|Home | About | Journals | Submit | Contact Us | Français|
1. Conception and design of the experiments
2. Collection, analysis and interpretation of data
3. Drafting the article or revising it critically for important intellectual content
A number of experiments in animals have shown that successful induction of plasticity can be abolished if an individually ineffective intervention is given shortly afterwards. Such effects are termed depotentiation/de-depression. These effects contrast with metaplasticity/homeostatic plasticity in which pretreatment of the system with one paradigm modulates the response to a second plasticity-inducing paradigm. Homeostatic plasticity maintains the balance of plasticity in the nervous system at a stable level whereas depotentiation/de-depression abolishes synaptic plasticity that has just occurred in order to prevent ongoing learning. In the present study, we developed novel protocols to explore the reversal of LTP- and LTD-like effects in healthy conscious humans based on the recently developed theta burst form of repetitive transcranial magnetic stimulation (TBS). The potentiation effect induced by intermittent TBS (iTBS) was completely erased by a short form of continuous TBS (cTBS150) given 1 min after iTBS, whereas the depressive effect of continuous TBS (cTBS) was successfully abolished by a short form of iTBS (iTBS150). The reversal was specific to the nature of the second protocol and was time dependent since it was less effective when the intervention was given 10 min after induction of plasticity. All these features are compatible with those of depotentiation and de-depression demonstrated in animal studies. The development of the present protocols would be helpful to study the physiology of the reversal of plasticity and learning and to probe the abnormal depotentiation/de-depression shown in animal models of neurological diseases (e.g. Parkinson’s disease with dyskinesia, dystonia and Huntingon’s disease).
The efficiency of synaptic transmission can be either potentiated or depressed by natural patterns of activity or by experimental intervention. Most investigations into these phenomena of synaptic long term potentiation/depression (LTP/LTD) have been performed in reduced animal preparations. However, recent work using repetitive transcranial stimulation has shown that it is also possible to produce lasting effects on physiology and behaviour in humans that may involve changes in synaptic function within the brain (Chen et al., 1997; Nitsche & Paulus, 2000; Stefan et al., 2000; Peinemann et al., 2004). Although most of the effects induced in conscious humans last for a shorter time than those described in animals, several lines of evidence, e.g. the dependence on the N-methyl-D-aspartic acid (NMDA) receptor and its interaction with motor learning, suggest that the underlying mechanisms are similar to those of synaptic LTP/LTD (Nitsche et al., 2004; Ziemann et al., 2004; Huang et al., 2007; Teo et al., 2007; Gentner et al., 2008; Huang et al., 2008).
Experiments in animals have shown that synaptic plasticity is controlled by several mechanisms. Thus, the ease with which it is possible to produce LTP/LTD depends on the previous history of activity in the system such that LTD is easier to produce after a history of high activity and LTP after a history of low activity (Bienenstock et al., 1982). This type of control is often termed homeostatic plasticity. A second type of control determines how long the LTP/LTD may last. Depotentiation/de-depression describes a mechanism by which changes in plasticity that have been induced by successful protocols are abolished by a second protocol. Thus a high frequency conditioning protocol might induce LTP which would then be abolished by subsequent application of a short period of low frequency stimulation (Larson et al., 1993; Kulla & Manahan-Vaughan, 2000; Huang et al., 2001). Importantly, the latter would have no effects on LTP/LTD when applied alone. Data show that the ease of reversing LTP/LTD is greatest if the reversing paradigms are applied immediately after the initial plasticity protocol and become gradually less successful with longer intervals (Larson et al., 1993; Staubli & Scafidi, 1999; Chen et al., 2001), suggesting that LTP/LTD “consolidates” over time. The phenomena of depotentiation/de-depression may account for retrograde interference with behavioural learning. In addition abnormal reversibility of LTP/D has been shown in animal models of Parkinson’s disease with drug-induced dyskinesia (Picconi et al., 2003), DYT1 dystonia (Martella et al., 2009) and Huntington’s disease (Murphy et al., 2000; Picconi et al., 2006).
The present experiments sought to identify possible depotentiation/de-depression in the human brain using TMS of the motor cortex. We employed TBS of the motor cortex, which has been shown to produce LTP/LTD-like effects that outlast the period of stimulation for 20-60 minutes (Huang et al., 2005; Huang et al., 2009). We then tried to abolish the after effects by applying at different times afterwards a second very short period of TBS that when delivered alone had no effect on motor cortex excitability.
