Improved isometric force endurance after transcranial direct current stimulation over the human motor cortical areas

Introduction

Neuromuscular fatigue is the exercise-dependent decrease in muscle force. It involves both peripheral and central factors (Gandevia, 2001). Whereas peripheral fatigue has been defined as a loss of force related to processes lying at or distal to the neuromuscular junction, central fatigue represents a progressive failure of the nervous system to drive the muscle maximally during exercise (Taylor et al., 2006). ‘Central fatigue’ is relevant for daily motor activities in normal conditions and has a major role in several disorders of the central nervous system. Central fatigue has been related both to changes in the activity of spinal motoneurons due to modulation of afferent inputs or intrinsic cellular properties (spinal fatigue) and to reduced supraspinal drive (supraspinal fatigue). Evidence from studies using transcranial magnetic stimulation (TMS) and transcranial electrical stimulation (TES) suggests that the development of reduced supraspinal drive is accompanied by changes in motor cortex excitability (for a review see Gandevia, 2001).

Transcranial direct current stimulation (tDCS) is a noninvasive technique for modulating brain function (Priori et al., 1998; Nitsche & Paulus, 2000; Priori, 2003; Paulus, 2004; Wassermann & Grafman, 2005). Weak electrical direct current (< 1.5 mA) applied over the scalp induces prolonged changes in brain excitability persisting long after the current has ceased (Nitsche & Paulus, 2001; Nitsche et al., 2003a). In particular, several studies, investigating changes in motor evoked potentials (MEPs) elicited by TMS, have shown that tDCS delivered over the motor cortex modulates corticospinal output (Priori et al., 1998; Nitsche & Paulus, 2000, 2001; Priori, 2003;Paulus, 2004; Ardolino et al., 2005). The after-effects of brain polarization on motor cortical excitability are due to both synaptic and nonsynaptic functional changes (Liebetanz et al., 2002; Ardolino et al., 2005).

In this study we investigated whether brain polarization modulates neuromuscular fatigue. To do so we evaluated the after-effects of anodal and cathodal tDCS over the motor areas of the cerebral cortex in healthy subjects on the endurance time (Et) for a submaximal isometric contraction of elbow flexors. We also assessed a group of control subjects without stimulation. Finally, in an additional experiment to investigate the mechanisms underlying the action of tDCS on fatigue, we delivered anodal tDCS and evaluated the changes in MEPs elicited by TMS.

Materials and methods

Force measurement

Subjects sat in a chair with the left elbow on a padded support, the elbow joint at a right angle, and the wrist connected to a load cell (BC 302; DS Europe, Milan, Italy; Fig. 1A). The motor task required the subject to exert an upward-directed force with the wrist activating the elbow flexor muscles (biceps brachii, brachialis, coracobrachialis, and brachioradialis). To determine the maximal voluntary contraction (MVC) force, subjects were instructed to increase the force from zero to maximum and to hold it for 3 s. Subjects performed nine MVC trials subdivided in three sessions and rested for 60 s between successive trials. The mean of the best three performances was taken as the MVC force and used as the reference to calculate the target force for the fatiguing contraction.

Muscle endurance was assessed as the time for which subjects could sustain an isometric contraction with the elbow flexor muscles at 35% of the MVC force (‘time to task failure’), previously determined. During the fatiguing task each subject received force feedback on a computer monitor showing a 35% MVC target. The subject was encouraged to keep the force at this level.

The contraction was terminated when the subject deviated from the target force (the subject remained below the force marker on the monitor) for > 3 s despite strong verbal encouragement.

The signal from the load cell (range: 0–60 kg; sensitivity, 2 mV/V; total error, < 0.5%; repeatability, < 0.1%) was preamplified and low-pass filtered (< 1000 Hz) with an analogue amplifier [Signal Conditioner Cambridge 1902; Cambridge Electronic Design (CED), Cambridge, England]. The output signal was digitized (Cambridge Micro 1402; CED) with a sampling rate of 2500 Hz and 12-bit quantization with 5 V range. The digitized signal was stored on a personal computer and displayed by the Spike2 software (version 5.11; CED).

