal interaction between them in two groups of healthy subjects, who underwent different types of training. One group underwent

© 2014 Federation of European Neuroscience Societies and John Wiley & Sons Ltd

2582 L. Avanzino et al. motor training with the right hand. The second group underwent ‘mirror training’, i.e. the same task as performed by the motor training group, but with the left hand inside a mirror box in order to create the illusion of moving the right hand, thus creating visual feedback training (Ramachandran & Altschuler, 2009).

Materials and methods Participants Eighteen healthy subjects (10 females and eight males; age, 19– 36 years; mean age, 24.5  1.39 years) without any history of neurological and orthopaedic disease were recruited for this study. All subjects were right-handed. Right-hand dominance was evaluated with the Edinburgh Handedness Inventory (Oldfield, 1971). Subjects had no contraindication to TMS. All subjects were na€ıve to the purpose of the experiment, and they gave written informed consent before participation. The experimental protocol was approved by the ethics committee of the University of Genoa, and was carried out in accordance with legal requirements and international norms (Declaration of Helsinki, 1964). Fig. 1. Experimental paradigm.

Experimental conditions Subjects were randomly assigned to two different groups (nine subjects each) matched for age and gender (group 1 – four males and five females, mean age of 24.5  1.42 years; group 2 – four males and five females, mean age of 24.5  1.50 years). The two groups executed different training protocols. In all of the experimental conditions, the motor task consisted of repetitive finger tapping movements (opposition of the thumb to the index) paced at 2 Hz (set by an acoustic cue) for 60 s. The task was repeated 10 times (giving a total of 10 min, i.e. a total of 1200 tapping movements). A resting period of 30 s between each 60-s block was used, to avoid any fatigue. In the current study, all measurements took place in the afternoon (between 14:00 h and 15:00 h), because, at a behavioural level, circadian variations have been observed for fine, skilled movements (Gueugneau et al., 2008). The experimental paradigm is summarized in Fig. 1. Group 1: motor training Subjects were asked to execute repetitive finger tapping movements with the right hand, with their eyes open. Vision was directed to the right hand, in order to give similar visual feedback to that provided during MVF training. Group 2: MVF training Subjects, with both hands inside a mirror box, were asked to perform the finger tapping task with the left hand. During this training, participants were required to carefully watch the mirror, placed in the middle of the box, in order to elicit the illusion of moving their right hand, thus creating MVF training. Recording procedures Electromyographic recording Electromyographic (EMG) activity was recorded from the right and left abductor pollicis brevis muscles, with silver disc surface electrodes. The ground electrode was placed at the elbow. EMG signals were amplified and filtered (20 Hz to 1 kHz) with a D360 amplifier

(Digitimer). The signals were sampled at 5000 Hz, digitized with a laboratory interface (Power1401; Cambridge Electronics Design), and stored on a personal computer for display and later offline data analysis. Each recording epoch lasted 400 ms; 100 ms of this preceded the TMS. Trials with background EMG activity were excluded from analysis. TMS Cortical activity changes were tested by means of TMS before and immediately after training. We tested cortical excitability of both the left and the right M1 by means of an input–output (IO) recruitment curve, and interhemispheric communication between the two M1s by assessing IHI. The order of the IO curve and IHI determination was random for each session (before and after training), in order to avoid spontaneous decay of the training-induced effect. For IO experiments, TMS was performed with a single Magstim 200 magnetic stimulator (Magstim Company) connected with a figure-of-eight coil (wing diameter, 70 mm). For IHI experiments, TMS was given through two Magstim 200 stimulators, one connected to a figure-of-eight coil with wing diameters of 70 mm (test stimulus) and the other connected to a smaller figure-of-eight coil with wing diameters of 50 mm (conditioning stimulus). The coils were placed tangentially to the scalp with the handle pointing backwards and laterally at 45° to the sagittal plane, inducing a postero-anterior current in the brain. This orientation was chosen because the lowest motor threshold is achieved when the induced electrical current flows approximately perpendicular to the line of the central sulcus. We determined the optimal position for activation of the left and right abductor pollicis brevis muscles by moving the coil in 0.5-cm steps around the presumed motor hand area. Resting motor threshold (RMT), defined as the minimum stimulus intensity that produced a motor evoked potential (MEP) of at least 0.05 mV in five of 10 consecutive trials, was found and expressed as a percentage of maximum stimulator output (MSO).

