Optimal motor performance requires the monitoring of sensorimotor input to ensure that the motor output matches current intentions. The brain is thought to be equipped with a “comparator” system, which monitors and detects the congruence between intended and actual movement; results of such a comparison can reach awareness. This study explored in healthy participants whether the cathodal transcranial direct current stimulation (tDCS) of the right premotor cortex (PM) and right posterior parietal cortex (PPC) can disrupt performance monitoring in a skilled motor task. Before and after tDCS, participants underwent a two-digit sequence motor task; in post-tDCS session, single-pulse TMS (sTMS) was applied to the right motor cortex, contralateral to the performing hand, with the aim of interfering with motor execution. Then, participants rated on a five-item questionnaire their performance at the motor task. Cathodal tDCS of PM (but not sham or PPC tDCS) impaired the participants' ability to evaluate their motor performance reliably, making them unconfident about their judgments. Congruently with the worsened motor performance induced by sTMS, participants reported to have committed more errors after sham and PPC tDCS; such a correlation was not significant after PM tDCS. In line with current computational and neuropsychological models of motor control and awareness, the present results show that a mechanism in the PM monitors and compares intended versus actual movements, evaluating their congruence. Cathodal tDCS of the PM impairs the activity of such a “comparator,” disrupting self-confidence about own motor performance.
Monitoring the more or less successful outcome of our actions is fundamental for optimal motor control and learning. Motor monitoring relies on internal representations of the intended, predicted, and actual action. According to “forward” models, when a motor command is issued to make a movement, such as a simple reaching or grasping movement, an efference copy of it is produced in parallel to predict the sensory consequences of the motor command, whenever the actual movement is made; hence, predictors are used to capture the forward, causal, relationship between actions and their outcomes. Conversely, according to the “inverse models,” controllers receive a desired configuration of the body and generate the appropriate motor commands, for instance, for reaching or grasping, so that the muscle activations achieve the desired trajectory (Wolpert & Kawato, 1998; Wolpert, 1997; review in Frith, Blakemore, & Wolpert, 2000). Optimal motor behavior requires both feed-forward and feed-back control by using an efference copy of the motor command; together with an internal model of the action goal, predictions of the consequences of actions can be generated; such predictions are used to monitor and control actions and constitute a signal based on which motor awareness is constructed (Blakemore, Wolpert, & Frith, 2002).
Motor awareness is largely dominated by representations of intended movements, rather than actual sensory information: Indeed, we are largely unaware of sensory feedback about the actual state of our motor system, as long as our motor intentions have been successfully realized (Jenkinson & Fotopoulou, 2010; Blakemore & Frith, 2003; Slachevsky et al., 2003; Blakemore et al., 2002; Frith et al., 2000; Fourneret & Jeannerod, 1998). We are usually aware of our motor performance and action goals, being able to detect any discrepancy between the planned and the performed movement; instead, we are typically unaware of all the fine motor adjustments necessary to perform a motor act (Blakemore & Frith, 2003).
Parietal and premotor areas of the right hemisphere seem to play a pivotal role in motor monitoring and awareness. Clues come firstly from classical neuropsychological evidence: Anosognosia for left hemiplegia, described by the French neurologist Josef Babinski in 1914 (see Langer & Levine, 2014; Papagno & Vallar, 2003), is a deficit in which awareness of the inability to perform movements is dramatically impaired by unilateral right brain damage. Anosognosic patients do not report their contralateral motor deficits, deny them when demonstrated by the examiner, even claiming that they are able to move their actually paralyzed limbs (Jenkinson & Fotopoulou, 2010; Vallar & Ronchi, 2006; Bisiach, Vallar, Perani, Papagno, & Berti, 1986). Anatomo-clinical correlation studies of anosognosia for hemiplegia indicate that a frontoparietal network in the right hemisphere, in particular the premotor cortex (PM), plays a major role in the monitoring and awareness of left-sided motor deficits (Berti et al., 2005; Pia, Neppi-Modona, Ricci, & Berti, 2004). On the other hand, lesions to the posterior parietal cortex (PPC) may alter the subjective experience of voluntary actions, in particular awareness of the intention to move (Sirigu et al., 2004).
