Formation of cortical plasticity in older adults following tDCS and motor training

EXPERIMENTAL DESIGN

This study was a double-blinded, cross-over sham controlled trial, whereby all participants were exposed to three acute tDCS conditions combined with motor training, with a 1 week washout period between each condition. Active tDCS included both unilateral and bilateral electrode arrangements, whereas the third condition was a sham tDCS. The delivery of each condition was randomized across participants and followed identical testing protocols.

Participants were required to complete a familiarization session 1 week prior to the commencement of the study to reduce the effect of learning the motor performance task. All participants were tested for baseline measures of corticospinal excitability and intracortical inhibition, as well as motor performance. Following baseline testing, participants were exposed to a total of 15 min of tDCS. After the first 5 min of stimulation, participants performed 5 min of visuomotor tracking. Time-course measures of corticospinal excitability, SICI, and visuomotor tracking error were taken immediately after and 30 min following the stimulation (Figure ​Figure11).

Open in a separate window

FIGURE 1

Schematic representation of the experimental protocol with measures obtained at baseline, immediately after and 30 min after tDCS.

TRANSCRANIAL DIRECT CURRENT STIMULATION

Transcranial direct current stimulation was applied over the M1 for 15 min with a fade-in-fade out of 5 s, to avoid alternating currents causing transient neuronal firing. Two 25 cm² electrodes, soaked in a saline solution (0.9% NaCl), were placed over the cortical representation of the extensor carpi radialis longus (ECRL) muscle, as explored and determined with TMS, and secured with a rubber strap. In all conditions, the anode was placed over the non-dominant M1 in the area corresponding with the participants’ non-dominant “ECRL optimal site.” In the unilateral condition, the cathode was placed over the contralateral supraorbital area. During the bilateral condition, the cathode was placed over the dominant M1 corresponding to the “ECRL optimal site.” For the sham condition, participants randomly received unilateral or bilateral electrode arrangements (i.e., 50% of participants were allocated to each arrangement) following the electrode arrangements described above. For all conditions, stimulation intensity was delivered at 1 mA (current density 0.040 mA/cm²) through a DC-stimulator (NeuroConn DC stimulator, Ilmenau, Germany). In order to obtain the participants perception of discomfort across the three tDCS conditions, a visual analogue scale (VAS) was used during the first 3 min of stimulation.

ASSESSMENT OF MOTOR PERFORMANCE

Participants were seated in an office armchair, upright in a neutral position. The elbow was flexed at 90°; shoulder abducted at 45°, and the wrist rested on a chair in the neutral anatomical position. This position allowed free movement of the wrist. Participants were fitted with a sensor icon (i.e., actual limb) driven by a single axis goniometer (3DM-GX2, Williston, VT, USA). Participants were instructed to perform voluntary wrist extension and replicate the movement of a target limb displayed on a PC monitor in front of them, as accurately as possible. The position of the participants wrist joint was displayed as a mirrored anatomical representation of their upper limb, which was positioned parallel to the target limb on the screen. The moving target consisted of three, 30 s unique frames that moved automatically in a vertical manner (i.e., wrist extension and flexion) across the screen with varied frequencies. Each frame was repeated twice, with a total testing time of 3 min.

MOTOR TRAINING

After the first 5 min of tDCS, all participants performed 5 min of visuomotor tracking. Five, 30 s frames were used with alternating movement frequencies, where each frame was repeated twice.

RECORDING OF EMG ACTIVITY

Surface electromyography (sEMG) was recorded from the ECRL muscle in both limbs using bipolar Ag-AgCL electrodes. Two electrodes were placed 2 cm apart on the mid belly of the ECRL, with a ground strap placed around the wrist as a common reference for all electrodes. All cables were fastened with tape to prevent movement artifact. The skin was prepared (i.e., shaved and swabbed with alcohol) prior to electrode placement to ensure a clear signal was obtained. sEMG signals were amplified (×100–1000), bandpass filtered (high pass at 13 Hz, low pass at 1000 Hz), digitized online at 2 kHz for 500 ms, recorded and analyzed using PowerLab 4/35 (ADInstruments, Bella Vista, NSW, Australia).