© 2014 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 40, 2581–2588

Mirror visual feedback and hemispheric balance 2583 IO curve. The IO curve was determined by measuring peak-to-peak amplitude (expressed in millivolts) of the MEPs elicited at different stimulus intensities of 5%, 10%, 15%, 20% and 25% of MSO above RMT. Seven trials were performed at each stimulus intensity, and the MEP size was taken as the average of the MEP amplitudes. The presentation of intensities was random, in order to minimize hysteresis effects (M€oller et al., 2009). Analysis of the steepness of the IO curve slope was used to assess relative changes from before training to after training (Rosenkranz et al., 2007a,b). For each subject, a regression model was applied on the linear part of the IO curves (from RMT to 15% of MSO above RMT) before and after the training protocol for the left and the right M1s, and the slopes of the linear fits were estimated. The change in the slope of the IO curve (DSlope) was calculated as the slope difference between the IO curves measured before and after training, and was expressed as a percentage (DSlope = [(Slopeafter – Slopebefore)/Slopebefore] 9 100). IHI. IHI both from left-to-right (LtoR) and from right-to-left (RtoL) M1s was tested with a previously reported randomized conditioningtest design (Ferbert et al., 1992). A suprathreshold-conditioning stimulus was given to one hemisphere 10 ms before a test stimulus delivered to the other side. The test stimulus was adjusted to produce an MEP with a peak-to-peak amplitude of ~1 mV. The conditioning stimulus was set at 130% of RMT. A 10-ms inter-stimulus interval and the conditioning and test intensities were chosen because they have been reported to be the most effective for studying IHI (or because the maximum short-latency IHI was found with these parameters) (Ni et al., 2009). Stimuli were randomly delivered in one set of 30 trials: 15 conditioned and 15 unconditioned. IHI was expressed as the ratio between the mean peak-to-peak MEP amplitude in conditioned vs. unconditioned trials. At each session (PRE and POST-training), conditioning stimulus was adjusted on the basis of the individual RMT found in the single session, and the test stimulus was adjusted to obtain a 1 mV MEP amplitude. Statistical analysis After a normal distribution of data had been verified, between-group differences of RMTs and IO curve slopes before and after training were analysed separately for each hemisphere by means of a repeated measures ANOVA, with group (1 and 2) as between-subjects factor, and time (before and after training) as within-subject factor. Then, we evaluated DSlope value differences between groups by means of a unpaired t-test. Paired t-tests were performed to compare: (i) the conditioned and the unconditioned trials of LtoR IHI and RtoL IHI at baseline (before training); and (ii) the unconditioned trials before training and those after training. This analysis was performed separately in each group. Normalized data (conditioned/unconditioned) of RtoL IHI and LtoR IHI were separately subjected to a repeated measures ANOVA with group (1 and 2) as between-subjects factor, and time (before and after training) as within-subject factors. The significance threshold was set at P < 0.05. If ANOVA showed a significant interaction effect, we performed post hoc comparisons by use of the least significance difference (Fisher’s) test to directly compare the experimental factors. Furthermore, to investigate a possible relationship between training-induced changes in IHI and in M1 excitability (i.e. between RtoL IHI and left M1 slope, and between LtoR IHI and right M1 slope), a normalized index for changes in M1 excitability (M1 index) and in IHI (IHI index) was calculated as follows: M1 index = [(IO slopeafter – IO slopebefore)/IO slopebefore] 9 100; IHI

index = [(IHIafter – IHIbefore)/IHIbefore] 9 100. Spearman’s correlation coefficient was applied to assess possible correlations. All statistical analyses were performed with SPSS 13.0. Data are presented as mean  standard error (SE).