The involvement of a premotor frontal-posterior parietal network in motor awareness is also supported by brain stimulation evidence. Studies with TMS in neurologically healthy individuals show that disrupting the activity of the left primary motor cortex (M1) substantially delays right-hand responses but has little effects on the overt judgment of such a delay, assessed by requiring participants to report the position of a rotating clock-hand, at which they have pressed the key. Conversely, TMS delivered to more anterior frontal areas (with the coil placed over FCz) produces smaller delays in actual RTs, but larger delays in overt judgments of the timing of the manual response, an ability related to motor awareness (Haggard & Magno, 1999). Another study showed no effects of TMS of the left PPC on both motor performance and motor awareness, when measured as the discrepancy between initial and reproduced visuomotor adjustments to a target shift (Johnson & Haggard, 2005). Desmurget and coworkers (2009) used direct electrical cerebral stimulation in patients with brain tumors, undergoing awake brain surgery, to assess the specific contribution of parietal and premotor cortical regions to conscious motor intention and awareness. Stimulation of the right PPC causes patients to desire and intend to move the left limb, hand, and foot and to report having moved, even in the absence of actual movement; stimulation of the left PPC has similar effects on lips–mouth movements and talking. Conversely, left and right PM stimulation triggers overt mouth and contralateral limb movements, which are not detected by the patients, who even firmly deny that they have moved (Desmurget et al., 2009). In a study in neurologically healthy participants, repetitive TMS over both M1 and dorsal PM brings about a movement illusion, in the absence of sensory feedback and accompanying movements (caused by ischemic nerve or spinal block); under the same conditions, repetitive TMS over the PPC does not evoke any sensation of movement. The movement illusion brought about by repetitive TMS to dorsal PM is unaffected by the block-induced absence of sensory feedback. Accordingly, stimulation of the dorsal PM may produce a corollary discharge, which is illusorily perceived as movement (Christensen et al., 2010). Taken together, neuropsychological studies in brain-damaged patients and direct and transcranial brain stimulation studies indicate that a PPC–PM network provides the neural underpinnings for the phenomenal experience of conscious motor intention (namely, the conscious desire to perform a motor act) and movement awareness, namely, the awareness that one is actually making, or has made, a movement (Desmurget & Sirigu, 2009).
On the basis of this evidence, we aimed at altering motor awareness by means of transcranial direct current stimulation (tDCS). tDCS is a neuromodulatory technique that, by delivering a weak current to the head, induces polarity-dependent shifts of spontaneous neuronal activity, ultimately modulating cortical excitability: Anodal tDCS increases cortical excitability, whereas cathodal tDCS decreases it (Brunoni et al., 2012; Paulus, 2011; Nitsche et al., 2008). This technique has been usefully used to modulate a number of cognitive and perceptual processes in healthy humans (Vallar & Bolognini, 2011). In the motor domain, motor execution, planning, and learning have been effectively enhanced or disrupted by tDCS in both unimpaired participants and brain-damaged patients (e.g., Bolognini et al., 2011, 2015; Convento, Bolognini, Fusaro, Lollo, & Vallar, 2014; Stagg et al., 2011; Reis et al., 2009; Boggio et al., 2006; Hummel et al., 2005). Instead, the potential of this technique for modulating action monitoring and awareness has not been explored so far.