TMS AND PERIPHERAL NERVE STIMULATION

Single and paired-pulse TMS were delivered over the cortical representation of the ECRL, using a figure-of-eight coil (external wing diameter 90 mm) attached via a BiStim unit, to two Magstim 200² stimulators (Magstim, Dyfed, UK). The coil was positioned over the M1 so that the current flowed in a posterior-anterior direction. Sites near the estimated center of the ECRL were explored to obtain the largest MEP amplitude (i.e., optimal site), and this area was marked by a small “X.” Participants maintained this mark throughout the intervention to ensure consistency and reliability of coil placement within and between sessions.

Measures of corticospinal excitability and intracortical inhibition included active motor threshold (AMT), MEP amplitudes at 130% AMT as well as SICI. All variables were collected at baseline, immediately following and 30 min following the cessation of tDCS. AMT was defined as the stimulator intensity at which at least five out of ten stimuli produced MEP amplitudes of greater than 200 μV. MEP amplitudes were evaluated by producing 10 stimuli at a test-intensity of 130% AMT. All MEPs were recorded during weak voluntary contraction whereby participants positioned their hand in line with their wrist (i.e., anatomically neutral). A constant level of contraction was maintained with reference to an oscilloscope (HAMEG, Mainhausen, Germany) that displayed the root mean square electromyography (rmsEMG) signal (equivalent to 5 ± 2% of maximal rmsEMG activity) in front of the participant. rmsEMG of the ECRL was obtained 100 ms prior to each TMS stimulus.

For the paired-pulse paradigm only, the test-intensity used was the stimulator output required to produce MEPs of ~1 mV. The test-intensity was adjusted if necessary, so that the test MEP amplitudes were always equivalent to ~1 mV (Cirillo et al., 2011). SICI was obtained by delivering a conditioning stimulus at 80% of AMT (subthreshold) followed by a test stimulus (~1 mV; suprathreshold), separated by a 3 ms inter-stimulus interval. Specifically, 10 test stimuli and 10 conditioned stimuli were delivered with the order of presentation randomized throughout the sessions. A rest period of 30 s was provided between stimuli sets to avoid muscular fatigue.

Direct muscle responses (M-waves), were obtained from the ECRL muscle by direct supramaximal electrical stimulation (pulse duration 1 ms) of the radial nerve under resting conditions. A high-voltage constant current stimulator (DS7, Digitimer®, Hertfordshire, UK) delivered each electrical pulse. Stimulation was delivered by positioning bipolar electrodes over the radial nerve on the distal, lateral shaft of the humerus. An increase in current strength was applied until there was no further increase in sEMG amplitude (M_MAX). To ensure maximal responses, the current intensity was increased an additional 20% and the average M_MAX obtained from five stimuli was delivered and recorded at 0.2 Hz.

All TMS and M-wave procedures were performed for both limbs at each time point and the order of limb testing was randomized across participants and conditions.

BASELINE CHARACTERISTICS

Table ​Table11 displays the mean ± SEM baseline values for measures of corticospinal excitability and intracortical inhibition in both hemispheres. No differences in rmsEMG, M_MAX, AMT, MEP amplitudes, and SICI ratios at baseline were observed across conditions, all p > 0.05. There were no hemispheric differences between the dominant and non-dominant M1 for baseline AMT t₃₂ = -1.88, p = 0.070, MEP amplitude, t₃₂ = -1.16, p = 0.255 and SICI ratio, t₃₂ = 1.54, p = 0.134.

rmsEMG

The average ± SEM rmsEMG (μV) prior to single and paired-pulse recordings was 0.053 ± 0.004 and 0.050 ± 0.003, respectively. There were no condition by time interactions for pre stimulus rmsEMG for single pulse, F_2.41,36.18 = 1.40, p = 0.261, paired-pulse, F_4,60 = 0.36, p = 0.835 signals. For the dominant hemisphere, the average ± SEM pre stimulus rmsEMG for single pulse MEPs was 0.055 ± 0.004 and 0.053 ± 0.004 for paired-pulse MEPs. There was no condition by time interaction for rmsEMG for single pulse, F_4,58 = 1.87, p = 0.127, or paired-pulse, F_4,60 = 1.74, p = 0.152 signals

M_MAX

There was no condition by time interaction for M_MAX in the non-dominant, F_2.63,39.46 = 0.77, p = 0.502, or dominant hemisphere, F_4,60 = 0.18, p = 0.947.