Results

None of the subjects complained of adverse reactions to tDCS.

Experiment 1

We found no significant differences in baseline MVC and Et between anodal, cathodal and controls (Et: P = 0.317; cathodal tDCS, 261.88 ± 18.9; anodal tDCS, 258.22 ± 22.76; controls, 226.47 ± 14.91 s; MVC: P = 0.73; cathodal tDCS, 156 ± 18; anodal tDCS, 154.4 ± 18.07; controls, 148.8 ± 6.64 N).

The two-way anova on MVC disclosed a significant effect of the factor time (P < 0.0001), baseline MVC being significantly higher than MVC evaluated after stimulation (T0, 152.29 ± 7.44 vs. T1, 140.18 ± 7.12 N). Neither MVC nor n-MVC were significantly affected by the factor stimulation (MVC: P = 0.9; anodal tDCS: T0, 154.4 ± 18.07; T1, 142.8 ± 16.62 N; cathodal tDCS: T0, 156 ± 18.75; T1, 141.86 ± 17.53 N; controls: T0, 148.8 ± 6.64; T1, 137.6 ± 7.36 N; n-MVC: P = 0.77; Anodal tDCS, −7.11 ± 2.36; cathodal tDCS, −9.44 ± 1.51; controls, −7.86 ± 2.09%) (Fig. 2A).

The two-way anova disclosed a significant effect of time (P < 0.0001) on Et, Et being significantly higher at T1 than at baseline for all groups (cathodal tDCS: T0, 261.88 ± 18.9; T1, 165.44 ± 12.08 s; anodal tDCS: T0, 258.22 ± 22.76; T1, 205.22 ± 24.92 s; controls: T0, 226.47 ± 14.91; T1, 136.13 ± 14.91 s; 2, 3). Whereas the factor stimulation was not significant (P = 0.071), the stimulation × time interaction was significant (P = 0.019). Although no difference was found between n-Et values at 1 h after cathodal tDCS and no stimulation (−35.77 ± 3.39 vs. −39.33 ± 3.32%; P = 0.53), anodal tDCS induced a significantly smaller decrease in Et at 1 h than did no stimulation (−21.11 ± 5.52%; P = 0.0029) and cathodal tDCS (P = 0.015; Fig. 2C).

EMG slope and y-intercept were significantly higher at T1 than at T0 for both the polarization and the control group (anova main factor time: slope, P = 0.0002; y-intercept, P < 0.00001). The two-factor anova disclosed no significant differences due to either the factor ‘stimulation’ or the time × stimulation interaction (Fig. 2D).

Experiment 2

After the end of anodal tDCS the amplitude of MEPs during a slight isometric biceps brachii contraction (5% of MVC) increased significantly from baseline (T0, 1.26 ± 0.1; T1, 1.59 ± 0.12 mV; P = 0.0156; Fig. 4).

Discussion

Our findings in healthy subjects show that anodal tDCS delivered over the cortical motor areas induced a relative increase in the flexor muscles' Et for a sustained submaximal isometric contraction: at the postconditioning evaluation (1 h after the baseline fatiguing task), the expected shortening of the Et owing to residual tiredness was reduced by ∼15% after anodal tDCS compared with cathodal tDCS and with no stimulation, showing that brain polarization can modulate neuromuscular fatigue.

Our data are in line with recent data demonstrating that anodal tDCS delivered over the motor cortex improves hand function in healthy subjects (Boggio et al., 2006) and in patients with stroke (Hummel & Cohen, 2005). Benwell et al. (2006) have shown that paired-pulse rTMS can modulate the rate of force loss in a brief MVC task of the hand, suggesting that the rTMS-induced manipulation of cortical excitability improves muscle force.