In this study, we aimed at reducing the ability of healthy participants to monitor their own motor performance, and consequently to evaluate it accurately, through the cathodal tDCS of the PM and the PPC of the right hemisphere. At variance with previous TMS studies (Johnson & Haggard 2005; Haggard & Magno, 1999), which used implicit measures of motor awareness (e.g., visuomotor adjustments, judgments about latencies of manual responses), we adopted a more direct, explicit, index of motor monitoring, namely the self-evaluation of the motor performance, as assessed with an ad hoc questionnaire. In particular, during a two-digit sequence motor task, single-pulse TMS (sTMS) was delivered over the right M1 to interfere online with motor execution, but not with motor awareness (Haggard & Magno, 1999): By evoking involuntary muscles twitches in the contralateral left hand, sTMS of M1 was used to disrupt the finger movements of the left hand, required to perform the two-digit sequence. After completing the motor task, participants rated on a five-item questionnaire their ability to execute the motor task during the sTMS interference. In different sessions, immediately before the motor task with sTMS, sham or active cathodal tDCS was delivered over the PM or the PPC of the right hemisphere, with the aim of modulating the process of motor monitoring, in turn altering the participants' ability of evaluating their own motor performance, as measured with the questionnaire.
Twenty-two healthy volunteers (eight men, mean age = 25.5 years, SD = 4.6, range = 21–38 years), all right-handed (Oldfield, 1971), took part in the study. Participants were naive as to the purpose of the experiment; none of them had history or evidence of neurological, psychiatric, other relevant medical problems or any contraindications to noninvasive brain stimulation, which was applied as indicated by international safety guidelines (Rossi, Hallett, Rossini, Pascual-Leone, & the Safety of TMS Consensus Group, 2009; Poreisz, Boros, Antal, & Paulus, 2007). All participants provided written informed consent before the experiment, which was performed in accordance with the ethical standards of the Declaration of Helsinki (BMJ 1991; 302: 1194) and was approved by the ethics committee of the University of Milano-Bicocca.
Motor Task and Motor Performance Awareness Questionnaire
Participants were seated on a chair, approximately 70 cm away from a computer screen. On each trial, a two-digit sequence was displayed on the computer screen, against a dark background (luminance = 0.1 cd/m2), indicating which keys had to be pressed on the computer keyboard. Each number indicated a key press with the (1) ring, (2) middle, or (3) index finger. Participants were instructed to look at the fixation point (a red cross) at the center of the monitor, which was displayed for a random duration between 5 and 7 sec; when the two numbers appeared, participants had to press the corresponding sequence of keys, as quickly and as accurately as possible with their left hand. The target two-digit sequence remained on the screen until the participant's response was given, namely, two keys pressed in sequence regardless of the correctness of the response. Participants did not receive any explicit feedback about their performance accuracy and speed both during each trial and at the end of the task. In this paradigm, an error corresponded to the erroneous reproduction of the two-digit sequence, either because the two correct keys had been pressed in the wrong order (e.g., target: 2 3, response: 3 2) or because a wrong key had been pressed (e.g., target: 2 3, response: 1 3); no omission could occur, because the next trial started only after a response was given. RTs (in sec) were also recorded, measuring the mean execution time required to reproduce the two-digit sequences, regardless of the correctness of the responses.
The motor task was performed in two sessions (see the description of the Experimental Procedure), between which tDCS was delivered. In the second session, to interfere with the participants' performance during the motor task, on each trial, at the appearance of the two-digit sequence, sTMS (intensity = 120% of the MT) was delivered to right M1, contralateral to the left-hand performing the task. In this way, by evoking involuntary muscle twitches in the contralateral left hand, sTMS of M1 disrupted the participant's trial-by-trial motor responses during the task. sTMS was delivered through a stimulating figure-of-eight coil (Ø = 80 mm), connected to a Magstim Super Rapid transcranial magnetic stimulator (Magstim Company, Whitland, UK). In every tDCS session (see below), the individual resting motor threshold was determined for each participant by stimulating the hand area of right M1 with sTMS, and it was defined as the minimum intensity that elicited visually detectable motor twitches in the fingers of the resting right hand in three of six consecutive sTMS pulses (e.g., Bolognini & Ro, 2010).