MOTOR PERFORMANCE

Figure ​Figure22 displays the average visuomotor tracking errors for each condition across time. There was a significant reduction in the proportion of error over time, F_1.38,40.13 = 19.01, p < 0.001 and a significant main effect for condition, F_2,19 = 4.04, p = 0.034, however the condition by time interaction was not significant, F_2.77,40.13 = 2.02, p = 0.131. Immediately following stimulation, only the unilateral and bilateral conditions improved motor performance by 13%, p = 0.006, and 21%, p < 0.001, respectively. At 30 min all conditions appeared to improve relative to baseline (sham 10%, p = 0.023; unilateral 12%, p = 0.012; bilateral 21%, p < 0.001). There were no differences between unilateral and bilateral stimulation for either time point, p > 0.05.

Open in a separate window

FIGURE 2

Average ± SEM tracking error (%) for all conditions across all time points. ^*p < 0.05 compared to baseline. No significant condition by time interaction p > 0.05.

MEPs

Figure ​Figure33 displays the average MEP amplitudes for each condition across time for the non-dominant hemisphere. There was a significant condition by time interaction F_4,60 = 4.29, p = 0.004. The change from baseline to immediately post stimulation for both the unilateral, 38%, p = 0.021, and bilateral, 53%, p < 0.001, conditions were significantly greater than the change for the sham condition, and this was sustained at 30 min (unilateral 49%, p = 0.014, and bilateral, 54%, p = 0.003). There were no significant differences in MEP amplitude between unilateral and bilateral conditions at any time point, p > 0.05. Figure ​Figure44 provides an illustration of the MEP sweeps recorded at 130% AMT across all conditions and time points for one participant.

Open in a separate window

FIGURE 3

Average ± SEM MEP amplitudes (% of M_MAX) at 130% AMT for all conditions across all time points. (A) non dominant M1, (B) dominant M1. ^*p < 0.05 compared to baseline. ^†p < 0.05 compared with sham.

Open in a separate window

FIGURE 4

MEP amplitude (130% AMT) sweeps recorded from one participant at baseline, immediately post and 30 min post stimulation for sham (A) unilateral (B) and bilateral (C) conditions.

For the dominant hemisphere there was a significant condition by time interaction, F_4,60 = 3.74, p = 0.009. There were no significant changes, p > 0.05, over time in either the sham or the unilateral tDCS condition but in the bilateral tDCS condition the decreases over time from baseline to the immediate and 30 min post time points were significant, 14%, p = 0.005 and 15%, p = 0.003 respectively.

MOTOR PERFORMANCE FOLLOWING tDCS

Our findings suggest that the tDCS electrode arrangement does not differentially modulate motor performance in older adults. Although motor improvements in older adults have been shown to be sustained following unilateral anodal tDCS (Hummel et al., 2010; Zimerman et al., 2012), a novel aspect to this study was that bilateral stimulation yielded similar improvements in an older population. Although previous findings have suggested a preferential increase in motor performance following bilateral tDCS, compared to unilateral and sham (Vines et al., 2008), this was not evident in the current study. The current findings suggest that the addition of motor training of the non-dominant limb may have augmented the excitatory response from the anode (in both conditions), and thus, modulated corticospinal excitability in a similar manner, contributing to the non-significant differences between unilateral and bilateral conditions.