Results Cortical excitability The resting motor threshold for TMS of the right and left hemispheres did not change after training in any of the experimental conditions. Accordingly, repeated measures ANOVA on RMT values did not show any significant effect of time (left M1, F1,16 = 0.59, P = 0.45; right M1, F1,16 = 0.71, P = 0.41) or time 9 group interaction (left M1, F1,16 = 0.73, P = 0.40; right M1, F1,16 = 0.54, P = 0.47). For each subject, in the different experimental sessions (before and after training), a linear relationship fitted with the IO curve (R2 always >0.7) for both the left and the right M1. The excitability of the left hemisphere, as tested with the IO curve, significantly increased in group 1 (motor training), whereas it did not change in group 2 (MVF training) (Fig. 2). Accordingly, repeated measures ANOVA on IO slope data showed a significant time 9 group intercation (F1,16 = 9.24, P = 0.008). Post hoc analysis revealed that only for the motor training group was there a significant increase in the steepness of the left M1 IO curve slope after training (P = 0.001); no change was observed in the MVF group (P = 0.74). The comparison of the training-induced changes in the slope of IO curves (DSlope values) showed a significant difference between group 1 and group 2 (P = 0.0008), with DSlope being significantly greater in group 1 than in group 2. The excitability of the right hemisphere, as tested with the IO curve, did not change in the two groups of subjects after training (Fig. 3). Repeated measures ANOVA on slope values showed neither a main effect of time (F1,16 = 0.38, P = 0.54) nor a significant time 9 group interaction (F1,16 = 0.031, P = 0.86). Accordingly, DSlope of the IO curve of the right hemisphere was not different between groups (P = 0.52). IHI The mean amplitudes of unconditioned and conditioned MEPs evoked in the IHI studies (both before and after training) are shown in Table 1. At baseline, when the test stimulus was preceded by a conditioning stimulus delivered 10 ms earlier in the contralateral hemisphere, a significant reduction in MEP size was observed in either the right or left hemisphere with respect to unconditioned trials (test stimulus alone) in all groups of subjects (group 1, LtoR IHI, P = 0.0005; group 1, RtoL IHI, P = 0.0003; group 2, LtoR IHI, P = 0.0002; group 2, RtoL IHI, P = 0.0003). Furthermore, there was no difference between the size of the unconditioned test pulses collected before training and those collected after training (P-value always >0.5; Table 1). After the different training sessions, LtoR IHI increased in all subjects (Fig. 4). Accordingly, repeated measures ANOVA revealed a significant effect of time (F1,16 = 2.45, P = 0.003) and no time 9 group interaction (F1,16 = 0.60, P = 0.44). After training, the increase in LtoR IHI, with respect to baseline values, was ~15% in all groups (group 1, P = 0.008; group 2, P = 0.040). In contrast, RtoL IHI did not change after training in either group (time, F1,16 = 0.013, P = 0.91; time 9 group, F1,16 = 0.12, P = 0.72).

© 2014 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 40, 2581–2588

2584 L. Avanzino et al. A

B

C

D

Fig. 2. IO curves of the left M1 before and after training. Data from the two groups of subjects who underwent different training (group 1, active motor training; group 2, MVF training) are shown. (A) Mean values of the slopes of the IO curves, measured before and after training. (B) Mean DSlope values. (C and D) MEP amplitudes before and after training. The slope was estimated from the linear part of the IO curve between RMT and 15% of MSO above RMT, as indicated by the shaded area. Asterisks indicate that, in the left M1, after training, the slope value was significantly steeper than before training only in group 1, and consequently DSlope was significantly greater in group 1 than in group 2 (**P < 0.01). Vertical bars indicate SEs.

Correlation between IHI and cortical excitability As only LtoR IHI was influenced by prior training (both motor and MVF training), correlation analysis was performed between the LtoR IHI index and right M1 index in the two groups of subjects. The training-dependent increases in LtoR IHI were significantly correlated with changes in right M1 excitability only for the MVF group (Spearman’s q = 0.75, P = 0.020) (Fig. 5). As shown in Fig. 5, those subjects with the smallest changes in LtoR IHI had the largest increases in right M1 excitability, whereas, when LtoR IHI substantially increased after training, right M1 excitability tended to decrease. No significant correlation was found in the motor training group (Spearman’s q = 0.38, P = 0.31).

Discussion MVF is an alternative rehabilitative strategy used in facilitating motor recovery after stroke (Ramachandran & Altschuler, 2009). The main aim of the present study was to investigate whether unilateral hand training based on MVF training is able to induce changes in the excitability of contralateral and ipsilateral M1s or in the transcallosal communication between them, and possible differences with respect to standard motor training. The main results of the study indicated that: (i) the excitability of the ‘active’ M1 (where ‘active’ means the M1 contralateral to the moving hand in the motor training and the M1 of the seen nonmoving hand in the MVF training) was significantly increased after