In each session (with or without sTMS, see below), the motor task comprised 24 trials and lasted about 4 min. Stimulus presentation, the delivery of sTMS, and response recording were under computer control (E-Prime software, Psychology Software Tools, Inc., Pittsburgh, PA, https://www.pstnet.com/eprime.cfm).
Immediately at the end of the motor task, in every session, the participants' awareness of motor performance was assessed with an ad hoc questionnaire, namely, the Motor Performance Awareness Questionnaire (MPAQ), comprising the following items: (1) “I made many errors”; (2) “The task was easy”; (3) “Coordinating fingers' movements was difficult”; (4) “My performance was fast”; (5) “I'm sure of the answers that I gave previously”. For each item, participants rated their agreement using a 100-mm Visual Analogue Scale (VAS), with the left end indicating “absolutely disagree” (0 mm) and the right end “absolutely agree” (100 mm).
Transcranial Direct Current Stimulation
A double-blind, sham-controlled design for tDCS application was used. tDCS was delivered using a battery-driven, constant current stimulator (BrainStim, E.M.S., Bologna, Italy; brainstim.it), through a pair of saline-sponge electrodes (25 cm2, 5 × 5 cm). For the stimulation of the right PPC, the cathodal electrode was placed over P4 (according to the 10/20 EEG system; e.g., Bolognini, Olgiati, Maravita, Ferraro, & Fregni, 2013), and for the stimulation of the PM, the cathodal electrode was placed 3 cm anterior to the C4 (e.g., Bolognini, Miniussi, Gallo, & Vallar, 2013; Boros, Poreisz, Münchau, Paulus, & Nitsche, 2008). In both cases, the reference electrode (anodal) was placed over the left supraorbital area. Cathodal, active, tDCS was delivered for 6 min at an intensity of 1.5 mA (i.e., fade-in/fade-out phases = 10 sec; Nitsche et al., 2003). Following a standard method for effective blinding (Gandiga, Hummel, & Cohen, 2006), sham tDCS was applied with the same stimulation parameters and the same electrode montage of active tDCS (i.e., placing the cathode over one of the two target areas in a random order across participants), but the current was ramped up over 30 sec and then immediately turned off. The sham and active modes were activated through the use of codes, which were set and conserved by the principal investigator (N.B.), who did not participate in data collection. Through such codes, the device was activated by the experimenter, showing on the display the intensity (1.5 mA) and duration (6 min) of tDCS, independent of the type of stimulation (active vs. sham). This method has been shown to be reliable for keeping both the experimenter and the participant blind to sham versus real tDCS (Gandiga et al., 2006).
Each participant underwent three tDCS sessions, counterbalanced across participants: (1) cathodal tDCS over the right PM, (2) cathodal tDCS over the right PPC, and (3) sham tDCS over the right PM or PPC (randomized across participants). The three sessions were separated by at least 24 hr to minimize carryover effects and to guarantee a sufficient washout of the effects of the previous run (e.g., Bolognini et al., 2015; Bolognini, Olgiati, et al., 2013).
In each tDCS session, firstly participants underwent a training session (18 trials) to familiarize with the motor task. Then, participants performed the motor task without online sTMS to the right M1 before the delivery of tDCS; immediately at the end of the motor task, they completed the MPAQ (i.e., pre-tDCS session). Next, tDCS was delivered. Immediately at the end of tDCS, participants performed again the motor task, during which sTMS to M1 was delivered on each trial; at the end of the task, participants completed again the MPAQ (i.e., post-tDCS session).
Overall, the duration of each tDCS session (comprising both the pre- and post-tDCS sessions) was about 40 min. The timeline of the experimental procedure is illustrated in Figure 1.
