Simultaneous application of slow-oscillation transcranial direct current stimulation and theta burst stimulation prolongs continuous theta burst stimulation-induced suppression of corticomotor excitability in humans
Sebastian H. Doeltgen,
The Robinson Institute, School of Paediatrics and Reproductive Health, University of Adelaide, Adelaide, Australia
The objective of this study was to assess whether the simultaneous application of slow-oscillation transcranial direct current stimulation enhances the neuroplastic response to transcranial magnetic theta burst stimulation. Motor evoked potential amplitude was assessed at baseline and at regular intervals up to 60 min following continuous theta burst stimulation, slow-oscillation transcranial direct current stimulation, and the simultaneous application of these paradigms. In addition, the electroencephalographic power spectra of slow and fast delta, and theta frequency bands recorded over the motor cortex were analyzed prior to and up to 5 min following each intervention. There was longer-lasting motor evoked potential suppression following the simultaneous application of continuous theta burst stimulation and slow-oscillation transcranial direct current stimulation compared with when continuous theta burst stimulation was applied alone. Slow-oscillation transcranial direct current stimulation applied alone did not modulate the motor evoked potential amplitude. No significant changes in spectral power were observed following slow-oscillation transcranial direct current stimulation. Simultaneous application of continuous theta burst stimulation and slow-oscillation transcranial direct current stimulation may provide an approach to prolong the induction of neuroplastic changes in motor cortical circuits by repetitive transcranial magnetic brain stimulation.
Non-invasive brain stimulation techniques, such as repetitive transcranial magnetic stimulation (rTMS), can be used to modulate motor cortical excitability. The resulting plastic changes are in many ways similar to those seen in the cortex during motor learning, and there is evidence that the two share similar neural mechanisms. For example, both the changes in motor cortical excitability induced by many forms of non-invasive brain stimulation (Teo et al., 2007; Huang et al., 2007; Stefan et al., 2002), and those seen following motor training (Hadj Tahar et al., 2004; Bütefisch et al., 2000) are N-methyl-d-aspartate-receptor dependent, providing evidence for the involvement of long-term potentiation (LTP)-like and long-term depression (LTD)-like mechanisms. Also, the changes induced by rTMS techniques can interact with those seen following motor training, providing evidence for the involvement of similar mechanisms in overlapping networks (Ziemann et al., 2004; Stefan et al., 2006). Given that experimentally induced motor plasticity has subtle effects on motor function in normal subjects (Huang et al., 2005) and patients with neurological impairment (Conforto et al., 2002; Uy et al., 2003; Talelli et al., 2007; Stagg et al., 2012), non-invasive brain stimulation may provide novel opportunities for neurorehabilitation. However, the therapeutic potential of these interventions is currently limited by the short duration of, and large variability in, subject responses (Sale et al., 2007; Ridding & Ziemann, 2010). Therefore, it is important to explore approaches to enhance the neuroplastic response to such experimental brain stimulation techniques.
Slow-frequency oscillatory activities in the brain can exert a significant influence on hippocampal memory encoding and consolidation. For example, slow-wave activity during sleep is thought to play a critical role in memory consolidation and, indeed, reduction in the amount of slow-wave activity during sleep leads to memory impairments (Lu & Goeder, 2011). However, slow-wave sleep not only interacts with hippocampus-dependent memory, but has also been shown to facilitate the consolidation of motor memory following motor-skill training (Nishida & Walker, 2007). In addition, motor cortical plasticity is associated with changes in slow-wave activity during sleep. For example, skilled motor training in rats increased slow-wave activity during subsequent non-rapid-eye-movement sleep (Hanlon et al., 2009). Likewise, LTP-like plasticity induced in humans by 5 Hz rTMS is associated with increased slow-wave activity (Huber et al., 2007). In contrast, LTD-like plasticity seen with arm immobilization is associated with reduced slow-wave activity during subsequent sleep (Huber et al., 2006). These studies underscore the association between LTP/LTD-like phenomena, memory consolidation and slow-oscillatory activity not only in relation to hippocampus-dependent memory, but also memory function within the sensorimotor system.