motor training, whereas no significant changes were noticed after MVF training; and (ii) transcallosal inhibition from the ‘active’ M1 to the opposite M1 increased after both motor training and MVF training. Corticospinal excitability An increase in the slope of the MEP IO curve of the left M1 was observed following motor training, indicating that neurons with a higher threshold for TMS (including neurons located at a greater distance from the centre of the coil) showed sufficiently large excitability changes. This finding is in line with other studies showing an increase in the IO slope consistent with an active training-dependent increase in cortico-motoneuronal excitability (Devanne et al., 1997; Ridding & Rothwell, 1997; Lotze et al., 2003). However, the lack of modulation of cortical excitability after MVF training contrasts with other findings in the literature. Recently, Nojima et al. (2012) showed that MVF intervention in normal volunteers improved motor behaviour and enhanced excitatory function of the M1. One possible explanation for the discrepancy between our findings and those of Nojima et al. (2012) may be related to the different characteristics of the motor task. In the current study, we used a finger tapping task, paced by a metronome, whereas, in the previous study, the authors used a ball-rotation task, in which the aim was to rotate two cork balls as quickly as possible in a counter-clockwise direction. Thus, both the complexity and the effort required by the movement were substantially different between the two tasks.

© 2014 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 40, 2581–2588

Mirror visual feedback and hemispheric balance 2585 A

B

C

D

Fig. 3. IO curves of the right M1 before and after training. Data from the two groups of subjects who underwent different training (group 1, active motor training; group 2, MVF training) are shown. (A) Mean values of the slopes of the IO curves, measured before and after training. (B) Mean changes in the DSlope values. (C and D) MEP amplitudes before and after training. The slope was estimated from the linear part of the IO curve between RMT and 15% of MSO above RMT, as indicated by the shaded area. Vertical bars indicate SEs.

Previous studies have shown that cortical excitability changes during a motor task are complexity-dependent and effort-dependent (Thickbroom et al., 1998; Kuhtz-Buschbeck et al., 2003; Roosink & Zijdewind, 2010; Gueugneau et al., 2013). In addition, it is worth noting that, in a recent study, Lepage et al. (2012) showed that observational training, based on observing someone else performing thumb abductions, did not lead to a significant increase in cortical excitability (Lepage et al., 2012). The scenario in that study was different from ours; the MVF illusion is unique, in that the person is performing the action with the opposite limb (thus receiving motor training and proprioception from that limb), but receives visual feedback from the opposite limb, whereas observational training such as that used by Lepage et al. (2012) is based on external visual feedback. However, the issues related to the training-induced changes in M1 excitability by means of visual feedback (action observation or mirror illusion) are still open and worthy of future study. IHI IHI refers to the neurophysiological mechanism whereby the M1 of one hemisphere of the brain inhibits the opposite M1, and arises from pyramidal neurons of cortical layer III in one M1 (transcallosal cells) projecting, via the corpus callosum, onto GABAergic inhibitory interneurons in the other M1, which in turn modulate the excitability of corticospinal pyramidal neurons of cortical layer V located in the same hemisphere (Ferbert et al., 1992). In our study, 10 min of unilateral right-hand motor training induced an increase in the transcallosal communication from the active M1 (left M1) to the opposite M1. As we did not observe any

Table 1. Values of unconditioned (Test) and conditioned (Conditioned) MEPs of LtoR and RtoL IHI paradigms Before Test*

After Conditioned Test*

Group 1: motor training IHI LtoR Mean  SD 1.24  0.61 0.67  0.40 IHI RtoL Mean  SD 1.27  0.25 0.67  0.24 Group 2: mirror visual feedback training IHI LtoR Mean  SD 1.15  0.18 0.58  0.30 IHI RtoL Mean  SD 1.17  0.25 0.51  0.14

PConditioned value

1.21  0.39 0.54  0.29 0.73* 1.29  0.24 0.67  0.28 0.62* 1.20  0.27 0.41  0.20 0.54* 1.26  0.41 0.57  0.27 0.22*

Mean data  standard deviation (SD) from the motor training (group 1) and MVF training (group 2) groups are shown. Asterisks indicate that paired t-tests were performed between the unconditioned trial results collected before training and the results collected after training in both groups of subjects (P-value always > 0.05).

excitability change in the right ipsilateral M1, we can suppose that the increase in LtoR IHI might depend on an increase in the excitability of transcallosal cells located in the left hemisphere, in parallel with the increase in excitability of corticospinal cells (increase in the IO slope). As both transcallosal and corticospinal neurons are modulated by a similar interneuron population (Trompetto et al., 2004; Avanzino et al., 2007), one possible explanation might involve changes in the excitability of interneurons that control both transcal-