Healthy non-medicated subjects gave their informed consent prior to participation. The experiments were performed with the approval of the Institutional Review Board of the Chang Gung Memorial Hospital, Taiwan. All subjects were naïve to the effects of TBS and unaware of the differences between iTBS and cTBS protocols and control experiments (see below) were performed. It was not possible to study the same subjects in all experiments, but within each protocol all subjects were the same and hence we could directly compare their behavior in test and baseline trials.
One limitation of the designs we used was that the experimenter was not blinded to the type of stimulation. This is difficult to achieve in a TMS experiment since the protocols can be distinguished by the sound of the stimuli in the coil. A potential way around this would be to apply the depotentiating/de-depressing stimuli through a sham coil placed over (or under) a real coil that delivered the initial TBS patterns. Unfortunately we do not possess one of the new sham coils that is truly indistinguishable from a real device (see O’Reardon et al., 2007) and were therefore unable to provide this “gold standard” control. However, it is an important point, which should be taken into account when considering the results to this, and many other, TMS studies.
Subjects were seated in a comfortable chair. EMGs were recorded using Ag-AgCl electrodes from the right (the dominant hand in all subjects) first dorsal interosseus muscle (FDI). EMG activity was recorded with a gain of 1000 and 5000 and filtered with a band-pass filter (3 Hz to 2k Hz) through Digitimer D360 amplifiers (Digitimer Ltd, Welwyn Garden City, Herts, UK). Signals were recorded with a sampling rate of 5 kHz and stored on a personal computer for later analysis by Signal software (Cambridge Electronic Design Ltd., Cambridge, UK) through a Power 1401 data acquisition interface (Cambridge Electronic Design Ltd., Cambridge, UK). Trials in which the target muscle was not relaxed (as monitored by an oscilloscope) were rejected online, and that stimulus condition was repeated.
Magnetic stimulation was given using a hand-held figure of eight coil with loop diameters of 70 mm (Magstim Co., Whitland, Dyfed, UK). Single pulse TMS was delivered by a Magstim 2002 machine, and TBS was delivered using a Magstim Rapid2 stimulator. Stimulation was delivered over the motor hand area with the coil tangential to the scalp and the handle pointing in the posterior direction. The motor hand area was defined as the location on the scalp where magnetic stimulation produced the largest MEP from the contralateral FDI when the subject was relaxed (the “motor hot-spot”). The stimulation intensity of TBS was defined in relation to the active motor threshold (AMT) of the subject. The AMT was defined for each Magstim machine separately as the minimum intensity of single pulse stimulation required to produce an MEP of greater than 200μV on more than five out of ten trials from the contralateral FDI while the subject maintained a voluntary contraction of the FDI to about 20% maximum. Visual feedback of EMG level was provided to help subjects maintain a constant level of contraction.
The protocols used for TBS are based on those that we previously reported (Huang et al., 2005). They were comprised of bursts of 3 pulses at 50Hz at an intensity of 80% AMT repeated at 200ms intervals (i.e. at 5Hz) and were all given over the “motor hot-spot”. Four TBS paradigms were used in this study: 1) intermittent TBS (iTBS): a 2s train of TBS repeated every 10s for 20 repetitions to have 600 pulses in total, 2) continuous TBS (cTBS): a 20s train of uninterrupted TBS containing 300 pulses, 3) a shorter form of cTBS containing 150 pulses (cTBS150): a 10s train of uninterrupted TBS and, 4) a shorter form of iTBS containing 150 pulses (iTBS150): a 2s train of TBS repeated every 10s for 5 repetitions. iTBS was used to potentiate motor cortical excitability, while cTBS was used to depress excitability. cTBS150 and iTBS150 applied alone had no effect on cortical excitability and were designed to test the reversibility of potentiation and depression respectively. Each experiment was performed at least 1 week apart in a pseudo randomised order. In all the following experiments, the intensity of stimulation for MEP assessment was set to that required to produce an MEP of approximately 1mV in the baseline condition.