Oscillatory activities within the human cortex can be modulated by using non-invasive brain stimulation techniques. For example, slow-wave activity in the cortex can be increased by applying 0.75 Hz transcranial oscillating electrical stimulation during slow-wave sleep, which, intriguingly, was associated with improved declarative memory (Marshall et al., 2006). Although it is generally thought that memory consolidation occurs most effectively off-line (i.e. during sleep), boosting slow-wave activity during wakefulness might also improve memory encoding. Kirov et al. (2009) demonstrated that slow-oscillating transcranial direct current stimulation (so-tDCS) at 0.75 Hz applied over fronto-lateral locations during the performance of a declarative memory task improves memory encoding. Given the demonstrated interaction between slow-wave activity and memory encoding in cortical motor networks, these findings suggest that it might be possible to enhance the neuroplastic response to rTMS by applying concurrent slow-oscillating electrical stimulation. Therefore, we hypothesized that so-tDCS simultaneously applied with an rTMS paradigm [continuous theta burst stimulation (cTBS)] that is known to induce LTD-like effects within the motor cortex results in an enhanced neuroplastic response when compared with that seen with cTBS applied alone. We further hypothesized that the application of so-tDCS results in an increase in electroencephalography (EEG) delta and theta power.
Materials and methods
Seventeen subjects (mean age 23.1 ± 5.1 years, 10 females) attended for three experimental sessions (cTBS plus sham so-tDCS, so-tDCS plus sham cTBS and simultaneous cTBS and so-tDCS), which were performed in pseudo-randomized order with a period of at least 48 h between sessions. All subjects were unaware of the experimental hypothesis. Subjects provided written, informed consent prior to inclusion in this study. No subjects met the exclusion criteria for transcranial magnetic stimulation, as assessed by a transcranial magnetic stimulation safety screen (Rossi et al., 2009). This study was approved by the University of Adelaide Human Research Ethics Committee and was conducted in accordance with the 1964 Declaration of Helsinki.
Subjects were seated comfortably in an armchair with their right arm and hand resting on the armrest of the chair, in a relaxed, pronated position supported by pillows. Following preparation of the skin using an alcohol swab, surface electromyography electrodes (silver/silver chloride monitoring electrodes, Ambu blue sensor, M-00-S/50) were attached to the skin overlying the first dorsal interosseous muscle in a belly-tendon montage. All electromyographic signals were sampled at 5 kHz (Cambridge Electronic Design, power 1401 interface), amplified (×1000) and filtered (20–1000 Hz). Subjects maintained relaxation of their hand, arm and shoulders throughout each experimental session and visual assessment of the ongoing electromyographic signal was used by the investigators to monitor muscle relaxation.
Transcranial magnetic stimulation
Changes in corticospinal excitability were assessed by single-pulse transcranial magnetic stimulation employing a mono-phasic stimulus applied through a figure-of-eight stimulation coil (90 mm external wing diameter) connected to a Magstim 2002 stimulator (Magstim Co., Whitland, UK). The coil handle was oriented at an angle of approximately 45° to the midline, in a tangential plane to the skull, with the handle pointing backwards, inducing a posterior–anterior current flow in the brain. The scalp was systematically explored for the coil position over which motor evoked potentials (MEPs) in the first dorsal interosseous muscle could be evoked consistently (hotspot), which was then marked with a pen.
At the beginning of each session, the resting motor threshold was determined in the relaxed first dorsal interosseous muscle. The resting motor threshold was defined as the lowest stimulus intensity (expressed as a percentage of maximal stimulator output) at which five out of ten magnetic stimuli applied at the hotspot evoked MEPs of at least 50 μV in amplitude. Sets of 15 MEPs were recorded with an interval of 7 s (10% variance) between successive stimuli. The stimulus intensity used for MEP assessment at all time-points was set to a level that evoked MEPs of ∼ 1 mV in amplitude at baseline.