© 2014 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 40, 2581–2588

2586 L. Avanzino et al. A

B

Fig. 4. Mean IHI values in the two groups of subjects (group 1, active motor training; group 2, MVF training) before training (white columns) and after training (grey columns). IHI from LtoR and RtoL hemispheres is expressed as the ratio of the mean peak-to-peak MEP amplitude in conditioned trials and that in unconditioned trials (MEP cond/MEP test). Asterisks indicate significant difference in LtoR IHI between before and after values in both groups of subjects. *P < 0.05. Vertical bars indicate SEs.

Fig. 5. Correlation analysis between MVF training-induced changes in IHI (LtoR IHI, x-axis) and in right M1 excitability (y-axis) (group 2: white circles). There is a significant positive correlation between the LtoR IHI index and the right M1 index (Spearman’s q = 0.75, P = 0.020), indicating that the training-induced increase in LtoR IHI was associated with a decrease in right M1 excitability.

losal and corticospinal neurons, inducing similar effects in the two neural systems. In agreement with this hypothesis, we recently showed that a short period of right-hand non-use induced a decrease in the cortical excitability of the left M1 and a reduction in IHI from

the left hemisphere to the right hemisphere (Avanzino et al., 2011, 2013). Several findings reported in the literature support the idea that interhemispheric communication between the M1s plays a role in the control of unilateral hand movements. Indeed, during the execution of unimanual finger movements, the contralateral M1 inhibits more deeply the ipsilateral one with respect to a rest condition (Duque et al., 2007). From a functional point of view, the increase in IHI from the active to the opposite M1 during unilateral hand movements seems to be instrumental in suppressing mirror activity (H€ ubers et al., 2008; Giovannelli et al., 2009). Thus, we can speculate that, in our experimental paradigm, the priority for the motor system was to maintain the strict laterality of the task, and that the increase in transcallosal inhibition from the active to the opposite hemisphere reflects an increase in synaptic strength between the neural populations involved. Interestingly, we observed clear modulation of transcallosal communication from the active to the opposite M1 by means of MVF training that was in the same direction as that observed after motor training (i.e. increase of LtoR IHI). The neurophysiological mechanisms of MVF are still not completely clear. It has been postulated that MVF might owe part of its efficacy to stimulation of the ‘mirror neuron system’, a cortical–subcortical network including the M1, premotor and parietal areas, basal ganglia, and cerebellum, which operates when we observe the actions performed by others (Ramachandran & Altschuler, 2009). According to this hypothesis, MVF might provide the visual input to activate motor neurons. Alternatively, visual motor imagery (imagining seeing oneself or another person performing actions with an exterior view) might be involved. When you observe or imagine an action, partial activation of the very same neural pathways as would be evoked during movement execution occurs (Kosslyn et al., 1983). The same holds true for the transcallosal pathway; indeed, neurophysiological data have shown that, when M1 is activated by unilateral imagined movements of the contralateral hand, there is an increase in the interhemispheric motor inhibition of the opposite M1 (Giovannelli et al., 2009; Gueugneau et al., 2013). According to this evidence, we can speculate that MVF training might have modulated the excitability of transcallosal cells in the left M1, through direct inputs to M1, or by means of cortico-cortical connections from premotor or parietal areas to M1. Whatever the case, our findings suggest that information related to unilateral hand movement coming from visual receptors of the eye, even in the absence of active motor training, would probably be sufficient to activate the population of transcallosal cells in M1, whose priority is to maintain the strict laterality of the task. A last finding to be discussed is the lack of excitability changes after MVF training in the corticospinal and transcallosal (RtoL IHI) neurons located in the right M1, controlling the left hand actually moving inside the mirror box, but hidden from the subject’s sight. We might speculate that the increased interhemispheric communication from the left to the right M1 was able to prevent the increase in excitability in the right M1 by potentiating GABAergic inhibition. This hypothesis is supported by the significant correlation that we found between the training-dependent increase in LtoR IHI and changes in right hemisphere excitability in the MVF group. In other words, those subjects with the smallest changes in LtoR IHI had the largest increase in right M1 excitability, and vice versa. In accordance with this finding, it has recently been shown that there is an increase in intracortical GABAergic inhibition in the M1 contralateral to the moving hand after MVF training (L€appchen et al., 2012). Also, we have to consider the demonstration that, in the case of dis-