In this study, the 13–21% improvement in motor performance following active tDCS is slightly larger than a previous investigation in older adults that observed a 2–10% improvement following an acute session of anodal tDCS applied in isolation (i.e., no motor training; Hummel et al., 2010). The greater performance gains in the current study could be explained by the fact that tDCS was applied not only in conjunction with, but prior to a visuomotor training task. It is established that the acute effects of tDCS induce spontaneous neuronal firing which primes the M1 and may consolidate the effects of motor training (Nitsche and Paulus, 2000, 2001; Vines et al., 2006). Although not quantified, the use of a visuomotor task may have activated other cortical and subcortical motor areas that are known contribute to motor performance and learning of a motor task (Bolam et al., 2000). In support of this, the magnitude of performance improvement observed in our study is consistent to the performance improvements induced by visuomotor training in a younger healthy population, without the addition of tDCS (Perez et al., 2006; Cirillo et al., 2011). Given the evidence supporting the reduced effectiveness of motor training alone in older compared with young adults (Rogasch et al., 2009; Zimerman et al., 2012), it appears the addition of tDCS has acted to improve the response to motor training in an older population. This finding suggests that the improved corticospinal activity following tDCS may be beneficial to facilitate improvement in motor performance. However, it should be noted, that interestingly we still observed an improvement in motor performance (10%) at 30 min following motor training alone. In light of this, it seems that the formation of motor performance improvement is facilitated following tDCS when combined with motor training, though, older adults still appear to benefit from motor training alone.

CORTICOSPINAL EXCITABILITY FOLLOWING tDCS

Motor improvements have been observed following unilateral anodal tDCS in older adults; however the mechanisms underpinning these after-effects have not been previously quantified. Further, no study has examined the effects of bilateral tDCS on modulating corticospinal excitability and motor performance in older adults. The current findings demonstrate no significant differences in the indices of M1 plasticity between unilateral and bilateral tDCS. These results are in agreement with Mordillo-Mateos et al. (2012) who also observed that corticospinal excitability in young adults was not differentially modulated by unilateral and bilateral tDCS. Given these findings, we speculate that the alternate tDCS electrode arrangements induce similar physiological mechanisms, underlying the improvements in motor performance.

Our findings show that in older adults, both unilateral and bilateral tDCS facilitated MEPs both immediately after and 30 min following stimulation. These findings are in agreement with studies performed in young adults whereby MEPs were increased following an acute bout of either unilateral or bilateral tDCS (Nitsche and Paulus, 2000, 2001; Lang et al., 2004; Williams et al., 2010; Mordillo-Mateos et al., 2012; Kidgell et al., 2013). In the dominant M1, MEP amplitudes were suppressed for the bilateral condition only, which supports previous literature in young adults (Williams et al., 2010). Reduced MEPs within the dominant M1 suggests that the cathode may have suppressed motor overflow of neuronal activity in the M1 ipsilateral to the trained limb (dominant M1), however this did not appear to differentially affect motor performance. In light of this finding, there were no differences in corticospinal excitability or motor performance of the non-dominant limb following either unilateral or bilateral tDCS suggesting the effect of the cathode over the dominant M1 may have been minimal. Although we observed a 15% reduction in corticospinal excitability of the dominant M1, previous data has suggested that individual variability in the response to tDCS, in particular cathodal stimulation, may be due to genetic factors such as the polymorphism of brain-derived neurotrophic factor (BDNF; Antal et al., 2010). Although not quantified in this study, the potential for these factors to contribute to the non-significant differences in between the responses to unilateral or bilateral tDCS cannot be overlooked. Therefore, it should be considered that there may be no optimal electrode arrangement for inducing plasticity and improving motor performance in this population.