In animal studies, depotentiation or de-depression is usually induced by a protocol milder than those used to induce long-term depression (LTD) or LTP, and which on its own has no effect on naive cortex (for review, see Zhou & Poo, 2004). Thus, depotentiation is usually produced by applying an LTD-like protocol (e.g. low frequency stimulation ranging 1 to 5 Hz) (O’Dell & Kandel, 1994; Staubli & Scafidi, 1999; Huang et al., 2001) at lower stimulus intensity or for a shorter time than usual. We decided to use a shorter form of cTBS (i.e. cTBS150) to test the reversibility of the effects induced by iTBS for three reasons: 1) We did not employ a lower intensity protocol since we were unsure whether stimulation at a reduced intensity would activate the same circuits as those involved in the LTP-like effects of iTBS. This was also the reason for not using protocols that are believed to act via different mechanisms, e.g. transcranial direct current stimulation (tDCS) or paired associative stimulation (PAS), as the reversal paradigm. 2) We chose to use 150 pulses as the stimulus since a previous study showed that trains of TBS shorter than 75 pulses tend to be excitatory (Huang et al., 2005). A cTBS train of 150 pulses was therefore a compromise between too short a train causing facilitation and too long a train causing inhibition. 3) A final possibility would have been to use regular rTMS at a fixed rate. However, this usually requires a lengthy period of stimulation which would have blurred the time course of the interaction we observed. As regards the protocol for de-depression, although less well studied in animal experiments, intermittent high frequency stimulation has been commonly used (Chen et al., 2004; Kumar et al., 2007). Two shorter forms of TBS, i.e. iTBS150 and cTBS150, were tested as reversal stimulation for de-depression.
Eight subjects (1 man, 7 women; mean age, 33.3±10.3 years) were recruited for this experiment. We first tested the effect of cTBS150 on the size of MEPs. Baseline MEP was determined as the average of 30 MEPs evoked by a standard TMS pulse delivered every 4.5-5.5 seconds. cTBS150 was then given, and the effect on motor cortex excitability was assessed using single pulses of TMS delivered in trains of 12 pulses given every 4.5-5.5 seconds every 1 minute for 6 minutes then every 2 minutes until 21 minutes after the end of TBS.
In separate experiments we examined how the effect of iTBS was modulated by cTBS150 given 1 min after the end of iTBS. Subjects came for two sessions in a random order. 1) The iTBS session: baseline MEPs were evoked using 30 pulses delivered every 4.5-5.5 seconds. iTBS was then applied to the subject. MEP size was assessed using single pulses of TMS delivered in trains of 12 pulses given every 4.5-5.5 seconds every 1 minute for 6 minutes, then every 2 minutes until 21 minutes after the end of iTBS. 2) The depotentiation (DePo) session: the procedure was similar to that of the iTBS session, except that cTBS150 was given over the motor hot-spot one minute after the initial assessment of MEP after the end of iTBS. In addition, a control study (DePo-i) was tested in 8 subjects (2 men, 6 women; mean age, 28.7±3.6 years)using a protocol similar to that of the DePo session, in which cTBS150 was replaced by iTBS150. This was a control experiment to test whether any short period of stimulation could reverse the plasticity induced by iTBS. It also provides useful information that differentiates reversal of plasticity from homeostatic plasticity.
Seven subjects (4 men, 3 women; mean age, 28.7±3.6 years) were recruited for this experiment. We first tested the effect of iTBS150 on the size of MEP as described above for cTBS150.
Then subjects were asked to come for a further two sessions in a random order to examine the effect of cTBS on MEP size and how this was modulated by iTBS150 given 1 min after the end of cTBS. 1) The cTBS session: the effect of cTBS was assessed as described above for the iTBS session. 2) The de-depression (DeDe) session: the procedure was similar to that of the cTBS session except that iTBS150 was given over the motor hot-spot one minute after initial assessment of MEPs after the end of cTBS. In addition, a third session (DeDe-c) was tested using a protocol similar to that of the DeDe session, in which iTBS150 was replaced by cTBS150.
Eight subjects (3 men, 5 women; mean age, 32.1±3.8 years), who did not participate Experiment 1 and 2, participated in this experiment. Subjects came for 2 sessions: 1) DePoX session: cTBS150 was given 10 min after the end of iTBS and 2) DeDeX session: iTBS150 was given 10 min after the end of cTBS. Thirty baseline MEPs were recorded as in previous experiments. After iTBS or cTBS was given, MEP size was assessed in trains of 12 pulses given every 4.5-5.5 seconds every 1 minute for 6 minutes, then every 2 minutes until 21 minutes. In between, cTBS150 or iTBS150 was given 10 min after the end of iTBS or cTBS, respectively.