© 2014 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 40, 2581–2588

Mirror visual feedback and hemispheric balance 2587 cordant proprioceptive–visual sensory information (as in the case of MVF), cortical motor activity relies on the viewed image of the hand rather than on proprioceptive information, suggesting a dominance of vision over proprioception in relation to motor programming (Touzalin-Chretien et al., 2010). Interestingly, this neurophysiological finding may have important clinical implications when the necessity is to prevent a possible an increase in cortical excitability related to ‘overuse’ of one hand (e.g. when the other hand is immobilized or plegic) (Avanzino et al., 2011). Conclusion and clinical perspectives Three potential limitations of the present study need to be mentioned. First, it would be interesting to also assess the modulation of IHI after left-hand training, in order to evaluate inter-manual differences in training-related IHI. This observation would support the application of these training procedures in a clinical setting. Second, the present study does not show whether the modulation in IHI is functionally relevant. Indeed, we did not assess possible improvements in motor skill, as this was beyond the aim of the present study. Nevertheless, in previous studies, we showed that the motor task adopted here (unilateral externally paced finger movements) could be susceptible to improvement in skill performance, probably involving motor learning processes (Avanzino et al., 2008, Avanzino et al., 2009a,b). This concept is of some relevance for the development of future protocols that, combining MVF with motor tasks able to induce motor learning, will have the final aim of enhancing plasticity in M1s. Third, in the present study we did not assess long-lasting effects of the different types of training on IHI. In general, we think that none of these issues disprove the findings of the present study; however, we also think that they should be addressed in future studies. In conclusion, our results support the use of training based on MVF in disorders where abnormal IHI is a potential target. These disorders include stroke, where the balance between the lesioned and the intact hemisphere plays a crucial role in motor recovery (Hummel & Cohen, 2006), and also other neurological diseases, such as focal hand dystonia, Parkinson’s disease, and multiple sclerosis, where abnormal interhemispheric communication is supported by clinical (Bonzano et al., 2008; Sitburana et al., 2009; Cox et al., 2012) and experimental (Boroojerdi et al., 1998; Li et al., 2007; Nelson et al., 2010; Wahl et al., 2011) data. In addition, these approaches may be useful in all conditions that lead to arm immobilization (musculoskeletal traumas, sport injuries, etc.) where hemispheric imbalance within the motor cortices may occur (Avanzino et al., 2011).

Acknowledgements This study was supported by a grant from the University of Genoa (Research Project 2012). The authors certify that they have no conflict of interest to declare.

Abbreviations EMG, electromyographic; IHI, interhemispheric inhibition; IO, input–output; LtoR, left-to-right; M1, primary motor cortex; MEP, motor evoked potential; MSO, maximum stimulator output; MVF, mirror visual feedback; RMT, resting motor threshold; RtoL, right-to-left; SE, standard error; TMS, transcranial magnetic stimulation; DSlope, change in the slope of the input–output curve.

References Avanzino, L., Teo, J.T. & Rothwell, J.C. (2007) Intracortical circuits modulate transcallosal inhibition in humans. J. Physiol., 15, 99–114. Avanzino, L., Bove, M., Trompetto, C., Tacchino, A., Ogliastro, C. & Abbruzzese, G. (2008) 1-Hz repetitive TMS over ipsilateral motor cortex