Statistical analyses were performed using Statistica Software (Statsoft, version 10.0, Tulsa, OK). The mean percentage of the error rate and the mean RT at the motor task, as well as the average VAS scores (in mm) at each item of the MPAQ, were analyzed by repeated-measures ANOVA, with tDCS (PM, PPC, Sham) and Time (pre-tDCS, post-tDCS) as within-subject factors. Multiple, post hoc comparisons were performed with the Newman–Keuls test. The partial eta square (pη2), which measures the proportion of total variance attributable to a main factor or to an interaction (Cohen, 1973), was also calculated for each repeated-measures ANOVA.
We looked also for an association between the participants' motor performance and its subjective evaluation at MPAQ, in both the pre-tDCS and post-tDCS sessions of each tDCS condition. To this aim, for each item of the MPAQ, Pearson's correlations were performed between the error rate at the motor task and the participants' evaluation of their performance, as assessed with the MPAQ's VAS. For Item 4, the correlation was performed considering RTs, because this item specifically inquired about the speed of motor performance. Correlation analyses were corrected for multiple comparisons (α/6 = 0.008).
Before running the analyses, for each data set (performance at motor task and VAS scores), the extreme Studentized deviate method (i.e., Grubbs' test; Grubbs, 1969) was used to detect outliers; no significant outlier was found in our data (all ps > .05); accordingly, all participants were included in the analyses.
Accuracy and Speed at the Motor Task
The analysis of error rate at the motor task showed a significant main effect of Time (F(1, 21) = 4.28, p < .05; pη2 = .2): Regardless of the tDCS condition, errors at the motor task increased in the post-tDCS session, when sTMS of M1 was applied during the motor task (post-tDCS, mean = 12%, SD = 9%), as compared with the performance in the pre-tDCS session, without sTMS (8% ± 10%). The main effect of tDCS (F(2, 42) = .86, p = .42; pη2 = .04) and the tDCS × Time interaction (F(2, 42) = 1.79, p = .18; pη2 = .21) did not reach significance. The analysis of the RTs at the motor task showed that the main effect of Time approached the .05 level of significance (F(1, 21) = 4.14, p = .055; pη2 = .16), with a trend toward an increase of the response time in every post-tDCS session (pre-tDCS, without M1-sTMS = 8.39 ± 1.53 sec vs. post-tDCS with online sTMS of M1 = 8.75 ± 1.67 sec). The main effects of tDCS (F(2, 42) = .07, p = .92; pη2 = .001) and the tDCS × Time interaction (F(2, 42) = 0.58, p = .56; pη2 = .13) were not significant. These results demonstrate that sTMS applied to the right M1 effectively interfered with performance accuracy and speed at the motor task, with no differences among tDCS conditions.
Self-evaluation of the Motor Performance at the MPAQ
Item 1 (“I made many errors”)
Only the main effect of Time was significant (F(1, 21) = 15.77, p < .001; pη2 = .43). The main effects of tDCS (F(2, 42) = 0.51, p = .60; pη2 = .02) and the tDCS × Time interaction (F(2, 42) = .21, p = .82; pη2 = .01) did not reach the significance level. Participants scored higher; hence, they strongly agreed to have committed many errors in the post-tDCS session, when the motor task was performed while receiving sTMS over M1 (39.9 mm), than in the pre-tDCS session, when the motor task was performed without the sTMS interference (23.7 mm).
Item 2 (“The task was easy”)
Again, only the main effect of Time was significant (F(1, 21) = 10.75, p < .01; pη2 = .34), whereas the main effects of tDCS (F(2, 42) = .24, p = .79; pη2 = .01) and the tDCS × Time (F(2, 42) = .40, p = .67; pη2 = .02) interaction were not. Participants judged the motor task to be easier (69.97 mm) in the pre-tDCS session than in the post-tDCS session (with online sTMS, 60.17 mm).