Given that short-interval intracortical inhibition (SICI) and intracortical facilitation (ICF) are modified by TBS (Huang et al., 2005; Huang et al., 2009), we tested if the reversal of the effects on MEP size could also be observed on SICI and ICF. SICI and ICF were tested before and at 5 min after the end of the TBS protocols in this part of the experiment. The conditioning stimulus was given at 80% AMT. The test stimulus was set to produce an MEP of 1 mV and was not adjusted after TBS conditioning since we expected no change in MEP size after the TBS protocols in this experiment. Subjects received in a random order either the test stimulus alone or conditioning-test stimuli at inter-stimulus interval (ISI) of 2, 3, 4, 7, 10 and 12 ms for a total of 10 trials per condition. Six subjects (2 men, 4 women; mean age, 30.3±3.1 years) came for 4 sessions with different TBS protocols: iTBS150, cTBS150, DePo (cTBS150 given at 1 min after iTBS) and DeDe (iTBS150 given at 1 min after cTBS).
Data were analyzed using SPSS. One way repeated measures ANOVA on the absolute amplitude of MEPs or two way repeated measures ANOVA on MEP amplitudes normalised to the pre-TBS baseline amplitude were used to examine the time course of changes in MEP amplitude and to test the effects of cTBS or iTBS and the influence of the cTBS150 or iTBS150 interventions. SICI and ICF before and after conditioning were compared with two-way repeated measures ANOVA. The data were also separated on a priori grounds into SICI (ISI of 2 and 3ms) and ICF (ISI of 10 and 12ms) and compared with t-tests. A P<0.05 was considered statistically significant.
A one-way ANOVA showed that cTBS150 and iTBS150 given alone did not change the size of MEPs (F(14,98)=0.601, p=0.859; F(14,84)=0.660, p=0.806, respectively). (Fig 2)
Figure 3a compares the after-effect of applying iTBS alone with iTBS followed by cTBS150 at 1 min (Depo) and iTBS followed by iTBS150 at 1 min (Depo-i). When it was applied alone, iTBS facilitated MEPs (one way ANOVA of all time points for iTBS, F(14,98)=1.799, p=0.049). Addition of cTBS150 abolished the facilitation. This was confirmed in the two way ANOVA of the time points after application of cTBS150 which showed a significant main effect of PATTERN (iTBS and DePo) (F(1,7)=11.712, p=0.011). Paired t-tests demonstrated that there was no difference between the initial amount of facilitation in the first minute after iTBS (t=0.689, p=0.513). In contrast to the effect of cTBS150, application of iTBS150 did not have any effect on the response to iTBS (two way ANOVA: no significant main effect of PATTERN (F(1,14)=0.025, p=0. 877) or PATTERN * TIME interaction (F(12,168)=0.689, p=0.761).
Figure 3b compares the result of cTBS alone with that of cTBS followed by iTBS150 (DeDe) and cTBS followed by cTBS150 (DeDe-c). When it was applied alone, cTBS suppressed MEPs (one way ANOVA of all time points for cTBS, F(14,84)=2.666, p=0.003). Addition of iTBS150 abolished the suppression. This was confirmed in the two way ANOVA of the time points after application of iTBS150 which showed a significant main effect of PATTERN (cTBS vs. DeDe) (F(1,6)=8.284, p=0.028). In contrast to the effect of iTBS150, application of cTBS150 did not have any effect on the response to cTBS (two way ANOVA: no significant main effect of PATTERN (F(1,6)=0.263, p=0. 626) or PATTERN * TIME interaction (F(11,66)=1.142, p=0.344)). One way ANOVA demonstrated that there was no difference between the initial amount of suppression in the first minute after cTBS in all three sessions (F(2,12)=0.697, P=0.517).
Figure 4a compares the result of DePoX session with that of iTBS session. There was neither a PATTERN (iTBS and DePoX) effect (F(1,14)=0.024, p=0.891) nor a PATTERN * TIME interaction (F(13,182)= 0.856, p=0.601) using two-way ANOVA. Further comparisons on the time points before and after cTBS150 were performed separately and confirmed that there was no significant difference between the two sessions. (F(1,14)=0.096, p=0.761; F(1,14)=0.220, p=0.646, respectively).