influences the performance of sequential finger movements of different complexity. Eur. J. Neurosci., 27, 1285–1291. Avanzino, L., Bove, M., Tacchino, A., Trompetto, C., Ogliastro, C. & Abbruzzese, G. (2009a) Interaction between finger opposition movements and aftereffects of 1 Hz-rTMS on ipsilateral motor cortex. J. Neurophysiol., 101, 1690–1694. Avanzino, L., Giannini, A., Tacchino, A., Pelosin, E., Ruggeri, P. & Bove, M. (2009b) Motor imagery influences the execution of repetitive finger opposition movements. Neurosci. Lett., 27, 11–15. Avanzino, L., Bassolino, M., Pozzo, T. & Bove, M. (2011) Use-dependent hemispheric balance. J. Neurosci., 31, 3423–3428. Avanzino, L., Pelosin, E., Abbruzzese, G., Bassolino, M., Pozzo, T. & Bove, M. (2013) Shaping cortex plasticity motor through proprioception. Cereb. Cortex, doi: 10.1093/cercor/bht139. [Epub ahead of print]. Bonzano, L., Tacchino, A., Roccatagliata, L., Abbruzzese, G., Mancardi, G.L. & Bove, M. (2008) Callosal contributions to simultaneous bimanual finger movements. J. Neurosci., 28, 3227–3233. Boroojerdi, B., Hungs, M., Mull, M., T€ opper, R. & Noth, J. (1998) Interhemispheric inhibition in patients with multiple sclerosis. Electroen. Clin. Neuro., 109, 230–237. Cox, B.C., Cincotta, M. & Espay, A.J. (2012) Mirror movements in movement disorders: a review. Tremor. Other Hyperkinet. Mov. (N Y), 2, PMCID: PMC3569961. Devanne, H., Lavoie, B.A. & Capaday, C. (1997) Input–output properties and gain changes in the human corticospinal pathway. Exp. Brain Res., 114, 329–338. Duque, J., Murase, N., Celnik, P., Hummel, F., Harris-Love, M., Mazzocchio, R., Olivier, E. & Cohen, L.G. (2007) Intermanual differences in movementrelated interhemispheric inhibition. J. Cognitive Neurosci., 19, 204–213. Ferbert, A., Priori, A., Rothwell, J.C., Day, B.L., Colebatch, J.G. & Marsden, C.D. (1992) Interhemispheric inhibition of the human motor cortex. J. Physiol., 453, 525–546. Giovannelli, F., Borgheresi, A., Balestrieri, F., Zaccara, G., Viggiano, M.P., Cincotta, M. & Ziemann, U. (2009) Modulation of interhemispheric inhibition by volitional motor activity: an ipsilateral silent period study. J. Physiol., 587, 5393–5410. Gueugneau, N., Mauvieux, B. & Papaxanthis, C. (2008) Circadian modulation of mentally simulated motor actions: implications for the potential use of motor imagery in rehabilitation. Neurorehabil. Neural Re., 23, 237–245. Gueugneau, N., Bove, M., Avanzino, L., Jacquin, A., Pozzo, T. & Papaxanthis, C. (2013) Interhemispheric inhibition during mental actions of different complexity. PLoS ONE, 8, e56973. H€ omberg, V. (2013) Neurorehabilitation approaches to facilitate motor recovery. Handb. Clin. Neurol., 110, 161–173. H€ ubers, A., Orekhov, Y. & Ziemann, U. (2008) Interhemispheric motor inhibition: its role in controlling electromyographic mirror activity. Eur. J. Neurosci., 28, 364–371. Hummel, F.C. & Cohen, L.G. (2006) Non-invasive brain stimulation: a new strategy to improve neurorehabilitation after stroke? Lancet Neurol., 5, 708–712. Kosslyn, S.M., Reiser, B.J., Farah, M.J. & Fliegel, S.L. (1983) Generating visual images: units and relations. J. Exp. Psychol., 112, 278–303. Kuhtz-Buschbeck, J.P., Mahnkopf, C., Holzknecht, C., Siebner, H., Ulmer, S. & Jansen, O. (2003) Effector-independent representations of simple and complex imagined finger movements: a combined fMRI and TMS study. Eur. J. Neurosci., 18, 3375–3387. Langhorne, P., Bernhardt, J. & Kwakkel, G. (2011) Stroke rehabilitation. Lancet, 14, 1693–1702. L€appchen, C.H., Ringer, T., Blessin, J., Seidel, G., Grieshammer, S., Lange, R. & Hamzei, F. (2012) Optical illusion alters M1 excitability after mirror therapy: a TMS study. J. Neurophysiol., 108, 2857–2861. Lepage, J.F., Morin-Moncet, O., Beaule, V., de Beaumont, L., Champoux, F. & Theoret, H. (2012) Occlusion of LTP-like plasticity in human primary motor cortex by action observation. PLoS ONE, 7, e38754. Li, J.Y., Espay, A.J., Gunraj, C.A., Pal, P.K., Cunic, D.I., Lang, A.E. & Chen, R. (2007) Interhemispheric and ipsilateral connections in Parkinson’s disease: relation to mirror movements. Movement Disord., 22, 813–821. Lotze, M., Braun, C., Birbaumer, N., Anders, S. & Cohen, L.G. (2003) Motor learning elicited by voluntary drive. Brain, 126, 866–872. M€ oller, C., Arai, N., L€ ucke, J. & Ziemann, U. (2009) Hysteresis effects on the input–output curve of motor evoked potentials. Clin. Neurophysiol., 120, 1003–1008. Murase, N., Duque, J., Mazzocchio, R. & Cohen, L.G. (2004) Influence of interhemispheric interactions on motor function in chronic stroke. Ann. Neurol., 55, 400–409.