The cTBS was delivered using a Magstim Super Rapid2 stimulator connected to an air-cooled figure-of-eight stimulating coil (Magstim Co.). Stimulation consisted of short bursts of three stimuli at 50 Hz applied every 200 ms for 40 s (600 stimuli in total) (Huang et al., 2005). The cTBS intensity was set to 80% of the active motor threshold. The active motor threshold was determined while the subject maintained an approximately 10% maximal voluntary contraction of the first dorsal interosseous muscle and was defined as the lowest level of maximal stimulator output required to evoke MEPs of at least 200 μV in amplitude in five out of ten successive trials. Sham cTBS was delivered using a sham figure-of-eight stimulation coil connected to the same Magstim Super Rapid2 stimulator. Both real and sham cTBS were applied over the hotspot.
Slow-oscillating transcranial direct current stimulation
The so-tDCS (0.8 Hz) was applied through conductive rubber electrodes that were contained within saline-soaked sponges (5 × 5 cm) and connected to a Magstim transcranial direct current stimulator plus (Eldith DC-stimulator, Neuroconn, Germany). The anode was positioned over the hotspot, and the cathode positioned on the forehead above the right orbit. The impedance was kept below 55 kΩ and so-tDCS was applied simultaneously with either real or sham cTBS. The sinusoidal current intensity was set to 1 mA (Nitsche & Paulus, 2000) and stimulation was applied for 40 s, consistent with the duration of cTBS. A current offset of 500 μA was applied so that the stimulation polarity did not change. These parameters did not exceed previously recommended safety limits (Agnew & McCreery, 1987). Some subjects reported a mild tingling sensation under the so-tDCS electrodes during real so-tDCS. During the real cTBS only protocol, so-tDCS electrodes were positioned as in the protocols applying real so-tDCS; however, the electrical stimulator remained switched off.
Silver/silver chloride disk electrodes (10 mm diameter) were placed according to the international 10–20 system at C3 and Fz in a bipolar montage. EEG recordings were made in blocks of 1 min before, immediately following and 5 min after each stimulation protocol. Participants were instructed to avoid movement and to keep their eyes open, looking straight ahead, during EEG recording. EEG data were sampled at a rate of 2048 Hz and the electrode impedance was kept below 5kΩ. The EEG was amplified (10 000), filtered (between 0.5 and 100 Hz) and recorded using the Signal software package (Cambridge Electronic Design, power 1401 interface), and stored on a personal computer for offline analysis. To facilitate the removal of eye blink and other artifacts, the EEG data were split into 2 s epochs and any artifact-contaminated epochs were removed from further analysis. Following this, for each EEG epoch the direct current component was removed and the data were zero padded to 20 480 samples for spectrum interpolation. Fourier transformations with a rectangular window were then performed (Rao et al., 2010). Subsequently, the mean power of the slow delta frequency band (summing power between 0.5 and 2 Hz), fast delta frequency band (summing power between 2 and 4 Hz) and theta frequency band (summing power between 4 and 8 Hz) was determined for each 2 s epoch. Finally, the mean spectral power in each frequency band for each 30 s epoch was calculated by averaging the power across 15 consecutive 2 s epochs. Thus, a total of six 30 s blocks of EEG were analysed, two before the intervention, two shortly after and two at 5 min post-intervention. The power in each frequency band following intervention was compared with pre-intervention baselines. In addition, for delta frequency bands, the ratio between the slow delta and overall delta power was determined for each epoch and compared with pre-intervention baselines. The slow and fast delta power were investigated as previous studies have shown changes in slow-oscillation frequency bands centered around the stimulation frequency (Kirov et al., 2009; Marshall et al., 2006). The power in the theta frequency band was examined as previous studies have demonstrated widespread changes in the theta power following so-tDCS applied to the frontalateral cortex, which was associated with enhanced encoding of hippocampus-dependent declarative memory (Kirov et al., 2009). Theta activity has also been linked to memory function in cortical regions, including the frontal cortex (Jensen & Tesche, 2002) and cortical areas associated with sensorimotor integration and spatial learning (Caplan et al., 2003; Cruikshank et al., 2012). Therefore, we sought to examine whether there were modulations in theta activity associated with so-tDCS and whether such modulations influenced the neuroplastic response to cTBS.