Figure 4b compares the result of DeDeX session with that of cTBS session. There was neither a PATTERN (cTBS and DeDeX) effect (F(1,13)=2.740, p=0.132) nor a PATTERN * TIME interaction (F(13,169)= 1.154, p=0.318) using two-way ANOVA. However, a further comparison on the time points after iTBS150 revealed a significant effect of PATTERN (F(1,13)=6.298, p=0.026) such that MEPs returned more quickly towards baseline after iTBS150. There was no significant PATTERN effect (F(1,13)=0.269, p=0.613) or PATTERN * TIME interaction (F(7,91)= 0.870, p=0.533) before the iTBS150 intervention.
The amplitude of control MEPs was not changed by cTBS150 (t=1.077, p=0.331 ), iTBS150 (t=0.925, p=0.397 ), DePo (t=-0.800, p=0.460 ) and DeDe (t=1.354, p=0.234). Figure 5 compares SICI/ICF before and after cTBS150 ((5a),5a), iTBS150 ((5b),5b), DePo ((5c)5c) and DeDe ((5d).5d). In each of the sessions, there was neither a main effect of CONDITION (before and after) (cTBS150: F(1,5)=0.000, p=0.989; iTBS150: F(1,5)=0.339, p=0.586; DePo: F(1,5)=0.287, p=0.615; DeDe: F(1,5)=0.044, p=0.842) nor a CONDITION * ISI (2, 3, 4, 7, 10 and 12ms) interaction (cTBS150: F(4,20)= 0.284, p=0.885; iTBS150: F(4,20)= 0.411, p=0.799; DePo: F(4,20)= 1.231, p=0.329; DeDe: F(4,20)= 1.208, p=0.338). Since the underlying mechanism is known to be different, we performed separate comparisons on SICI (at ISI of 2 and 3 ms), ICF (10 and 12ms) and the effect at an ISI of 7ms. Neither SICI, ICF nor the effect at an ISI of 7ms were changed by cTBS150 (p= 0.807,0.651 ,0.962 , respectively), iTBS150 (p= 0.161,0.766 ,0.782 , respectively), DePo (p= 0.482,0.958 ,0.319 , respectively) or DeDe (p= 0.541,0.451 ,0.383 , respectively). Note that in 5c,d SICI/ICF was first evaluated at baseline prior to any TBS. It was tested again after iTBS+cTBS150 or cTBS+iTBS150. At this time SICI/ICF was not different from baseline.
As reported previously (Huang et al., 2005; Huang et al., 2007; Huang et al., 2008), iTBS potentiated the amplitude of MEPs for approximately 20 min after the end of stimulation whereas cTBS depressed the amplitude of MEPs for 20 min or so. The shorter protocols of cTBS150 and iTBS150 had no after-effects on the size of MEPs nor on the amount of SICI and ICF. Despite this, both of them could interact with the longer lasting after effects of conventional cTBS/iTBS. Thus, cTBS150 abolished potentiation produced by iTBS if it was given 1 min after the end of iTBS; similarly, depression produced by cTBS was abolished when iTBS150 was delivered 1 min after the end of cTBS. Importantly there was no change in SICI/ICF measured at the time that the MEP amplitude had returned to the baseline level after depotentiation/de-depression. This suggests that all the effects of iTBS and cTBS reported previously (Huang et al., 2005) were abolished by the second stimulation and that the reversal of MEP size was not due to a shift in balance between SICI and ICF. Importantly, the effect of iTBS was not changed by applying a short session of iTBS150 rather than cTBS150 and the effect of cTBS was not changed by cTBS150, indicating that depotentiation and de-depression were specific to the nature of the second protocol. Finally, these effects depended on the time interval between protocols. Thus cTBS150 had no effect on facilitation if it was applied 10 min after iTBS. The reversal of depression was still present if iTBS150 was applied at 10 min after cTBS, but in contrast to the very quick reversal when iTBS150 was given at 1 min, it now took 5 min to build up.
As noted in the Methods, although subjects were blinded to the experimental design, the experimenters were not. This is a potential limitation that is difficult to overcome with present TMS technology. However, advances in coil designs that have been piloted in clinical studies of depression (O’Reaedon et al., 2007) may make this feasible in future studies.