© 2014 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 40, 2581–2588

2588 L. Avanzino et al. Nelson, A.J., Hoque, T., Gunraj, C., Ni, Z. & Chen, R. (2010) Impaired interhemispheric inhibition in writer’s cramp. Neurology, 75, 441–447. Ni, Z., Gunraj, C., Nelson, A.J., Yeh, I.J., Castillo, G., Hoque, T. & Chen, R. (2009) Two phases of interhemispheric inhibition between motor related cortical areas and the primary motor cortex in human. Cereb. Cortex, 19, 1654–1665. Nojima, I., Mima, T., Koganemaru, S., Thabit, M.N., Fukuyama, H. & Kawamata, T. (2012) Human motor plasticity induced by mirror visual feedback. J. Neurosci., 32, 1293–1300. Oldfield, R.C. (1971) The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia, 9, 97–113. Ramachandran, V.S. & Altschuler, E.L. (2009) The use of visual feedback, in particular mirror visual feedback, in restoring brain function. Brain, 132, 1693–1710. Ridding, M.C. & Rothwell, J.C. (1997) Stimulus/response curves as a method of measuring motor cortical excitability in man. Electroen. Clin. Neuro., 105, 340–344. Roosink, M. & Zijdewind, I. (2010) Corticospinal excitability during observation and imagery of simple and complex hand tasks: implications for motor rehabilitation. Behav. Brain Res., 213, 35–41. Rosenkranz, K., Kacar, A. & Rothwell, J.C. (2007a) Differential modulation of motor cortical plasticity and excitability in early and late phases of human motor learning. J. Neurosci., 27, 12058–12066. Rosenkranz, K., Williamon, A. & Rothwell, J.C. (2007b) Motorcortical excitability and synaptic plasticity is enhanced in professional musicians. J. Neurosci., 27, 5200–5206.

Rothgangel, A.S., Braun, S.M., Beurskens, A.J., Seitz, R.J. & Wade, D.T. (2011) The clinical aspects of mirror therapy in rehabilitation: a systematic review of the literature. Int. J. Rehabil. Res., 34, 1–13. Shiri, S., Feintuch, U., Lorber-Haddad, A., Moreh, E., Twito, D., TuchnerArieli, M. & Meiner, Z. (2012) Novel virtual reality system integrating online self-face viewing and mirror visual feedback for stroke rehabilitation: rationale and feasibility. Top. Stroke Rehabil., 19, 277–286. Sitburana, O., Wu, L.J., Sheffield, J.K., Davidson, A. & Jankovic, J. (2009) Motor overflow and mirror dystonia. Parkinsonism Relat. D., 15, 758– 761. Takeuchi, N., Oouchida, Y. & Izumi, S. (2012) Motor control and neural plasticity through interhemispheric interactions. Neural Plast., 2012, 823285. Thickbroom, G.W., Phillips, B.A., Morris, I., Byrnes, M.L. & Mastaglia, F.L. (1998) Isometric force-related activity in sensorimotor cortex measured with functional MRI. Exp. Brain Res., 121, 59–64. Touzalin-Chretien, P., Ehrler, S. & Dufour, A. (2010) Dominance of vision over proprioception on motor programming: evidence from ERP. Cereb. Cortex, 20, 2007–2016. Trompetto, C., Bove, M., Marinelli, L., Avanzino, L., Buccolieri, A. & Abbruzzese, G. (2004) Suppression of the transcallosal motor output: a transcranial magnetic stimulation study in healthy subjects. Exp. Brain Res., 158, 133–140. Wahl, M., H€ ubers, A., Lauterbach-Soon, B., Hattingen, E., Jung, P., Cohen, L.G. & Ziemann, U. (2011) Motor callosal disconnection in early relapsing–remitting multiple sclerosis. Hum. Brain Mapp., 32, 846–855.

© 2014 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 40, 2581–2588