Experiment 1 – motor evoked potential amplitude
All 17 subjects participated in this experiment, which evaluated the effects of cTBS plus sham so-tDCS, so-tDCS plus sham cTBS and simultaneous cTBS and so-tDCS. The MEP amplitude was assessed before, and at 2, 6, 10, 15, 30, 45 and 60 min following each intervention. EEG was recorded before the intervention, and shortly after (30–60 and 61–90 s) and at 5 min following each intervention (300–330 and 331–360 s) (Fig. 1).
Experiment 2 – immediate changes in electroencephalography power spectra
Due to technical limitations in Experiment 1, we were unable to examine changes in EEG power in the first 30 s following the interventions. Thus, very short-lasting changes in the post-intervention EEG may have been missed. Therefore, we conducted a second experiment to specifically examine whether so-tDCS modified the power in each frequency band in the period immediately following stimulation. Eight of the subjects tested in Experiment 1 (mean age 22.4 ± 4.1 years, four females) attended for an additional session in which so-tDCS was applied as described in Experiment 1. The EEG power was calculated for each frequency band and subjected to the same data analysis procedures as described above, providing the mean power for six 30 s epochs (baseline 1, baseline 2, 0–30, 31–60, 300–330 and 331–360 s). As in Experiment 1, the power in each frequency band following intervention was compared with pre-intervention baselines. In addition, for the delta frequency bands, the ratio between slow delta and overall delta was determined for each epoch and compared with the pre-intervention baseline. MEPs were recorded at baseline and 6 min following so-tDCS (Fig. 1).
The MEP amplitudes were averaged across the 15 trials recorded at each time-point. The MEP amplitudes were examined using a two-way repeated-measures anova using the variables Stimulation Type (cTBS alone, so-tDCS alone, simultaneous cTBS and so-tDCS) and Time (baseline, 2, 6, 10, 15, 30, 45 and 60 min). In addition, a separate two-way repeated-measures anova of the MEP data was performed comparing only the cTBS alone and simultaneous cTBS and so-tDCS conditions using the variables Type (cTBS alone, simultaneous cTBS and so-tDCS) and Time. Subject to significant main effects or interactions, paired-samples t-tests were performed to compare the post-intervention outcome measures with baseline (corrected for multiple comparisons using Bonferroni adjustments).
The EEG data were contaminated by frequent eye blink artifacts in two subjects and their data were excluded from the analysis. For statistical analyses of the delta and theta power of the remaining 15 subjects, the two 30 s epochs recorded at baseline were averaged to establish the baseline power values. Each of the post-intervention 30 s epochs was then compared with this baseline value. First, a three-way repeated-measures anova using the independent variables of Stimulation Type (so-tDCS, cTBS, simultaneous so-tDCS and cTBS), Frequency Band (slow delta, fast delta and theta) and Time (baseline, 30–60, 61–90, 300–330 and 331–360 s post-intervention) was conducted. Subject to significant main effects, separate two-way repeated-measures anova s were conducted to examine the spectral power in each frequency band using the independent variables of Stimulation Type and Time. Subsequently, paired-samples t-tests comparing post-intervention epochs with baseline were performed subject to significant main effects or interactions.
The MEP amplitudes were calculated by averaging across the 15 trials recorded at each time-point (i.e. before and at 6 min after so-tDCS) and compared using a paired-samples t-test. Changes in the spectral power in each frequency band (slow delta, fast delta and theta) were examined by conducting a two-way repeated-measures anova first, using the variables Frequency Band (slow delta, fast delta and theta) and Time (baseline, 0–30, 31–60, 300–330 and 331–360 s post-intervention). Subject to significant main effects or interactions, three separate one-way repeated-measures anova s using the variable Time were then performed for each frequency band. Subject to a significant main effect, post-intervention epochs were compared with baseline using paired-samples t-tests.
Resting motor threshold and stimulus intensities
Baseline resting motor thresholds were not significantly different between experimental sessions [overall mean 46% (8.2); F2,36=1.3, P =0.285]. Likewise, the baseline active motor threshold did not differ across experimental sessions [overall mean 52.8% (9.8); F2,36=1.26, P =0.297].