Our hypothesis is that these effects are compatible with the phenomena of depotentiation and de-depression that have been described in experiments on brain slices in areas including hippocampus, cerebellum and corticostriatal pathways (Huang et al., 2001; Picconi et al., 2003; Coesmans et al., 2004; Kumar et al., 2007) as well as in the brains of freely moving animals (Staubli & Scafidi, 1999; Kulla & Manahan-Vaughan, 2000).
It is important to distinguish depotentiation and de-depression from the similar but unrelated phenomenon of synaptic metaplasticity, which describes how synaptic plasticity can be modulated by prior synaptic activity (Abraham & Bear, 1996). The latter involves giving a priming protocol that changes the plasticity induced by a subsequent protocol whereas depotentiation/de-depression require that plasticity is first induced by a standard protocol but is then abolished by a second protocol. In metaplasticity, the priming protocol may or may not cause a long term effect on the system under test (Abraham, 2008). On the contrary, the abolition protocol in depotentiation/de-depression on its own usually has no long term effect (Zhou & Poo, 2004). Metaplasticity prevents LTP/LTD from driving a system into instability whereas depotentiation/de-depression is thought to abolish previously acquired learning. For example, Xu et al. (1998) demonstrated that LTP induced in the hippocampus of freely-moving rats was completely erased when the rat entered a new environment.
There have been many studies in humans of the interaction between various types of rTMS and tDCS (transcranial direct current stimulation) paradigms, but all of these have been designed to test for metaplasticity rather than depotentiation/de-depression. In some cases, a weak priming protocol was followed by a reliable test protocol (Iyer et al., 2003; Hamada et al., 2008; Todd et al., 2009), but in many cases, each paradigm alone was capable of producing long term effects on cortical excitability (Siebner et al., 2004; Quartarone et al., 2005; Muller et al., 2007; Potter-Nerger et al., 2009).
As far as we are aware, there is only one previous study in humans in which an effective priming protocol was followed by an ineffective test protocol. Siebner et al (2004) studied the interaction between an effective tDCS protocol and a 15-min train of 1 Hz rTMS at 85% rest motor threshold which had no significant effect on MEPs when given alone. In this case, the 1 Hz rTMS did not simply abolish the effect of preceding tDCS, as expected from depotentiation/de-depression; instead it reversed the effect. Thus Siebner et al (2004) interpreted their results in terms of a homeostatic interaction between a potentiating or depressing pretreatment with tDCS and a potentially suppressive effect of 1 Hz rTMS. In fact, as Siebner et al (2004) pointed out, although 1Hz rTMS did not produce a significant effect on MEPs in their study, several other studies had found that very similar 1Hz rTMS protocols, sometimes at a slightly higher intensity (e.g. 90% rest motor threshold), produced long-term effects on MEP size (Romero et al., 2002; Lang et al., 2006) and SICI/ICF (Romero et al., 2002; Modugno et al., 2003). Indeed, this form of 1 Hz rTMS can have clinical benefits in patients with levodopa induced dyskinesia (Brusa et al., 2006).
It should be noted that in the present study, the protocols for abolishing previously induced plasticity (cTBS150 and iTBS150) are much milder and shorter than those used by Siebner et al (2004) and produced no visible effect on either MEPs or SICI/ICF. Thus we think that the present protocol is a close parallel to the depotentiation/de-depression protocols used in animal literature. We cannot exclude the possibility that the results are an example of “partial homeostatic” interactions in which a slightly more effective protocol would have led to a reversal rather than abolition of the priming effect. However, this is less likely since iTBS did not prime iTBS150 to produce inhibition and cTBS did not prime cTBS150 to produce facilitation or at least to block the ongoing potentiation from iTBS and depression from cTBS.
Experiments in reduced preparations have shown that depotentiation/de-depression have specific underlying cellular mechanisms. Depotentiation for example is not simply the summation of LTP and LTD. Studies have shown that some forms of depotentiation may activate protein phosphatase −1 and 2A and reverse the phosphorylation of Ser831 on the GluR1 subunit of the AMPA receptor caused by LTP induction and thereby stop ongoing LTP (Huang et al., 2001). In contrast, the Ser845 on GluR1, which is dephosphorylated to produce LTD, is unaffected during depotentiation (for review, see Zhou & Poo, 2004). Recent studies suggest that metaplasticity occurs through different mechanisms in which the priming stimuli modify NMDA receptors (NMDARs) and NMDARs-mediated calcium influx (Abraham, 2008) and may also change the activity of some types of potassium channel (Surmeier & Foehring, 2004).