The finding of tDCS induced corticospinal excitability outlasting the stimulation period, appears reflective of LTP-like mechanisms that have been observed following motor learning in rats (Sanes and Donoghue, 2000; Rioult-Pedotti et al., 2000), and has more recently have been suggested to occur in humans (Ziemann et al., 2004). Previous pharmaceutical investigations have applied dextromethorphane (an NMDA receptor antagonist) to alter the after-effects of tDCS (Liebetanz et al., 2002; Nitsche et al., 2003), therefore it can be proposed that the after-effects of tDCS are indicative of modifications in NMDA receptor dependent neurotransmission. The small and non-significant increase in corticospinal excitability following motor training with sham tDCS, supports the reduced ability for older adults to form use-dependant plasticity following motor training alone (Sawaki et al., 2003; Rogasch et al., 2009). The current results speculate that during natural aging, there may be a limited response to plasticity inducing protocols that reflect the involvement of LTP-like processes. In vivo studies have certainly shown that in an aging rat model, the interaction between dopamine, GABA and glutamate in the basal ganglia is decreased, which may be reflective of decreased activity in glutamate receptor binding sites (i.e., NMDA receptor; Segovia and Mora, 2005; Mora et al., 2008). Additionally, in aged human brain tissue, a reduction in both dopamine uptake and NMDA receptor activity have been observed (Villares and Stavale, 2001). Given that these neurotransmitters located in the basal ganglia are important for the acquisition and performance of motor patterns, degeneration of these structures may contribute to the delayed onset of motor performance improvement observed in this study. Collectively, the current evidence suggests the additive combination of tDCS and motor training may have improved sensitivity and unmasking of excitatory synapses at the post-synaptic membrane, improving synaptic efficacy, and neural transmission along the corticospinal pathway (Nielsen and Cohen, 2008).

INTRACORTICAL INHIBITION FOLLOWING tDCS

It is established that GABA dependent neurotransmission plays an important role in shaping excitatory output, and is partially modulated by NMDA receptor activity (Ziemann et al., 1998, 2001; Zoghi et al., 2003). Therefore, the balance of GABA mediated intracortical inhibition is vital for efficient and coordinated movement (Rothwell et al., 2009; Marneweck et al., 2011). Previously, the effect of tDCS on intrinsic inhibitory circuits has not been investigated in older adults. The current study demonstrated a release of SICI in the non-dominant M1, following both unilateral and bilateral tDCS relative to sham, but importantly, there were no differences between the two active tDCS conditions.

The current finding that tDCS reduced SICI by 21–36% is comparable to studies in young adults and post stroke patients, which demonstrates that improvements in corticospinal excitability may be modulated by reduced intracortical inhibition (Hummel et al., 2005; Nitsche et al., 2005; Edwards et al., 2009; Batsikadze et al., 2013). Contrary to our findings, Williams et al. (2010) combined bilateral tDCS with motor training in healthy young adults and found no effect on SICI. Although comparisons to this study should be viewed with caution (due to different tDCS parameters used), the age of the cohort used in the current study may have contributed to the tDCS induced changes in SICI. Certainly, there is evidence for age-related deficits in SICI circuitry (Sale and Semmler, 2005; Fujiyama et al., 2009), and therefore it is possible that older adults may respond more favorably to tDCS.

Based upon the current findings it can be speculated that the reduction in GABAergic inhibition has improved the synaptic efficacy between intracortical and corticospinal neurons (Nitsche et al., 2003). In support of this, previous data in the rat M1 suggests that synaptic plasticity is enhanced by a release of intracortical inhibition (Hess et al., 1996). Interestingly, there was a non-significant change in SICI in the dominant M1, possibly contributing to the non-significant differences between the release of SICI following the two active tDCS conditions. Although, a recent study in an aging population observed differences in spatial activation between bilateral and unilateral tDCS (Lindenberg et al., 2013), it is conceivable that contribution from other motor control pathways such as interhemispheric networks, basal ganglia circuits and the posterior cingulate cortex involving GABAergic synapses may contribute to corticospinal excitability and motor performance improvements following tDCS (Bolam et al., 2000; Baudewig et al., 2001; Williams et al., 2010; Lindenberg et al., 2013). Irrespective of this, our results support the contribution of released GABA-related inhibitory activity in M1, on the overall net excitatory output and improved motor performance of the non-dominant limb in older adults.

Dawson J. Kidgell is supported by an Alfred Deakin Postdoctoral Award.