Motor evoked potential amplitude
There were significant main effects for Stimulation Type (F2,32=8.28, P =0.001) and Time (F7,112=5.79, P <0.001), as well as a significant interaction between these variables (F14,224=2.25, P =0.007) (Fig. 2). Direct comparison of the cTBS alone and simultaneous cTBS and so-tDCS conditions revealed a significant main effect of Type (F1,16=4.7, P =0.045) and Time (F7,112=9.43, P < 0.001), but no significant interaction between these variables (F7,112=1.75, P = 0.104). Paired-samples t-tests revealed a significant reduction of MEP amplitude following cTBS alone (sham so-tDCS) at 30 min (t16 = 3.1, P =0.007) and 45 min (t16 = 6.1, P <0.001), and following simultaneous cTBS and so-tDCS at 0 min (t16 = 3.23, P =0.005), 5 min (t16 = 3.28, P =0.005), 10 min (t16 = 3.75, P =0.002), 30 min (t16 = 4.9, P <0.001), 45 min (t16 = 6.88, P <0.001) and 60 min (t16 = 4.96, P <0.001). The MEP amplitudes were significantly smaller following simultaneous cTBS and so-tDCS compared with cTBS alone (sham so-tDCS) at 60 min (t16 = 3.15, P =0.006) (Fig. 2).
Electroencephalography delta power
Three-way repeated-measures anova revealed significant main effects of Stimulation Type (F2,28 = 4.5, P =0.019) and Frequency Band (F2,28 = 23.5, P <0.001).
Subsequent two-way repeated-measures anova revealed no significant main effects on the slow delta power (Stimulation Type –F2,28 = 2.8, P =0.078; Time –F4,56 = 2.11, P =0.091), or interaction between these variables (F8,112 = 0.947, P =0.414). For the fast delta power, there was a significant main effect of Stimulation Type (F2,28 = 3.5, P =0.042), but no effect of Time (F4,56 = 1.5, P =0.237) or interaction between these variables (F8,112 = 0.36, P =0.805). Subsequent separate one-way repeated-measures anova s for each stimulation type revealed no significant main effects of Time on the fast delta power (P >0.184) (Fig. 3).
There were no significant main effects on the ratio between the slow delta power and overall delta power (Stimulation Type –F2,28 = 0.16, P =0.853; Time –F4,56 = 1.14, P =0.346), or interaction between these variables (F8,112 = 1.099, P =0.369).
Electroencephalography theta power
There was a significant effect of Stimulation Type (F2,28 = 3.45, P =0.046), but no effect of Time (F4,56 = 1.5, P =0.237) or interaction between these variables (F8,112 = 0.88, P =0.449). Subsequent separate one-way repeated-measures anova s for each stimulation type revealed no significant main effects of time (P >0.314) (Fig. 3).
Motor evoked potential amplitude
In line with the findings of Experiment 1, so-tDCS did not significantly affect the MEP amplitude at 6 min following the intervention (t8 = 0.45, P =0.66).
Electroencephalography delta power
Two-way repeated-measures anova revealed a significant main effect of Frequency Band (F2,16 = 8.16, P =0.02) but no effect of Time (F4,32 = 2.8, P =0.11) or interaction between these variables (F8,64 = 2.4, P =0.12). Subsequent, one-way repeated-measures anova s for each frequency band revealed no significant main effects of Time on either the slow delta power (F4,32 = 2.8, P =0.104) or fast delta power (F4,32 = 2.4, P =0.13) (Fig. 4).
Electroencephalography theta power
One-way repeated-measures anova also revealed no significant main effect of Time on the theta power (F4,32 = 0.93, P =0.402) (Fig. 4).