As in animal studies, our experiments showed that reversibility of potentiation caused by iTBS is time-dependent. When cTBS150 was given at 10min, instead of 1min after iTBS, it was no longer able to abolish potentiation: the facilitated MEP remained unchanged for another 10 min and was not different to that observed in iTBS session alone. In contrast, depression induced by cTBS was still reversible by iTBS150 at 10 min after the end of cTBS, although there was a few minutes delay before MEPs returned to the baseline. This difference in the time course of the effects of cTBS150 and iTBS150 may relate to experiments in animals which have shown that the time window during which depotentiation may occur depends on a number of factors including the protocol used to induce LTP, the protocol for induction of depotentiation and the pathway under study: it may range from a few minutes (Larson et al., 1993; Huang et al., 1999) to tens of minutes or even more (Burette et al., 1997). The effect of the depotentiation decreases as the interval between LTP induction and delivery of depotentation stimuli is increased (Chen et al., 2001). A weaker depotentiation paradigm leads to a narrower time window for depotentiation (Staubli & Scafidi, 1999). Less is known about the time dependency of de-depression although it is usually assumed that it shares common features with depotentiation. Given these data, our current results could be explained by postulating that the baseline LTP/LTD-like effects of iTBS/cTBS are equally powerful but that the reversing effect of cTBS150 is slightly stronger than iTBS150. Thus, iTBS150 cannot reverse cTBS at 10min whereas cTBS could reverse iTBS effects at a similar timing.
It has been suggested that reversal of LTP and LTD in the adult might be important in preventing acquisition of inappropriate learning and might contribute to the initial instability of memory before final consolidation (Huang & Hsu, 2001); it may also play a role in developmental refinement of neural circuits (Zhou & Poo, 2004). In the motor system it is well known that motor memories are unstable after a period of learning and vulnerable to interference by subsequent practice of a different task (Walker et al., 2003; Robertson et al., 2004; Krakauer & Shadmehr, 2006). This is compatible with the properties of the reversibility of LTP and LTD. However, the direct connection between synaptic depotentiation and de-depression and interference learning has not been studied. It may be that the paradigms developed in the present study would be helpful in understanding such linkage and probing the underlying physiological implications of the reversal of synaptic plasticity.
In conclusion, we have shown for the first time that it is possible to reverse LTP/LTD-like effects in the conscious human brain with short bursts of stimulation that on their own have no apparent effect on cortical excitability. We suggest that this is an example of the phenomena of depotentiation/de-depression described in the animal literature. This will be helpful to probe natural processes of memory consolidation and forgetting. Moreover, animal models of several pathological conditions are characterised by an abnormality of mechanisms that reverse plasticity such as PD with levodopa induced dyskinesia, DYT1 dystonia and Huntington’s disease (Murphy et al., 2000; Picconi et al., 2003; Picconi et al., 2006; Martella et al., 2009). The present methods may be able to probe whether similar abnormalities can be demonstrated in human patients with these conditions, and could be used as a tool to evaluate the therapeutic benefit of other procedures on these conditions (Koch et al., 2005).
Plasticity describes the fact that the nervous system, in particularly the brain, is flexible in response to internal and external demands. This flexibility arises because connections between circuits can be potentiated or depressed, thus allowing the system to learn new tasks, or to re-organise after injury. In animal studies, such potentiation/depression can be reproduced by artificial stimulation of nerve pathways; importantly, it can be abolished by a second mild stimulation applied within a certain time window after the plasticity is induced. Behaviourally this might correspond to the fact that it easier to reverse bad habits if we act before they stabilise, whereas it is much more difficult at a later time. Protocols to test early reversal of plasticity have never been demonstrated experimentally in conscious humans. The present experiments identify possible method to investigate reversal of plasticity in the human brain using transcranial magnetic stimulation of the motor cortex.
The authors would like to thank the National Science Council of Taiwan (Contract Nos. NSC 94-2314-B-182A-057/ NSC 95-2314-B-182A-012/ NSC 96-2314-B-182A-003) and Royal Society of the UK (UK-Taiwan Joint Project Grant) for financially supporting this research.