Studies using TMS and TES have reported exercise-dependent changes in motor cortex excitability. During and immediately after brief or fatiguing exercise, whereas motor responses to TMS increase in amplitude (Brasil-Neto et al., 1993; Samii et al., 1996), those elicited by TES decrease in size (Modugno et al., 1998; Gandevia et al., 1999). These data suggest a motor cortical facilitation related to voluntary contraction which offsets alpha-motoneuron relative inhibition. The increased cortical excitability outlasts the fatiguing effort by a few minutes (postcontraction facilitation) and is followed by a depression of the MEPs (postcontraction depression) not related to spinal motoneuron inhibition (Brasil-Neto et al., 1993). Further support for the existence of focal cortical changes related to central fatigue comes from studies on the cortical silent period. During prolonged MVC or submaximal efforts the silent period which follows the motor response evoked by TMS progressively lengthens (Taylor et al., 1996; Taylor et al., 1999). This phenomenon has been interpreted as an increase in local intracortical inhibition. Because the relationship between these cortical events and the development of supraspinal fatigue and task failure is unclear, some have suggested that changes in cortical excitability are epiphenomena and not causal factors in motor fatigue (Gandevia et al., 1996; Gandevia, 2001). Our findings on the relative Et prolongation induced by anodal tDCS of motor cortex support the hypothesis that manipulation of cortical function by tDCS may play a role in determining task failure during sustained effort.

The Et changes we observed after anodal tDCS could have several explanations. The most obvious is that anodal tDCS could modulate supraspinal fatigue by balancing the aforementioned changes in motor cortex excitability. Anodal and cathodal scalp tDCS induce well-known changes in the MEP size (Nitsche & Paulus, 2000; Priori, 2003) and TMS motor threshold (Ardolino et al., 2005). Because we used a different electrode setup (motor cortex–right deltoid) from previous studies, to evaluate the effect of anodal tDCS on motor cortex excitability we conducted an additional experiment. We found that the amplitude of MEPs evoked by TMS had increased by nearly 30% at the end of anodal stimulation. This result is similar to that observed by others using the ‘motor cortex–contralateral forehead’ setup (Liebetanz et al., 2002; Lang et al., 2005), and confirms an increase in motor cortex excitability related to anodal polarization. Hence, anodal tDCS may increase supraspinal drive by inducing a prolonged facilitation of corticospinal neurons, thereby prolonging Et.

As expected for the presence of a residual tiredness due to the basal fatiguing task, EMG slope and y-intercept were significantly higher at postconditioning evaluation for both the polarization and the control group but statistical analysis did not show any differences between groups related to the factor ‘stimulation’. Why surface EMG variables in biceps brachii remained unchanged despite the Et prolongation remains unclear. First, the evaluated EMG parameters may be too insensitive to detect subtle changes in central recruitment of spinal motorneurons. Second, the effect of anodal tDCS on the force endurance may arise not from a postsynaptic effect on corticomotoneuronal projections but could be related to a presynaptic effect on the motor cortex interneuronal network. Moreover, we cannot rule out possible concomitant mechanisms involving cortical areas neighbouring the active cephalic electrode. In this regard, imaging studies showed that polarization of motor cortex induces widespread activation changes (Baudewig et al., 2001; Lang et al., 2005). The decline in voluntary activation during prolonged effort is thought to be related to ‘upstream’ failure of the motor cortical neuron (Gandevia et al., 1996; Gandevia, 2001; Taylor et al., 2006). Furthermore, brain activation measured by functional magnetic resonance imaging during a sustained maximal contraction shows an initial increase followed by a significant reduction, seen not only in the primary sensorimotor areas but also in the secondary and association cortices (Liu et al., 2002). Hence, t-DCS might act by modulating nonprimary premotor motor areas. A further possibility is that the prolonged Et may be related to the tDCS-induced modulation of the feedback inhibitory systems limiting motor cortical output to ‘protect’ the motor system from overload. tDCS of the motor cortex has been reported to decrease pain (Fregni et al., 2006a) and pain afferent input from muscle tissue is probably involved in muscle fatigue (Gandevia et al., 1996; Gandevia, 2001). The increased Et after anodal tDCS could therefore depend on a reduced pain sensation arising from muscle tissue during sustained and prolonged voluntary contraction. A similar mechanism is involved in the well-known ergogenic effects of caffeine during prolonged efforts (Plaskett & Cafarelli, 2001). Another possible mechanism is increased voluntary activation. Because tDCS delivered over the dorsolateral prefrontal cortex improves depression (Fregni et al., 2006b), although none of our subjects reported consistent mood changes or euphoria after brain polarization, tDCS could increase motivation and the voluntary drive for sustained and prolonged muscle contraction. Finally, because tDCS induces persistent increases in intermuscular coherence (Power et al., 2006), the prolonged Et could also arise from a more optimal coupling of synergistic flexor muscles (biceps brachii, brachialis, coracobrachialis and brachioradialis).