Previous research has demonstrated that experimentally enhancing cortical slow-oscillatory activity with so-tDCS can facilitate the encoding and consolidation of memory (e.g. Lu & Goeder, 2011; Marshall et al., 2006). Given previously demonstrated interactions between slow-wave activity and memory encoding and consolidation in cortical motor networks (Nishida & Walker, 2007; Huber et al., 2007; Hanlon et al., 2009; Kirov et al., 2009), we investigated whether so-tDCS applied over the primary motor cortex can increase the neuroplastic response to a simultaneously applied rTMS protocol. The novel finding of the current study is that there was a trend toward more robust, and significantly longer lasting, reduction of MEP amplitudes following the simultaneous application of cTBS and so-tDCS, than when cTBS was applied alone. When so-tDCS was applied alone, there was no significant effect on MEP amplitudes. There were no significant changes in the EEG delta and theta power in the immediate post-stimulation period.
Effects of continuous theta burst stimulation alone and slow-oscillating transcranial direct current stimulation alone on motor evoked potential amplitude
In line with previous studies, cTBS applied alone suppressed MEP amplitudes for up to 45 min (Huang et al., 2005; McAllister et al., 2011; Doeltgen & Ridding, 2011). It is of note that MEP suppression only reached the statistical significance level at the 30 min time-point following the intervention, whereas previous studies have reported an earlier onset of MEP suppression post-cTBS (Huang et al., 2005). The reasons for this discrepancy are unclear although it is known that there is considerable variability in subject responses to such paradigms (see Ridding & Ziemann, 2010) and therefore subject characteristics may be responsible.
When so-tDCS was applied alone there were no significant changes in MEP amplitude. Previous studies have demonstrated that anodal so-tDCS applied for longer periods (up to 20 min) and at higher intensities (1.5 mA) can induce changes in MEP amplitude that outlast the stimulation period (Groppa et al., 2010; Bergmann et al., 2009). Therefore, it is likely that the lack of effect on MEPs when so-tDCS was applied in isolation in the present study is due to the very short duration (40 s) and lower stimulation intensity (1 mA) employed. In support of this, Nitsche & Paulus (2000) demonstrated that a stimulation duration of at least 3 min is required for transcranial direct current stimulation applied at 1mA to induce lasting effects on MEP amplitude.
Effects on electroencephalography delta and theta power
We hypothesized that the power in the delta and theta frequency bands would increase during the application of so-tDCS; however, due to technical limitations imposed by a large so-tDCS artifact, we were unable to assess the EEG power during periods of stimulation. However, there is some evidence to suggest that the application of so-tDCS can result in changes in slow-oscillation and theta power that outlast the stimulation period (Marshall et al., 2006; Kirov et al., 2009). Therefore, to provide evidence that so-tDCS modulated cortical slow-oscillatory activity in the present study, we examined whether there were changes in the delta and theta power that lasted beyond the application of so-tDCS in Experiment 1. Analysis of the EEG data provided no evidence that so-tDCS altered the slow or fast delta or theta power in the period examined following the intervention. However, technical limitations in Experiment 1 precluded the examination of EEG data in the 30 s period immediately following so-tDCS application; hence, short-lasting changes in the delta (and theta) EEG power may have been missed. Therefore, we further investigated the effects of so-tDCS on EEG power immediately following stimulation in Experiment 2. Again, this experiment provided no evidence of significant changes in the delta and theta power in the period immediately following so-tDCS (Fig. 4). This finding was somewhat unexpected given that other groups have reported so-tDCS effects on the EEG that outlast the stimulation period (Kirov et al., 2009). The reasons that we failed to see an effect on EEG power following the stimulation period are not clear. However, it should be noted that in the studies by Kirov et al. (2009) and Marshall et al. (2006), the change in the power of EEG frequency bands was seen in stimulation-free periods between longer and repeated 5 min stimulation epochs. Therefore, longer periods of stimulation may be necessary in order to see lasting effects on oscillatory EEG activities. Interestingly, there appeared to be a similar, yet small trend toward increased power in the theta frequency band following the application of so-tDCS alone (Fig. 3C), as reported previously by Kirov et al. (2009).