Item 3 (“Coordinating the fingers' movements was difficult”)
The main effect of Time (F(1, 21) = 80.45, p < .0001, pη2 = .79) was significant, whereas both the main effect of tDCS (F(2, 42) = .12, p = .89; pη2 = .01) and the tDCS × Time interaction (F(2, 42) = .27, p = .76; pη2 = .02) were not. Participants judged more difficult to coordinate the movements of the fingers in the post-tDCS session, during online sTMS (60.80 mm), as compared with the pre-tDCS session, without sTMS (32.09 mm).
Item 4 (“My performance was fast”)
The main effect of Time (F(1, 21) = 9.93, p < .005; pη2 = .34) was significant, whereas the main effect of tDCS (F(2, 42) = .42, p = .66; pη2 = .02) and the tDCS × Time (F(2, 42) = .05, p = .94; pη2 = .001) interaction did not attain significance. Participants rated their performance faster in the pre-tDCS session (without sTMS, 73.30 mm) than in the post-tDCS (with online sTMS, 65.92 mm).
Item 5 (“I'm sure of the answers that I gave previously”)
The main effect of Time (F(1, 21) = 10.41, p < .005, pη2 = .25) was significant, whereas the main effect of tDCS did not attain significance (F(2, 42) = 1.49, p = .2; pη2 = .07). The tDCS × Time interaction (F(2, 42) = 3.74, p < .03; pη2 = .33) was significant. Post hoc comparisons showed that participants reported to be more insecure about their own responses given at the previous MPAQ's items after cathodal tDCS of PM (post-tDCS = 74.09 mm, p < .001), as compared with the pre-tDCS session (87.27 mm); no difference between pre-tDCS and post-tDCS sessions emerged for sham tDCS (84.32 vs. 81.05 mm, p = .52) and PPC tDCS (85.18 vs. 82.18 mm, p = .61). The VAS score in the post-tDCS session of PM tDCS was significantly lower, as compared with all the other conditions (all ps < .02).
Correlations between Motor Performance and Its Evaluation
As shown in Figure 3, significant positive correlations were found between the VAS score at Item 1 (“I made many errors”) and the error rate in the post-tDCS sessions of sham tDCS (r = .59, p = .004) and of PPC tDCS (post-tDCS, r = .69, p = .001). This finding indicates that participants were aware of their worsened motor performance after sham tDCS or cathodal tDCS over PPC, because they congruently reported to have made more errors (due to the sTMS interference) when they had actually performed worst. Instead, after PM tDCS, the correlation between error rate and the VAS score at Item 1 was not significant (r = .22, p = .32). This finding is suggestive of a disruption of motor monitoring by PM tDCS: After cathodal stimulation of PM, participants became less reliable in judging the amount of errors committed at the motor task, being unable to detect to have made more errors (due to sTMS of M1) during the task, and thus underestimating error rate, in the face of a worsened performance at the motor task. Correlations between the VAS score at Item 1 and error rate in the pre-tDCS sessions were not significant (Sham tDCS, r = .43; PPC tDCS, r = .53; PM tDCS, r = .54, all ps > .05). This finding, together with the significant correlations in the Sham and PPC post-tDCS sessions, indicates that sTMS increases error rate in the motor task and awareness of it, supporting the validity of the approach used in this study.
No other correlation was significant (error rate and VAS scores at Item 2 p > .42, Item 3 all ps > .12, Item 5 all ps > .15; RT and VAS score at Item 4, all ps > .11).