The net effect (increased or reduced excitability) produced by tDCS depends on several technical issues including the relative position of the reference electrode as well as the orientation of excitable tissue with respect to the electrical field (Priori, 2003). In this study, we used an extracephalic reference electrode positioned on the right deltoid (Priori et al., 2007). This experimental setup differs from that used in previous works (Nitsche & Paulus, 2000; Priori, 2003; Ardolino et al., 2005; Hummel & Cohen, 2005; Boggio et al., 2006). Modelling studies suggested that current flow is localized beneath the stimulating electrodes, both the anode and the cathode (Nathan et al., 1993; Miranda et al., 2006). Thus, anodal tDCS of one cortical area is combined with cathodal stimulation of another, and vice versa (Nitsche et al., 2007). Using the conventional ‘motor cortex–contralateral forehead’ setup with a complex motor behaviour such as a fatiguing task would generate confusion regarding the source of the observed effect as it could originate from the active as well as the cephalic reference electrode. In this regard, a single cephalic electrode on the scalp helps to resolve this ambiguity, as the observed aftereffects are unlikely to be related to the extracephalic reference electrode.

Among critical technical points to consider in choosing the extracephalic electrode setup are safety criteria and possible charge flow to brainstem (Nitsche et al., 2003b). Given our stimulating dipole geometry, we cannot a priori exclude brainstem involvement. Nevertheless, this possibility seems unlikely for several reasons. First, in accordance with the results reported by Accornero et al. (2007) our experimental setup left heart rate and body temperature unchanged and none of our subjects showed signs or symptoms attributable to the modulation of brainstem activity. Second, most of the current delivered by transcranial stimulation gets shunted through the scalp and the current density delivered to the cortex by the stimulation decreases rapidly with depth (Nathan et al., 1993). Lippold & Redfearn (1964) reported one case of disturbed breathing, speech arrest and psychosis possibly related to brainstem excitability changes after direct current stimulation using an extracephalic electrode setup (active electrodes above each eyebrow and reference electrode above the right knee). Nonetheless, these authors applied battery-delivered small currents (100–500 µA) for long time periods (up to 5 h), delivering values of total charge a thousand-fold higher than in our experiments. Moreover, previous experiences on humans during surgery demonstrated that for eliciting a motor evoked potential by directly stimulating the human brainstem a current density of ∼ 2–9 mA/cm² is required (Cedzich et al., 1998); this is far above the current density we delivered.

In conclusion, anodal tDCS applied over the right motor cortical areas prolonged the Et for contralateral elbow flexors in a submaximal isometric task, thus showing that brain polarization can modulate human neuromuscular fatigue. Anodal tDCS could improve the Et through several different mechanisms, for example by increasing motor cortical excitability, acting on premotor areas, decreasing fatigue-related muscle pain, inducing behavioural disinhibition and increasing motivation or improving synergist muscle coupling. Our observations open the possibility of improving muscle performance and decreasing muscle fatigue in normal (i.e. sports medicine) and pathological (i.e. stroke and primary or secondary chronic fatigue syndrome, rehabilitation) conditions.

Acknowledgements

The authors wish to thank Mr Vito Antonio Di Pace for building the experimental apparatus, Lorenzo Rossi, PhD, for setting the experimental instrumentation, and Dr Simona Mrakic-Sposta, Dr Ettore Accolla and Dr Francesca Mameli for their helpful cooperation. S.M. is a PhD student supported by Fondazione IRCCS Ospedale Maggiore, Policlinico, Mangiagalli and Regina Elena di Milano (Italy) at the Dipartimento di Bioingegneria, Politecnico di Milano.

Abbreviations