The fact that we were unable to demonstrate an effect of so-tDCS on delta EEG activities does not necessarily rule out effects on the EEG during so-tDCS stimulation. Indeed, Marshall et al. (2006) demonstrated the entrainment of slow-oscillatory activity with so-tDCS when applied during early nocturnal non-rapid-eye-movement-sleep. However, in the absence of online recordings of cortical oscillations during stimulation, it is currently not possible to conclusively prove that so-tDCS modulated the slow-oscillatory cortical activity during stimulation.
Possible mechanisms by which slow-oscillating transcranial direct current stimulation might modify the response to continuous theta burst stimulation
The mechanisms responsible for the enhanced cTBS response when applied with so-tDCS are not clear. However, several possibilities exist, which we will discuss in turn. Firstly, cellular gating mechanisms may contribute to our findings, in that the application of so-tDCS may have modulated the level of excitability of cortical motor neurons simultaneously targeted by cTBS. Secondly, synchronization of the excitability state of the targeted neuron population may contribute to the reported findings in that it may align the stimulated neuron population’s response to cTBS. Thirdly, longer lasting MEP suppression may be related to the engagement of different LTD-like molecular mechanisms during the simultaneous application of so-tDCS and cTBS.
Slow-oscillating transcranial direct current stimulation-induced gating of the continuous theta burst stimulation response
Synaptic gating describes a neuronal mechanism by which neuroplastic changes are dependent on a variable threshold of post-synaptic depolarization (for review, see Ziemann & Siebner, 2008). Specifically, the induction of LTD-like plasticity is facilitated when post-synaptic depolarization is above a certain LTD threshold, but below a (higher) threshold above which LTP induction is facilitated (Artola et al., 1990). Synaptic gating mechanisms are thought to be involved in the modulation of motor cortical excitability that is observed in response to motor learning tasks performed during the simultaneous application of non-invasive brain stimulation techniques. For example, performance on a sequential finger movement task improved more when motor training occurred during the simultaneous application of 10 Hz rTMS than when compared with a sham condition (Kim et al., 2004). Likewise, anodal transcranial direct current stimulation applied over the primary motor cortex during a serial reaction time task resulted in improved performance (Nitsche et al., 2003).
Similarly, the simultaneous application of so-tDCS and cTBS in the present study may have resulted in ‘gating’ of the cTBS-induced neuroplastic response. cTBS has previously been shown to induce LTD-like effects in cortical motor networks (Huang et al., 2005, 2007). The slowly oscillating depolarization of neuronal networks by means of anodal so-tDCS may have facilitated the cTBS-induced response in these networks by exceeding the threshold critical for LTD-like plasticity. Gating mechanisms may also explain the apparently contradictory recent report that the simultaneous application of non-oscillating anodal transcranial direct current stimulation reversed the inhibitory effects of cTBS toward facilitation (Hasan et al., 2011). These authors suggested that hyperpolarization of the post-synaptic (dendritic) membrane produced by anodal transcranial direct current stimulation might reduce Ca2 + entry during the prolonged cTBS train and favor the induction of a facilitatory effect, similar to that reported with shorter cTBS trains (Gentner et al., 2008). Given that Hasan et al. (2011) used the same peak current as in the present study, but in the form of a non-oscillating, direct-current waveform, twice as much total current was delivered in their study. Therefore, it is possible that their transcranial direct current stimulation protocol may have exceeded the LTP induction threshold in the motor networks concurrently targeted by cTBS, resulting in the induction of LTP-like changes.