Cathodal tDCS of the right PM impairs motor monitoring, making participants more insecure in explicitly judging their own motor performance, as well disrupting the association between level of accuracy at the motor task and the explicit estimation of committed errors. In particular, when motor execution is disrupted by online sTMS of the right M1 (contralateral to the left hand used to perform the motor task), participants congruently report the increase of task difficulty (post-tDCS session), with respect to the judgments they gave when the motor task was performed without the sTMS interference (pre-tDCS session). Indeed, at the MPAQ, participants report to have made more errors during the motor task (Item 1), which was rated as more difficult (Item 2), and an increased difficulty in coordinating fingers' movements (Item 3); they also judged their performance as slower (Item 4). These reports were consistent across the tDCS sessions, after sham or active tDCS over the PPC and the PM. However, when specifically asked to rate their self-confidence in such judgments (Item 5: “I'm sure of the answers that I gave previously”), participants report a greater uncertainty about their motor performance evaluation only after cathodal tDCS over PM. This effect is absent after sham and PPC tDCS, where self-confidence in motor performance evaluation does not change, as compared with the pre-tDCS assessment. These findings suggest an impairment of motor monitoring mechanisms brought about by the cathodal tDCS modulation of PM activity in the right hemisphere.
Moreover, only after the sham tDCS and the PPC tDCS, a significant positive correlation between increased error rate induced by M1-sTMS and the participants' report of having committed more errors (MPAQ's Item 1) was found. Such a correlation became weaker and no longer significant after PM tDCS. This evidence further supports the claim that cathodal tDCS of the right PM impairs the participants' ability to detect the discrepancy between planned and performed movements, as indexed by the increased error rate in the motor task.
Optimal motor performance requires a mechanism that monitors sensorimotor inputs to ensure that motor output matches current intentions. Computational models of the motor system suggest that the brain is equipped with “comparators,” one of which monitors and detects the congruence between intended and actual movement (Blakemore et al., 2002; Frith et al., 2000). In this view, motor monitoring is usually implicit and automatic but becomes conscious whenever there is a mismatch between expected and realized sensorimotor states (Fink et al., 1999). A neural locus for such a “comparator” for motor performance in the human brain is suggested by functional neuroimaging experiments, showing that right dorsolateral prefrontal cortex (DLPFC) is activated when actions are to be maintained in the face of a conflict between motor intention and sensory outcome (Fink et al., 1999). Within this framework, by evoking motor twitches in the hand performing the motor task, sTMS of M1 produces a mismatch between the expected (the two-sequence digit to be reproduced) and the actual motor output (altered by the motor twitches); this mismatch provides the “comparator” with an error signal normally used to inform participants and to update their awareness of motor performance (Jenkinson & Fotopoulou, 2010; Blakemore et al., 2002; Frith et al., 2000). The cathodal tDCS of PM may disrupt the activity of this “comparator,” directly affecting the operation of the motor monitoring system that consists in the detection of the mismatch between the actual motor output and intended action goal. The consequence is that participants become less confident in estimating the reliability of their judgments (Item 5 of the MPAQ), which are related to their performance accuracy and speed at the motor task (Items 1 and 4), its difficulty (Item 2), and the ability in coordinating the fingers' movements (Item 3). Instead, during sham and cathodal PPC tDCS, the discrepancy between intended and actual movements is appropriately detected by the “comparator,” with participants feeling confident about their judgments at the MPAQ (Item 5). The proposal of a breakdown of the “comparator” by stimulation of the right PM is further supported by the lack of an association between the participants' increased error rate at the motor task and their subjective report of having committed more errors following cathodal PM tDCS: This finding is indicative of impaired motor awareness.
Likely due to its feature of being a neuromodulatory technique (Miniussi & Ruzzoli, 2013; Brunoni et al., 2012; Nitsche et al., 2008), premotor tDCS does not completely abolish motor monitoring (i.e., no change for MPAQ's Items 1–4); rather, its effect is one of reducing the participants' self-confidence in the evaluation of their motor behavior (i.e., change at Item 5).
The finding that the increase of error rate, brought about by sTMS of M1 in every tDCS session, was small (on average 4%), may be accounted for in terms of the easiness of the two-digit sequence motor task, where participants likely learned soon how to compensate for the motor twitches. However, it is also possible that cathodal tDCS of both PM and PPC failed to interfere with rapid error corrections based on discrepancies between intended and predicted motor output; this hypothesis, based on a negative finding, deserves further investigations.