Synchronization of the neuronal population
A second potential mechanism that might explain the current findings is related to the rhythmic synchronization of membrane excitability. Brain oscillations have the potential to ‘bind’ populations of neurons by synchronizing their excitability and activity (Buzsaki & Draguhn, 2004). For example, oscillations can synchronize the activity in widely dispersed but interconnected neuronal populations and therefore provide a potential mechanism for information transfer between remote, but functionally connected, brain regions (Buzsaki & Draguhn, 2004; Bazhenov et al., 2008). Such temporal coordination of cell assemblies is thought to be important for memory encoding (Buzsaki & Draguhn, 2004) and it has been suggested that, by synchronizing convergent input, oscillations may provide temporal ‘windows’ for synaptic integration (Jutras & Buffalo, 2010). As such, neuronal oscillations have been implicated in the modulation of spike-timing-dependent plasticity (Jutras & Buffalo, 2010). Therefore, we speculate that, if so-tDCS were capable of temporarily synchronizing the excitability state of the neuronal populations targeted by cTBS, then this may be a potential mechanism by which cTBS-induced changes in motor cortical excitability can be enhanced. The lack of significant changes in oscillatory activity that outlasted the stimulation period in the present study does not provide any support for this hypothesis. However, it has previously been shown that so-tDCS can, under certain circumstances, entrain slow-oscillatory brain activity to the oscillatory current of so-tDCS (Marshall et al., 2006). In addition, differences in motor cortical excitability have been demonstrated between the up and down states of slow oscillations during sleep (Bergmann et al., 2012). Synchronization of the excitability states of cortical neurons may facilitate their overall response to simultaneously applied cTBS by increasing the proportion of cells that respond in a similar fashion. In contrast, these neurons may be in more diverse excitability states in the cTBS alone condition and therefore produce an overall smaller effect. In this scenario, so-tDCS would act as a catalyst for cTBS-induced inhibition by synchronizing the responsiveness of the targeted neuronal populations during stimulation. In the absence of further experimental evidence of the entrainment of cortical oscillatory activity during so-tDCS, this hypothesis remains speculative.
Nature of the effects induced by combined continuous theta burst stimulation and slow-oscillating transcranial direct current stimulation
It is interesting to note that the response to cTBS when applied with concurrent so-tDCS displayed a trend toward greater MEP suppression and was longer lasting (at least 60 min) than that seen with cTBS alone. This raises the possibility that different mechanisms and processes may be involved in the response to combined cTBS and so-tDCS. Most studies have reported cTBS effects that last for approximately 45 min and this is consistent with the data from the present study. It is likely that this MEP suppression reflects early-phase LTD-like mechanisms, as blocking of N-methyl-d-aspartate receptors by the N-methyl-d-aspartate receptor antagonist memantine abolishes the effect of cTBS (Huang et al., 2007). However, given that MEPs were still significantly suppressed at 60 min following simultaneous so-tDCS and cTBS, it is possible that this combined stimulation induced a longer lasting form of LTD. Whereas early-phase LTD is characterized by functional changes such as the internalization of [2-amino-3-(5-methyl-3-oxo-1,2- oxazol-4-yl) propanoic acid] receptors in the post-synaptic cell (Snyder et al., 2001; Brown et al., 2005), late-phase LTD is additionally dependent on changes in gene transcription and protein synthesis, which ultimately lead to more persistent morphological changes (Zhou et al., 2003; Becker et al., 2008). It is possible that the application of a slowly oscillating current and concurrent modulation of N-methyl-d-aspartate receptor activation by cTBS may have facilitated the molecular mechanisms responsible for late-phase LTD, expressed as longer lasting inhibition of MEP amplitude. This hypothesis could potentially be tested in the future.
In summary, we have demonstrated that the simultaneous application of so-tDCS and cTBS prolongs the cTBS-induced inhibition of motor cortical excitability. The enhanced encoding of cTBS-induced plasticity, facilitated by synaptic gating mechanisms or differences in molecular mechanisms underlying cTBS-induced LTD-like effects, may contribute to the present findings. The simultaneous application of so-tDCS and cTBS may prove to be a feasible approach to enhance the functional effects reported previously for cTBS. Further research is warranted to evaluate this approach in neurologically healthy and impaired elderly cohorts and to establish behavioral correlates.
S.H.D. is a Postdoctoral Biomedical Research Fellow of the National Health and Medical Research Council (NHMRC) of Australia. S.M.M. is supported by an Australian Postgraduate Award. M.C.R. is a Senior Research Fellow of the NHMRC of Australia. This work was supported by a grant from the NHMRC of Australia (Grant ID 565301). The authors gratefully acknowledge the support of Ruiting Yang in the analysis of EEG data. The authors declare that they have no conflict of interest.
continuous theta burst stimulation
motor evoked potential
repetitive transcranial magnetic stimulation
slow-oscillating transcranial direct current stimulation