Our results are in agreement with neuropsychological models of anosognosia for hemiplegia. An influential account proposes that patients with this neurological disorder are able to form appropriate representations of their intended movements, but they are unaware of the discrepancy between intended and actual movement (Garbarini, Piedimonte, Dotta, Pia, & Berti, 2013; Fotopoulou et al., 2008; Vallar & Ronchi, 2006). The failure to detect such a discrepancy is the result of a damaged comparator mechanism, located in the right PM, which is dedicated to the conscious monitoring of motor acts (Berti et al., 2005). Our results support this proposal, if one hypothesizes that sTMS of M1 disrupts motor execution, but not motor intention, roughly simulating the motor impairment of hemiplegia (namely, the inability to move) and cathodal tDCS over PM brings about a malfunctioning of the “comparator,” then impairing the self-monitoring of the TMS-induced motor impairment.
Cathodal tDCS over the right PPC does not affect motor monitoring. Rather, in healthy participants, this stimulation brings about behavioral changes resembling unilateral spatial neglect (Giglia et al., 2011; Sparing et al., 2009). Instead, when the left hemisphere is targeted, cathodal tDCS of the PPC impairs the ability to plan a motor act (Convento et al., 2014). In line with these findings in healthy participants, in stroke patients anodal tDCS over the PPC of the right damaged hemisphere ameliorates left visuospatial neglect (Sparing et al., 2009), whereas anodal tDCS over the PPC of the left damaged hemisphere improves ideomotor apraxia (Bolognini et al., 2015).
Gandola and coworkers (2014) have reported a transient improvement of anosognosia for hemiplegia in a chronic stroke patient with a bilateral cerebral lesion, after anodal tDCS over the right prefrontal cortex. The improvement of motor awareness by tDCS was observed only when the patient had to actually perform a movement with eyes open. No change took place when the patient did not attempt to perform the movement or when the movement was executed with eyes closed. In this single case study, the dependence of the tDCS effects on the visual feedback, which informed the anosognosic patient that his hand was not moving, suggests that motor awareness relies on the interactions between top–down and bottom–up signals: Sensory inputs are likely to be used for comparing expected and actual movements based on which motor awareness is built up (Berti, Spinazzola, Pia, & Rabuffetti, 2007). In this perspective, the anodal tDCS of areas involved in motor monitoring may be able to restore the ability of the motor “comparator” system to correctly appreciate afferent information, consequently inducing a transient remission of unawareness of left hemiplegia (Gandola et al., 2014). In our study, no sensory feedback was provided to participants; the manipulation of sensory feedback, given online and offline during the motor task, could be valuable to further explore, and maybe to amplify, the effects of tDCS found in our study.
In conclusion, goal-directed action necessitates a mechanism that monitors sensorimotor inputs to ensure that motor output matches current intentions. Such a monitoring is made less efficient by modulating, through cathodal tDCS, the activity of the right PM; the result is an impaired ability to evaluate motor performance accurately. This novel evidence paves the way to the use of tDCS for the rehabilitation of action monitoring and awareness deficits, which represent invalidating symptoms of many neurological and psychiatric disorders, such as anosognosia for hemiplegia, anarchic hand syndrome, utilization behavior, and delusions of control (Vallar & Ronchi 2006; Frith et al., 2000).
We are grateful to Vito Giorgio Catania and Martina Oldrini for their assistance and Daniele Romano for helpful suggestions. This work has been supported in part by a FAR grant (2014-ATE-0209) from the University of Milano-Bicocca to N. B. and by Ricerca Corrente grants from the IRCCS Istituto Auxologico Italiano to G. V.
Reprint requests should be sent to Nadia Bolognini, Department of Psychology, University of Milano-Bicocca, Building U6, Piazza Dell'Ateneo Nuovo 1, 20126, Milano, Italy, or via e-mail: firstname.lastname@example.org.